Public health in Vietnam: here's the data, where's the action ?

Supplement 2, 2013

Public health in Vietnam: here's the data, where's the action ?

Guest Editors: Nguyen Duc Hinh, Hoang Van Minh and Lucia D'Ambruoso

Global Health Action

Global Health Action (www.globalhealthaction.net) is an international peer-reviewed Open Access journal affiliated to the Umeå Centre for Global Health Research (UCGHR), Sweden, and published by Co-Action Publishing. The UCGHR (www.globalhealthresearch.net) is associated with the Division of Epidemiology and Global Health, at the Department of Public Health and Clinical Medicine, Umeå University.

Public health challenges in a global context are to be found in the widening gap between the winners and losers of globalisation. To meet these challenges it is crucial not only to act constructively on what is already known and evaluate the results, but also to establish what we have yet to learn and still need to implement. The Journal therefore specifically welcomes papers that report on results and evidence derived from practical implementations of current knowledge, but also papers suggesting strategies for practical implementations where none already exist. Thus the aim of Global Health Action is to contribute to fuelling a more concrete, hands-on approach to global health challenges. The journal particularly invites articles from low- and mid-income countries, while also welcoming South-South and South-North collaborations. All papers are expected to address a global agenda and include a strong implementation or policy component.

Global Health Action welcomes Original Articles, Short Communications, Debate Articles, as well as PhD Reviews in which recently graduated PhD students review their thesis. Study Design Articles, for example on longitudinal infrastructures for population research, may well precede the publication of results. If manuscripts are received that are not up to the standard required for publication in the journal but relate to work that has publication potential, the author(s) will be offered mentorship from a senior academic who will work collaboratively to make the article publishable. In such a case, the mentor would not expect to become a co-author but would be named in the published article as a Contributing Editor.

EDITORIAL TEAM
Chief Editor:
Stig Wall
Umeå University, Sweden
Email: stig.wall@epiph.umu.se

Deputy Editor:
Peter Byass
Umeå University, Sweden
Email: peter.byass@epiph.umu.se

Managing Editor:
Nawi Ng
Umeå University, Sweden
Email: nawi.ng@epiph.umu.se
Editors:
Lucia D’Ambruoso, Umeå University, Sweden
Maria Emmelin, Umeå University, Sweden
Malin Eriksson, Umeå University, Sweden
Anneli Ivarsson, Umeå University, Sweden
Lars Lindholm, Umeå University, Sweden
Nawi Ng, Umeå University, Sweden
Joacim Rocklöv, Umeå University, Sweden

INTERNATIONAL ADVISORY BOARD
Tedros Adhanom Ghebreyesus, Addis Ababa, Ethiopia
Heiko Becher, Heidelberg, Germany
Ruth Bonita, Auckland, New Zealand
Vinod Diwan, Stockholm, Sweden
Wendy Graham, Aberdeen, Scotland
Jane Menken, Colorado, USA
Tom Pearson, Rochester, USA
Stig Pramming, London, UK
Osman Sankoh, Accra, Ghana
Rainer Sauerborn, Heidelberg, Germany
Krisela Steyn, Cape Town, South Africa

PUBLICATION INFORMATION

Global Health Action, ISSN 1654-9880 (online), Supplements ISSN 1654-9880 (online), 1654-9716 (print), is an Open Access journal and applies the Creative Commons Attribution-NonCommercial 3.0 Unported License. Under this license, copyright is retained by the author who grants to users the right to download, reuse, reprint, modify, distribute and copy articles for any non-commercial purpose as long as the original author(s) and source are properly cited. No permission is required from the publishers or author(s).

All articles are posted online immediately they are ready for publication and will be assigned a DOI number (Digital Object Identifier) whereby they become searchable and citable without delay.

Global Health Action full text articles are accessible via the journal’s website at www.globalhealthaction.net as well as via international search engines such as Google Scholar.

Global Health Action is indexed/tracked/covered by the following services: CABI Global Health Collection, CINAHL, Current Contents/Social and Behavioral Sciences, DOAJ, EMCare, Google Scholar, HINARI, Intute: Health and Life Sciences collection, J-gate, JournalSeek, Medline, PubMed, Science Citation Index Expanded, SCIMago, Social Sciences Citation Index, Ulrich’s Periodicals Directory

To submit manuscripts and for information about publication fees, see www.globalhealthaction.net

For information on reprints, advertisements and other commercial sales, please contact info@co-action.net

Typeset by Datapage (India) Private Limited, Chennai, India
Printed and bound by Hobbs the Printer, Southampton, UK

Cover design: Julie Bindslev

image

The paper used in this publication contains pulp sourced from forests independently certified to the Forest Stewardship Council (FSC) principles and criteria. Chain of custody certification allows the pulp from these forests to be tracked to the end use (see www.fsc-uk.org).

 

Global Health Action

Supplement 2, 2013

CONTENTS

Foreword

Lars Weinehall

Foreword

Nguyen Duc Hinh

Acknowledgment

Editorial

Public health in Vietnam: scientific evidence for policy changes and interventions

Nguyen Duc Hinh and Hoang Van Minh

Hot spot detection and spatio-temporal dispersion of dengue fever in Hanoi, Vietnam

Do Thi Thanh Toan, Wenbiao Hu, Pham Quang Thai, Luu Ngoc Hoat, Pamela Wright and Pim Martens

Assessing the household financial burden associated with the chronic non-communicable diseases in a rural district of Vietnam

Hoang Van Minh and Bach Xuan Tran

Survival probability and prognostic factors for breast cancer patients in Vietnam

Nguyen H. Lan, Wongsa Laohasiriwong and John F. Stewart

Cost of treatment for breast cancer in central Vietnam

Nguyen Hoang Lan, Wongsa Laohasiriwong, John Frederick Stewart, Nguyen Dinh Tung and Peter C. Coyte

Knowledge of the health consequences of tobacco smoking: a cross-sectional survey of Vietnamese adults

Dao Thi Minh An, Hoang Van Minh, Le Thi Huong, Kim Bao Giang, Le Thi Thanh Xuan, Phan Thi Hai, Pham Quynh Nga and Jason Hsia

Alcohol consumption and household expenditure on alcohol in a rural district in Vietnam

Kim Bao Giang, Hoang Van Minh and Peter Allebeck

Alcohol-related harm among university students in Hanoi, Vietnam

Pham Bich Diep, Ronald A. Knibbe, Kim Bao Giang and Nanne De Vries

Road traffic injury among young people in Vietnam: evidence from two rounds of national adolescent health surveys, 2004–2009

Linh Cu Le, and Robert W. Blum

Factors associated with health risk behavior among school children in urban Vietnam

Tran Bich Phuong, Nguyen Thanh Huong, Truong Quang Tien, Hoang Khanh Chi and Michael P. Dunne

Exploring quality of life among the elderly in Hai Duong province, Vietnam: a rural—urban dialogue

Nguyen Thanh Huong, Le Thi Hai Ha, Nguyen Thai Quynh Chi, Peter S. Hill and Tara Walton

Social capital and mental health among mothers in Vietnam who have children with disabilities

Nguyen Thi Minh Thuy and Helen L. Berry

The association and a potential pathway between gender-based violence and induced abortion in Thai Nguyen province, Vietnam

Phuong Hong Nguyen, Son Van Nguyen, Manh Quang Nguyen, Nam Truong Nguyen, Sarah Colleen Keithly, Lan Tran Mai, Loan Thi Thu Luong and Hoa Quynh Pham

An analysis of interprovincial migration in Vietnam from 1989 to 2009

Le Thi Kim Anh, Lan Hoang Vu, Bassirou Bonfoh and Esther Schelling

Handwashing among schoolchildren in an ethnically diverse population in northern rural Vietnam

Le Thi Thanh Xuan and Luu Ngoc Hoat

Factors associated with job satisfaction among commune health workers: implications for human resource policies

Bach Xuan Tran, Minh Van Hoang and Nguyen Duc Hinh

 

In addition to the mentorship and editing provided by Guest Editors, Nguyen Duc Hinh, Hoang Van Minh and Lucia D’Ambruoso, each paper has been subjected to regular peer review.

 

FOREWORD

Published: 25 February 2013

Glob Health Action 2013. © 2013 Lars Weinehall. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 20446 - http://dx.doi.org/10.3402/gha.v6i0.20446

 

Bilateral research co-operation and research training between Vietnam and Sweden, aimed at focusing on important development issues, began in the 1970s. Today, global health and health impacts of climate change are on top of the agenda:

  • Vietnam is experiencing a double burden of diseases, including both infectious (HIV/AIDS, bird flu, dengue fever, food poisoning, etc.) and non-communicable diseases (cancer, cardiovascular diseases, etc.), and is predicted to be greatly impacted by climate change in the future.
  • Sweden also faces significant health challenges, in terms of chronic diseases, and will probably at the same time – like many similar countries in the West – become greatly affected by climate change.

Research cooperation between the two countries is essential to develop common knowledge and to find strategies to move from words to action. Therefore, the collaborative research spanning more than 10 years between the Hanoi Medical University (HMU), Vietnam, and Epidemiology and Global Health (EGH), Umeå University, Sweden, plays an important role. Over the years, a number of staff members from EGH have visited Hanoi as teachers, while HMU researchers have lectured in Umeå.

Scientific workshops have been conducted both in Vietnam and Sweden, and more than 20 staff members from HMU have completed their PhD or Master of Public Health at Umeå University. Some have also spent postdoctoral periods in Umeå. This cooperation has produced more than 60 scientific papers that have been jointly published in international peer-reviewed journals.

This supplement represents a new milestone in the long-term cooperation between HMU and EGH. Supported by funding from the Swedish International Development Cooperation Agency (SIDA), 15 scientific papers have been published, with special focus on how new scientific knowledge can be transformed into policy, thereby helping to improve people's living conditions. The studies represent important steps in our collective effort to implement public health research in public health policy and thereby, hopefully, reduce the know–do gap.


Lars Weinehall
Head of Epidemiology and Global Health
Umeå University
Umeå, Sweden

 

FOREWORD

Published: 25 February 2013

Glob Health Action 2013. © 2013 Nguyen Duc Hinh. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 20433 - http://dx.doi.org/10.3402/gha.v6i0.20433

 

I am very pleased to introduce the Global Health Action cluster of scientific papers on ‘Public health in Vietnam: here's the data, where's the action?’ by the Hanoi Medical University, Hanoi, Vietnam, and Epidemiology and Global Health, Umeå University, Umeå, Sweden. The objectives of the supplement are to: 1) provide scientific evidence for policy changes and interventions; 2) share research findings on various critical public health issues from Vietnam with the rest of the world; and 3) improve the capacity of Vietnamese scientists for writing international scientific articles.

This cluster, consisting of 15 papers, not only provides up-to-date scientific evidence on the situation of health status and health care in Vietnam but also suggests recommendations for policy and public health actions to improve these situations. The cluster, one of a few clusters of peer-reviewed scientific papers from Vietnam, demonstrates that international scientific publications are both necessary for and possible from Vietnam.

I wish to thank the Swedish International Development Cooperation (Sida), through the Swedish Embassy in Hanoi, for funding this supplement. I acknowledge Dr. Lucia D'Ambruoso, Prof. Stig Wall, Assoc. Prof. Nawi Ng, Global Health Action, Sweden, Prof. Ta Thanh Van and Assoc. Prof. Hoang Van Minh, Assoc. Prof. Kim Bao Giang, Hanoi Medical University, Vietnam for their editorial support and coordination. I am grateful to Prof. Lars Weinehall from Epidemiology and Global Health, Umeå University, for his technical and managerial contribution to the preparation of this cluster. I deeply appreciate the continued efforts by the authors, mentors and reviewers who have made these papers possible.

I sincerely hope that this cluster of papers will inspire Vietnamese scientists to do more research for international peer-reviewed publications. I also expect that the results published in this supplement will result in proactive public health actions to respond to public health issues in Vietnam and elsewhere!

Nguyen Duc Hinh
President of the Hanoi Medical University
Hanoi, Vietnam

 

Acknowledgment

Published: 25 February 2013

Glob Health Action 2013. © 2013 Glob Health Action. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 20434 - http://dx.doi.org/10.3402/gha.v6i0.20434

 

The cluster of scientific papers ‘Public health in Vietnam: here's the data, where's the action?’ was published with the support of the Sida-funded project ‘Public Health Preparedness and Response to Critical Global Health Issues in Vietnam and Sweden’ which is being implemented by the Center for Health System Research, Hanoi Medical University, Hanoi, Vietnam, and Epidemiology and Global Health, Umeå University, Umeå, Sweden. The cluster was only possible because of the hard work carried out by the editors, mentors, and reviewers*:


Ann Fitzmaurice University of Aberdeen, UK
Fredrik Norström Umeå University, Sweden
Heiko Becker University of Heidelberg, Germany
Hans Stenlund Umeå University, Sweden
Hoang Van Minh Hanoi Medical University, Vietnam
Irina Campbell NJ Child Health Project, USA
Jenni Hislop Newcastle University, UK
Joacim Rocklöv Umeå University, Sweden
Kim BaoGiang Hanoi Medical University, Vietnam
Lars Lindholm Umeå University, Sweden
Lars Weinehall Umeå University, Sweden
Lena Wistrand Co-Action Publishing, Sweden
Lennarth Nyström Newcastle University, UK
Laura Ternent Newcastle University, UK
Lucia D'Ambruoso Umeå University, Sweden
Malin Eriksson Umeå University, Sweden
Margareta Norberg Umeå University, Sweden
Maria Nilsson Umeå University, Sweden
Marie Lindkvist Umeå University, Sweden
Nawi Ng Umeå University, Sweden
Nguyen QuynhAnh Hanoi School of Public Health, Vietnam
Nguyen Thanh Huong Hanoi School of Public Health, Vietnam
Nguyen Van Huy Hanoi Medical University, Vietnam
Pham NganGiang Ministry of Health of Vietnam
Barbara Schumann Umeå University, Sweden
Stig Wall Umeå University, Sweden
Ta Thanh Van Hanoi Medical University, Vietnam
Tran Thanh Huong Hanoi Medical University, Vietnam
Tran Xuan Bach Hanoi Medical University, Vietnam

*All editors, mentors, and reviewers involved in the peer review process had no competing interests or involvement with the papers to which they contributed as reviewers at the pre-submission stages.

 

EDITORIAL

Public health in Vietnam: scientific evidence for policy changes and interventions

Published: 25 February 2013

Glob Health Action 2013. © 2013 Nguyen D. Hinh and Hoang Van Minh. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 20443 - http://dx.doi.org/10.3402/gha.v6i0.20443

 

Vietnam has made impressive progress toward improving the health status of the population, with progress that equals or surpasses that of many neighboring countries. Life expectancy in Vietnam is 72.8 years (70.2 for men and 75.6 for women), a level that is considerably higher than that in many countries with similar levels of GDP per capita. From 1990 to 2009, the infant mortality rate fell from 44.4% to 16.0%, the under-five mortality rate dropped from 58.0% to 24.5%, and the maternal mortality ratio declined from 233 to 69 maternal deaths per 100,000 live births. Estimated to be around 18% in 2010, the rate of under-five malnutrition has also fallen dramatically. These improvements are attributable to a widespread health care delivery network, increasing numbers of qualified health workers, and expanding national public health programs (1, 2).

Although many significant achievements have been made, Vietnam's health care system still faces many difficulties and challenges. Recent health sector reviews have identified a number of health issues in this regard. These include an emerging a double burden of noncommunicable diseases (cardiovascular diseases, cancer, diabetes, etc.) and infectious diseases (HIV/AIDS, H1A1, etc.), an ageing population, inadequate capacity of the health system, and problems of inequities in access to health and health care (16). The findings from this cluster of papers provide further insights into today's health issues in Vietnam.

The fact that infectious diseases remain a public health concern is illustrated by Toan et al. in a study demonstrating that Dengue fever remains pervasive and that the geographical scope of the disease has expanded in recent years (7). Vietnam is also undergoing epidemiological transition whereby the burden of disease attributable to chronic noncommunicable conditions is rising rapidly. Minh et al.'s research demonstrates that chronic diseases are highly prevalent in rural populations and that households are more likely to face catastrophic health expenditure and impoverishment for chronic noncommunicable disorders (8). Lan et al. observe that the survival probability for breast cancer is lower in Vietnam than it is in countries with similar distributions of stage at diagnosis (9). Furthermore, a study led by Lan highlights the substantial costs of breast cancer treatment in central Vietnam, especially among patients who lack health insurance (10).

Smoking is common in Vietnam, partly due to the general population's poor knowledge of its health consequences. An et al. demonstrate that adult smokers, especially those belonging to ethnic minority groups, have low levels of knowledge regarding the harmful effects of smoking. Alcohol misuse is also a rising public health problem (11). In their study, Giang et al. report on common problems related to alcohol consumption among men in Vietnam and report that the share of total household expenditure on alcohol is remarkable, especially among poorer households (12). Diep et al. also suggest that alcohol-related harm presents a serious public health problem among young and educated individuals (13). Linh et al. (14) discuss the relationship between alcohol use and road traffic accidents, and Phuong et al. (15) demonstrate the association between harmful use of alcohol and suicidal thoughts.

Vietnam is also experiencing a rapidly ageing population. It is therefore critical to have an in-depth understanding about quality of life and associated factors among elderly groups. In their study on ageing populations, Huong et al. find that quality of life among the elderly is strongly correlated with issues related to finances and economics, as well as with social relationships and familial support (16).

In the coming years, equity-oriented reform will be a major focus for the health system in Vietnam. Despite substantial achievements, there are still large health status disparities across regions and between demographic and socioeconomic groups. In this regard, Thuy et al. illustrate the extreme marginalization and distress among Vietnamese mothers whose children have disabilities. These authors identify modest levels of social capital among this population group, although relatively better mental health is also detected (17). Otherwise, Phuong et al. find a high prevalence of gender-based violence in Vietnam. This study observes that abused women are more likely than nonabused women to report contraceptive use and unintended pregnancies and that these factors are in turn associated with increased risks of induced abortion (18). In addition, Anh et al. observe the effects of unequally growing Vietnamese labor markets on migration and identify corresponding infrastructure improvements and public service needs in these areas. Analysis of migration can provide useful information for planning health and social services and for policymaking for national economic development (19).

Vietnam has over eight million people belonging to ethnic minorities, the majority of which live in remote and mountainous areas. These populations are relatively more disadvantaged in terms of socioeconomic and health status. The study by Xuan et al. finds poor hand washing with soap behaviors among schoolchildren in a multi-ethnic population of Vietnam, a potential cause of a number of health conditions (20). Human resources for health are an important building block of health system. The number of health workers in Vietnam has increased substantially over the past 10 years, but there are still severe shortages in remote and disadvantaged areas. In their study, Bach et al. find generally low levels of work-related satisfaction among of primary health care staff, particularly regarding salary and incentives, equipment, and the working environment. Predictors of job satisfaction identified by these authors include age, areas of work and expertise, professional education, location, and the sufficiency of staffing (21).

Based on the empirical evidence, all contributing authors have developed recommendations for policy changes and interventions in Vietnam. These recommendations are comprehensive and include primary, secondary and tertiary approaches, as well as policy-level interventions (721). Both the findings and the policy recommendations documented in these papers are highly relevant to health system stakeholders in Vietnam. The evidence is intended to help health system stakeholders, especially health policy makers and managers, to understand the implementation and impact of the policies and interventions that they introduce. The recommendations are intended to provide health system stakeholders with more options as they change or refine these measures.

Policy makers, managers, health staff and other health system stakeholders in Vietnam are committed to ensuring that all people attain a level of health that enables them to participate actively in the social and economic life of the communities in which they live. An important factor that can help the health system achieve this goal is the availability and quality of information on which decisions are based. As academics and scientists, we have conducted research to generate robust scientific evidence to support health planning and decision making in Vietnam. We hope that health system stakeholders will find this cluster of papers useful. We enthusiastically stress that scientific evidence on health is crucial for policy changes and interventions and, when the evidence is compelling, actions toward better health and health care should be taken.

Nguyen Duc Hinh
President of the Hanoi Medical University, Hanoi Vietnam
Head of the Department of Medical Ethics and Social Medicine
Institute for Preventive Medicine and Public Health
Center for Health System Research
Hanoi Medical University, Hanoi Vietnam

Hoang Van Minh
Lecturer
Department of Health Economics
Institute for Preventive Medicine and Public Health
Center for Health System Research
Hanoi Medical University, Hanoi Vietnam

References

  1. Ministry of Health of Viet Nam, Health Partnership group. Join annual health review 2011. Hanoi: Ministry of Health of Vietnam; 2011.
  2. Nam MoHoV. Plan for the protection, care and promotion of the people's health 2011–2015. Hanoi: Ministry of Health of Vietnam; 2010.
  3. Ministry of Health of Viet Nam, Health Partnership group. Join annual health review 2008. Hanoi: Ministry of Health of Vietnam; 2008.
  4. Ministry of Health of Viet Nam, Health Partnership group. Join annual health review 2009. Hanoi: Ministry of Health of Vietnam; 2009.
  5. Ministry of Health of Viet Nam, Health Partnership group. Join annual health review 2010. Hanoi: Ministry of Health of Vietnam; 2010.
  6. Partnership for Action in Health Equity (PAHE). Health equity in Vietnam: a civil society perspective. Hanoi: PAHE; 2011.
  7. Toan DT, Hu W, Thai PQ, Hoat LN, Wright P, Martens P. Hot spot detection and spatiotemporal dispersion of dengue fever in Hanoi, Vietnam. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18632. Publisher Full Text
  8. Minh HV, Bach TX. Assessing the household financial burden associated with the chronic non-communicable diseases in a rural district of Vietnam. Glob Health Action 2013; 5. DOI: 10.3402/gha.v5i0.18892. Publisher Full Text
  9. Lan NH, Laohasiriwon W, Stewart JF. Survival probability and prognostic factors for mortality in patients with breast cancer in Vietnam. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18860. Publisher Full Text
  10. Lan NH, Laohasiriwong W, Stewart JF, Tung ND, Coyte PC. Cost of treatment for breast cancer in Central Vietnam. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18872. Publisher Full Text
  11. An DTM, Minh HV, Huong LT, Hai PT, Giang KB, Xuan LTT, et al. Knowledge of health consequences of tobacco smoking: a cross-sectional survey of Vietnamese adults. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18707. Publisher Full Text
  12. Giang KB, Minh HV, Allebeck P. Alcohol consumption and household expenditure on alcohol in a rural district in Viet Nam. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18937. Publisher Full Text
  13. Diep PB, Knibbe RA, Giang KB, Vries ND. Alcohol- related harm among university students in Hanoi, Viet Nam. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18857. Publisher Full Text
  14. Linh LC, Blum RW. Road traffic injury of young people in Vietnam: perspective and evidence from two rounds of national adolescent health surveys, 2004–2009. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18757. Publisher Full Text
  15. Phuong TB, Huong NT, Tien TQ, Chi HK, Dunne MP. Factors associated with health risk behavior among school children in urban Vietnam. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18876. Publisher Full Text
  16. Huong NT, Ha LTH, Chi NTQ, Hill PS, Walton T. Exploring quality of life among the elderly in Hai Duong province Viet Nam: a rural- urban dialogue. Glob Health Action 2013; 5. DOI: 10.3402/gha.v5i0.18874. Publisher Full Text
  17. Thuy NTM, Berry HL. Social capital and mental health among mothers in Vietnam who have children with disabilities. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18886. Publisher Full Text
  18. Phuong NH, Son NV, Manh NQ, Nam NT, Keithly S, Lan TM, et al. The association and a potential pathway between gender-based violence and induced abortion in Thai Nguyen province, Vietnam. Glob Health Action 2013; 5. DOI: 10.3402/gha.v5i0.19006. Publisher Full Text
  19. Anh LTK, Lan VH, Bonfoh B, Schelling E. An analysis of inter-provincial migration in Viet Nam from 1989–2009. Glob Health Action 2013; 5. DOI: 10.3402/gha.v5i0.9334. Publisher Full Text
  20. Xuan LTT, Hoat LN. Handwashing among school children in an ethnically diverse population in Northern rural Vietnam. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18869. Publisher Full Text
  21. Xuan BT, Minh HV, Hinh ND. Factors associated with job satisfaction among commune health workers: Implications for human resource policies. Glob Health Action 2013; 6. DOI: 10.3402/gha.v6i0.18619. Publisher Full Text

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Hot spot detection and spatio-temporal dispersion of dengue fever in Hanoi, Vietnam

Do Thi Thanh Toan1*, Wenbiao Hu2, Pham Quang Thai3, Luu Ngoc Hoat1, Pamela Wright4 and Pim Martens5

1Institute of Training for Preventive Medicine and Public Health, Hanoi Medical University, Hanoi, Vietnam; 2School of Population Health, The University of Queensland, Queensland, Australia; 3National Institute of Hygiene and Epidemiology of Vietnam, Hanoi, Vietnam; 4The Medical Committee Netherlands – Vietnam, Amsterdam, the Netherlands; 5International Centre for Integrated assessment and Sustainable development, Maastricht University, Maastricht, the Netherlands

Abstract

Introduction: Dengue fever (DF) in Vietnam remains a serious emerging arboviral disease, which generates significant concerns among international health authorities. Incidence rates of DF have increased significantly during the last few years in many provinces and cities, especially Hanoi. The purpose of this study was to detect DF hot spots and identify the disease dynamics dispersion of DF over the period between 2004 and 2009 in Hanoi, Vietnam.

Methods: Daily data on DF cases and population data for each postcode area of Hanoi between January 1998 and December 2009 were obtained from the Hanoi Center for Preventive Health and the General Statistic Office of Vietnam. Moran's I statistic was used to assess the spatial autocorrelation of reported DF. Spatial scan statistics and logistic regression were used to identify space–time clusters and dispersion of DF.

Results: The study revealed a clear trend of geographic expansion of DF transmission in Hanoi through the study periods (OR 1.17, 95% CI 1.02–1.34). The spatial scan statistics showed that 6/14 (42.9%) districts in Hanoi had significant cluster patterns, which lasted 29 days and were limited to a radius of 1,000 m. The study also demonstrated that most DF cases occurred between June and November, during which the rainfall and temperatures are highest.

Conclusions: There is evidence for the existence of statistically significant clusters of DF in Hanoi, and that the geographical distribution of DF has expanded over recent years. This finding provides a foundation for further investigation into the social and environmental factors responsible for changing disease patterns, and provides data to inform program planning for DF control.

Keywords: dengue fever; hotspots; dynamic dispersion; Hanoi; Vietnam

Received: 25 April 2012; Revised: 21 December 2012; Accepted: 21 December 2012; Published: 24 January 2013

Glob Health Action 2013. © 2013 Do Thi Thanh Toan et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 18632 - http://dx.doi.org/10.3402/gha.v6i0.18632

Summary of Policy Recommendations

For Hanoi Preventive Medicine Center:

  • Hotspots analysis for DF should be widely used in DF surveillance since it can help reallocate the resource to deal with the outbreak more effectively.

For Preventive Medicine System:

  • Case-based surveillance system should use GPS data to track the disease outbreak and the effect of intervention.

Dengue fever (DF), a mosquito transmitted viral infection, is a serious public health concern worldwide, particularly in developing countries. Annually, the number of DF cases has been estimated to range from 50 to 100 million cases worldwide; of which, up to 500,000 cases result in dengue haemorrhagic fever (1, 2). The infection remains a major threat to the community well-being because it is associated with an increased risk of premature mortality and incurs significant health care cost to society (3). In Vietnam, a number of recent outbreaks of DF have also generated significant concerns among international health authorities (4, 5).

As the risk of DF varies with space and time, it is important to have precise knowledge of the regions at risk, the level of risk, risk factors, and the population exposed. In recent years, the development of geographic information systems (GIS) has provided a supportive spatial analytical tool that has enabled epidemiologists to include more simply a spatial component in epidemiologic studies (6). In the field of infectious and vector-borne diseases such as malaria or DF, GIS have been widely used for disease mapping of different pathologies, in analysis of space and space–time distribution of disease data, in identifying risk factors, and in mapping risk areas (5, 7). Moreover, to test whether any clusters can be detected or if the point process is purely randomly distributed, temporal, spatial, and space–time scan statistics (SaTScan) have recently come into common use (8, 9). The advantages of using SaTScan are that it can adjust for confounding variables, it can reduce pre-selection bias as it searches for clusters without specifying their size or location, it gives a single p-value as the likelihood-ratio-based test takes account of multiple testing, and finally, it can be applied to a whole region to detect significant clusters in that region (10).

With increasing concern about the threat posed by DF in Hanoi, a clearer picture of the epidemiology and important risk factors was needed. This study aimed to fill that need. Using spatial scan statistics and GIS, we investigated the spatial distribution of confirmed cases of DF and investigated the areas of high risk within all 14 districts of Hanoi. In addition, we have used GIS and spatial scan statistics to detect the hot spots and identify the disease dynamics dispersion of DF over period between 2004 and 2009 in Hanoi.

Methods

Study site

Hanoi is located in the north of Vietnam, in the low lying and densely populated Red River delta. Hanoi, before its merging with part of neighbouring provinces in 2008 (known as ‘old Hanoi’), had 14 districts divided into 229 postcode areas. In 2009, the old city had a population of 3.5 million; the population density was quite high, at 1,943 people/km2 (11). In this study, the geographic area of ‘old Hanoi’ was selected as a study site because it made it possible to use consistent data for all of the study time periods (2004–2009) (Figure 1).

Fig 1

Fig. 1.   Location of the study area – Hanoi, Vietnam.

There is a rapid population growth in Hanoi due to the influx of workers from rural areas. Previous studies in Vietnam showed that people living in poorer areas and old tenement housing, including houses for transient workers, generally tend to have unhygienic conditions with a lack of water supply and absence of window screens, both of which may promote Aedes mosquito proliferation and contact. The rapid movement of population between communities may have promoted the largest outbreak of dengue in Hanoi: in 2009, the city recorded almost 8,000 DF patients, representing a 15-fold increase from the previous year.

Hanoi experiences the typical climate of northern Vietnam, where summers are hot and humid, and winters are relatively cool and dry. The summer months from May to September receive the majority of rainfall in the year (1,682 mm rainfall/ year). The winter months from November to March are relatively dry, although spring then often brings light rains. The minimum winter temperature in Hanoi can dip as low as 6–7°C (43–45°F) not including the wind chill, while summer can get as hot as 38–40°C (100–104°F). The period from May to September is suitable for the development of the mosquito vector: high temperatures and high precipitation favour increased rates of mosquito development and a decreased length of reproductive cycle, as well as providing more sites for egg deposition and larva development (12).

With the above characteristics, Hanoi provides very favourable conditions for the existence, circulation and development of infectious diseases such as DF and dengue haemorrhagic fever.

Data sources

Daily data on DF cases between January 1998 and December 2009 in old Hanoi were obtained from the Hanoi Center for Preventive Health. Data included the onset date and place of onset of the notified cases of DF infection, age, sex, occupation of the patients and laboratory test results. The criteria for notification of DF disease are based on guidelines from the Ministry of Health, 1999, on surveillance, diagnosis, and treatment of dengue, in which individuals suspected to have dengue are those who have acute febrile illness (≥38°C) of 2–7 days duration with two or more of the following non-specific manifestations of DF: headache, retro-orbital pain, myalgia, arthralgia, rash, haemorrhagic manifestations, and leucopenia (11). The total number of cases was 25,483 after excluded cases where the residential address was unspecified (the years 1999–2002 have no addresses). The locations of patient's residence were taken from a Garmin GPS60 (Garmin Corporation, Taipei County, Taiwan) global positioning system while doing a field survey. A total of 51.4% (13,092 cases) of the DF records are available as case event data.

The population data of each postcode area for 2004 and 2009 were obtained directly from the decennial national Population and Housing Census, conducted by the General Statistic Office of Vietnam (GSO; http://www.gso.gov.vn).

Data analysis

Descriptive analysis

Descriptive statistics of numbers of Hanoi areas with notified DF cases in three periods (2004–2005, 2006–2007, and 2008–2009), incidence rate of DF in these periods, or choropleth maps were analysed to describe the dynamics of the disease. The monthly distribution of DF according to seasonality was also performed using boxplots.

Spatial autocorrelation analysis

The spatial autocorrelation of the expected incidence rates of DF in three different periods was assessed using Moran's I statistic in the program ArcGIS 9.2. Spatial autocorrelation was considered significant if the p<0.05. Moran's I ranges from −1 to 1 and can be interpreted as follows: a value close to 0 indicates spatial randomness, while a positive value indicates positive spatial autocorrelation and a negative value indicates negative spatial autocorrelation.

Cluster analysis

Scan statistics were used to detect and evaluate the clusters of cases in either a purely temporal, purely spatial, or space–time setting. This was achieved by gradually scanning a window across time and/or space, noting the number of observed and expected observations inside the window at each location. For each location and size of the scanning window, the alternative hypothesis was that there was an elevated risk within the window compared to outside.

In SaTScan software, the scanning window was an interval (in time), a circle or an ellipse (in space), or a cylinder with a circular or elliptic base (in space–time). Multiple different window sizes were used. The window with the maximum likelihood was the most likely cluster, that is, the cluster least likely to be due to chance. A p-value was assigned to this cluster. The standard purely spatial scan statistic imposed a circular window on the map. The window was in turn centred on each of several possible grid points positioned throughout the study region. For each grid point, the radius of the window varied continuously in size from zero to some upper limit specified by the user. In this way, the circular window was flexible both in its location and size, while each circle was a candidate cluster. The space–time scan statistic was defined by a cylindrical window with a circular (or elliptic) geographic base and with its height corresponding to the time. The base was defined exactly as for the purely spatial scan statistic, while the height reflected the time period of potential clusters. The cylindrical window was then moved in space and time, so that for each possible geographical location and size, it also visited each possible time period. In effect, an infinite number of overlapping cylinders of different sizes and shapes were obtained, which jointly cover the entire studied region, where each cylinder reflected a possible cluster. The temporal scan statistic used a window that moved in one dimension, time, defined in the same way as the height of the cylinder used by the space–time scan statistic. This meant that it was flexible in both start and end date. For purely spatial and space–time analyses, SaTScan also identified secondary clusters in the data set in addition to the most likely cluster and lined them up by their likelihood ratio test statistic. For purely temporal analyses, only the most likely cluster was reported. No geographic overlap was used as a default setting, so secondary clusters would not overlap the most significant cluster. In order to scan from small to large clusters, the maximum cluster size was set to 50% of the total population at risk. To ensure sufficient statistical power, the number of Monte Carlo replications was set to 999.

Dynamic dispersion of DF

In this study, we try to identify whether changes in DF varied with latitude and longitude of villages centroids in the three periods. Logistic regression models can be constructed with the dichotomous outcome variable defined as whether or not an increase of DF occurred in each village between the three periods. Longitude and latitude of village centroids were entered as explanatory variables. Spatial dispersions can be expressed in terms of odds ratios (OR) for longitude and latitude, with 95% confidence interval (CI).

Results

Descriptive analysis

Monthly average numbers of postcode areas with notified dengue cases in Hanoi for three periods 2004–2005, 2006–2007, and 2008–2009 are summarised in Table 1. It was shown that there is a clear trend of geographic expansion of dengue transmission in Hanoi through periods (with mean equal to 41.0, 49.33, and 79.33, respectively).


Table 1.  Descriptive statistics of monthly numbers of postcode areas with notified dengue cases
Period Mean SD Minimum Q1a Median Q3b Maximum
2004–2005 41.00 32.36 5 15.75 36 59 95
2006–2007 49.33 32.47 8 24.25 47 72.25 104
2008–2009 79.33 45.53 17 36.75 74 122.5 146
aQ1, first quartile value.
bQ3, third quartile value.

Figure 2 presents a striking variation in the monthly numbers of dengue cases and monthly numbers of postcode areas with dengue from 2004 to 2009. A large peak of dengue incidence occurred in October and November 2009 (3,696 cases and 2,698 cases, respectively). Peaks in incident cases coincided with high monthly numbers of postcode areas with DF cases.

Fig 2

Fig. 2.   Numbers of dengue cases (____) and postcode areas with dengue notifications (-----) between January 2004 and December 2009 in Hanoi.

Boxplots of the monthly numbers of postcode areas with dengue are shown in Fig. 3. The results indicated a strongly seasonal pattern (with a peak in autumn and early winter) and suggested that there was an upward trend of dengue incidence from 2004 to 2009.

Fig 3

Fig. 3.   Boxplots of the seasonal distribution of numbers of postcode areas with dengue infection in three periods in Hanoi. The boxplots display the values of the 25th, 50th, and 75th percentiles.

Figure 4 shows the geographic distribution of the raw incidence of notified dengue cases in Hanoi in three time periods. There was an expansion of postcode areas with dengue to the west-northern wards in Hanoi between 2004 and 2009. Dengue incidence ranged from 4.55 to 2887.6/100 000 and kept increasing, from 97 to 132 and subsequently 160 postcode areas in the three periods.

Fig 4

Fig. 4.   Map showing raw dengue incidence rates in three periods.

Spatial autocorrelation of DF

A significant positive spatial autocorrelation of dengue incidence for all three periods is presented in Table 2, where Moran's I index was 0.19 (expected Moran's I=−0.011, p<0.001) during 2004–2005 (expected Moran's I=−0.007, p=0.001), 0.32 during 2006–2007, and 0.22 (expected Moran's I=−0.006, p=0.001) during 2008–2009. This means that villages closer together tend to have more similar baseline incidence rates than those further apart.


Table 2.  Spatial autocorrelation analysis for dengue in Hanoi, 2004–2009
Period Annual incidence Moran's I E(I) p
2004–2005 226.32 0.19 −0.011 0.001
2006–2007 1049.78 0.32 −0.007 0.001
2008–2009 3774.49 0.22 −0.006 0.001

Purely temporal clustering

The results of the purely temporal clustering analysis in each year also indicate the seasonal tendency of dengue transmission. Table 3 shows the temporal clusters of DF cases in the study area from 2004 to 2009. There were significantly temporal clusters in all years (range of RR from 1.34–14.26), p=0.001). The highest RR was found in 2005 (RR: 14.26, p=0.001), which suggested there were strongest temporal clusters in 2005. The temporal clusters of DF in Hanoi often covered 6 months (June to December).


Table 3.  The clusters of dengue cases detected using the purely temporal analysis
Year Cluster time frame Total days of cluster Obsa Expb Relative risk LLRc p
2004 4/8–1/12 118 216 160.30 1.90 17.68 0.001
2005 5/6–1/12 177 246 167.56 14.26 74.0 0.001
2006 5/6–1/12 177 1,679 1,556.38 2.30 45.39 0.001
2007 5/7–31/12 177 1,051 937.56 1.88 33.36 0.001
2008 4/8–16/12 133 992 886.41 1.60 23.84 0.001
2009 20/6–16/12 179 7,725 7,634.84 1.34 12.64 0.001
aThe number of observed cases in a cluster.
bThe number of expected cases in a cluster.
cLog likelihood ratio.

Purely spatial clustering

Analysis of purely spatial clustering of dengue cases from 2004 to 2009, with the maximum spatial cluster size of 50% of the total population, identified the most likely cluster for each of the 6 years. However, only in 2004 and 2005 did the result show a random distribution of dengue in space of Dong Da, Hai Ba Trung, Thanh Xuan and Hoang Mai districts (Table 4). For the years 2006–2009, it was found that the risk is the same inside and outside the cluster since p>0.05.


Table 4.  The clusters of dengue cases detected using the purely spatial analysis
Year   Location Obsa Expb Relative risk LLRc p
2004 A Thanh Tri 175 150.08 1.33 3.56 0.932
2005 A Hai Ba Trung, Hoang Mai 65 37.10 2.01 10.4 0.002
  B Thanh Xuan, Dong Da 128 96.19 1.66 8.18 0.017
2006 A Dong Da, Tay Ho, Long Bien 833 806.05 1.06 0.81 1.000
2007 A Dong Da, Ba Dinh 658 609.28 1.17 3.89 1.000
2008 A Tu Liem, Cau Giay 238 215.30 1.13 1.41 1.000
2009 A Hoan Kiem, Hai Ba Trung 2,346 2323.1 1.02 0.16 1.000
aThe number of observed cases in a cluster.
bThe number of expected cases in a cluster.
cLog likelihood ratio.
A: Most likely cluster; B: Secondary cluster.

Space–time clustering

The space–time clustering analysis of the dengue data from 2004 to 2009 was also tested. Figure 4 and Table 5 illustrate the clusters in all districts of Hanoi at a 5% significant level (p<0.05) in this period.


Table 5.  SaTScan statistics for space–time clusters with significantly higher incidence in Hanoi from 2004 to 2009 (most likely cluster)
Location Radius (km) Time frame Relative risk p
Hoang Mai 2.50 4/8–19/7/2004 2.03 0.001
Thanh Xuan 0.68 17/11–31/12/2005 2.66 0.001
Dong Da 0.64 2–16/11/2006 3.51 0.001
Dong Da 0.99 17/11–16/12/2007 3.38 0.001
Hoang Mai 0.29 19/8–2/10/2008 5.58 0.001
Hoan Kiem, Hai Ba Trung 1.56 2–16/11/2009 6.41 0.001

The results reveal a high significance of space–time association with DF transmission. It is revealed that six out of the 14 districts of Hanoi had significant cluster patterns, in which Dong Da, Hoang Mai, and Thanh Xuan have the highest number of space–time clusters. The most likely cluster was found to differ during all three year periods. In 2004–2005, only a few clusters were found, distributed over a large distance and time. In 2006–2007, the most likely cluster occurred in Dong Da, with 149 cases and within 14 days in November. In August 2008, the most likely cluster was reported in Hoang Mai and was limited to 250 m (RR=5.58, p=0.001). In November 2009, the highest number of dengue cases (553) was again found in Dong Da and Hoan Kiem, within the radius of 1,560 m. The RR within the most likely cluster was 6.41 (p=0.001). The secondary clusters reported in Hoang Mai, Tay Ho, and Hai Ba Trung were also limited at 1,000 m and within 29 days (Fig. 5).

Fig 5

Fig. 5.   Scatterplot of significant space–time clusters in Hanoi from 2004 to 2009.

Dynamic dispersion of DF

Logistic regression models were constructed to identify whether changes in DF varied with latitude and longitude of postcode centroids in the three periods (Table 6). The results suggest that changes in DF were significantly associated with latitude (OR 1.17, 95% CI 1.02–1.34) between the periods 2008–2009 and 2004–2005. However, there was no association between DF and longitude in any period.


Table 6.  Changes of dengue fever in latitude and longitude, Hanoi, 2004–2009
  Latitude Longitude
Change in periods OR 95% CI OR 95% CI
Period 3–Period 1 1.17 1.02–1.34 1.06 0.93–1.34
Period 2–Period 1 0.94 0.85–1.03 0.98 0.87–1.11
Period 3–Period 2 1.09 0.99–1.22 0.98 0.89–1.06
Note: Period 1: 2004–2005; Period 2: 2006–2007; Period 3: 2008–2009.

Discussion

The results of this study revealed significant spatio-temporal variation in the distribution of DF in Hanoi, Vietnam. Previous studies showed that, in Vietnam, the dengue epidemic often had a cycle of 3–5 years and was expected to reach a 10-year peak (5, 13, 14). These peaks were in 1987, 1998 and 2009, as also recorded in this study, showing an upward trend of dengue cases with the largest outbreaks in 2009. Most of the dengue cases occurred between June and November, when the rainfall and temperature are highest of the year. This time period appeared in our analysis of purely temporal clustering, showing high-risk months for dengue each year between 2004 and 2009. The results are consistent with those of a study conducted in a Central Highlands province of Vietnam, which found that dengue was most prevalent in the wet season (11). As other tropical countries, Vietnam's climate is favourable for the transmission of DF. A warm temperature is crucial to the mosquito's life and gonotrophic cycle, and to virus replication. In addition, stagnant water and higher humidity could augment the epidemic during a rainy season (9, 12, 1519).

Results from space–time clustering identified the high-risk areas over the larger region and over the years. Using the maximum spatial cluster size of 50% of the total population, and the maximum temporal cluster size of 50% of the total population, we identified six among 14 districts of Hanoi as having significant cluster patterns within a period of 29 days and limited at 1,000 m on average. The areas recording the highest numbers of space–time clusters were Dong Da, Hoang Mai, and Thanh Xuan, with some expansion to the north-western wards of Hanoi, where a higher population is concentrated. Similar results were reported in Malaysia, where geographical weighted regression analysis revealed that the spatial distribution of DF was closely related to population distribution (9). Moreover, the result of their space–time permutation scan statistics showed that most of the clusters were in medium or high population areas. Spatial clustering of disease is almost inevitable, since human populations generally live in spatial clusters rather than random distributions.

Finally, spatial autocorrelation and logistic regression analysis are valuable tools for studying how spatial patterns change over time (2026). In this study, we found that DF had high spatial autocorrelation in three different time periods. The patterns were closely related to the topography of the environment, in that the villages closer together tended to have more similar baseline incidence rates.

To the best of our knowledge, this is the first study to apply a spatial scan technique, using SaTScan software, to investigate the temporal and geographical clustering of DF disease in Vietnam, in this case in Hanoi. The study provides useful information on the prevailing epidemiological situation of DF in Hanoi city. This new knowledge about the presence of hotspots of DF in the city can help Hanoi Preventive Medicine Center to intensify their remedial measures in the identified areas of high DF prevalence and chalk out future strategies for more effective DF control.

The study does have some potential limitations. The disease case data were not survey-based but used sentinel surveillance data, which only records patients presenting at hospitals. The study analysed statistically significant clusters of DF in Hanoi but did not examine their causes. Future research should focus on the effect of various socio-economic and environmental factors that could affect disease transmission.

Conclusion and policy implications

The study has shown the presence of long-term hotspots of DF occurrence, which was highest in Dong Da, Hoang Mai, and Thanh Xuan districts of Hanoi. We have also shown the expansion of geographic distribution of DF over recent years. The results demonstrate the necessity to further improve our understanding of the impact of socio-environmental change and ecosystem stress on the transmission of DF. The study has illustrated how, using existing health data, spatial scan statistic and GIS can provide public health officials with necessary information about the prevalence of statistically significant hotspots of DF in the city, thus enabling them to chalk out more effective strategies to contain this scourge. Moreover, hot spot analysis using GIS should be widely used in DF surveillance since it can help reallocate the resource to deal with the outbreak more effectively. This effort will contribute to dengue control strategy.

Acknowledgements

We thank Dr. Nguyen Nhat Cam, Vice-Director of Hanoi Preventive Medicine Center and his staff for assistance in collecting surveillance data.

Conflict of interest and funding

This study was funded by the Netherlands Higher Education (NPT) project on ‘Strengthening teaching and research capacity in preventive medicine in Vietnam’.

References

  1. Hopp MJ, Foley JA. Global-scale relationships between climate and the dengue fever vector, Aedes aegypti. Clim Chang 2001; 48: 441–63. Publisher Full Text
  2. Honorio NA, Nogueira RMR, Codeco CT, Carvalho MS, Cruz OG, Magalhães MAFM, et al. Spatial evaluation and modeling of dengue seroprevalence and vector density in Rio de Janeiro, Brazil. PLOS Negl Trop Dis 2009; 3: e545. PubMed Abstract | PubMed Central Full Text
  3. Shepard DS, Suaya JA, Halstead SB, Nathan MB, Gubler DJ, Mahoney RT, et al. Cost-effectiveness of a pediatric dengue vaccine. Vaccine 2004; 22: 1275–80. PubMed Abstract | Publisher Full Text
  4. Thai KTD, Cazelles B, Nguyen NV, Vo LT, Boni MF, Farrar J, et al. Dengue dynamics in Binh Thuan province, Southern Vietnam: periodicity, synchronicity and climate variability. Negl Trop Dis J 2010; 4: e747. Publisher Full Text
  5. Nguyen LAP, Clements ACA, Jeffery JAL, Yen NT, Nam VS, Vaughan G, et al. Abundance and prevalence of Aedes aegypti immatures and relationships with household water storage in rural areas in southern Vietnam. Int Health 2011; 3: 115–25. Publisher Full Text
  6. Lawson AB, Williams FLR. An introductory guide to disease mapping. Chichester, UK: John Wiley & Son; 2001.
  7. Hu W, Clements A, Williams G, Tong S. Spatial analysis of notified dengue fever infections. Epidemiol Infect 2011; 139: 391–9. PubMed Abstract | Publisher Full Text
  8. Jeefoo P, Tripathi NK, Souris M. Spatio-temporal diffusion pattern and hotspot detection of dengue in Chachoengsao province, Thailand. Int J Environ Res Public Health 2011; 8: 51–74. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  9. Seng SB, Chong AK, Moore A. Geostatistical modelling, analysis and mapping of epidemiology of dengue fever in Johor state, Malaysia; 2005. Available from: http://www.business.otago.ac.nz/sirc/conferences/2005/13_su.pdf [cited 15 January 2012].
  10. Kulldorff M. (2006) SaTScan user guide (http://www.satscan.org) v7.0, 19–20.
  11. Ministry of Health. Guidelines on the surveillance, diagnosis, and treatment of dengue hemorrhagic fever. Medical Publishing House. Hanoi, Vietnam; 1999.
  12. Martens P, Huynen MMTE. Will global climate change reduce thermal stress in the Netherlands? Epidemiology 2001; 12: 753–4. 11: 172. PubMed Abstract | Publisher Full Text
  13. Pham HV, Doan HTM, Phan TTT, Minh NNT. Ecological factors associated with dengue fever in a central highlands province, Vietnam. BMC Infect Dis 2011; 11: 172.
  14. Cuong HQ, Hien NT, Duong TN, Phong TV, Cam NN, Farrar J, et al. Quantifying the emergence of dengue in Hanoi, Vietnam: 1998–2009. PloS Negl Trop Dis 2011; 5: e132. Publisher Full Text
  15. Martens P. Health transitions in a globalising world: towards more disease or sustained health? Futures 2002; 37: 635–48. Publisher Full Text
  16. Martens P, Huynen MMTE. A future without health? Health dimension in global scenario studies. Bull World Health Organ 2003; 81: 896–901. PubMed Abstract | PubMed Central Full Text
  17. Martens P. Integrated assessment models. In: Valkering P, Amelung B, Vander Brugge R, Rotmans R, eds. More puzzlesolving for policy: integrated assessment from theory to practice. Maastricht University. Maastricht: ICIS; 2006.
  18. Cazelles B, Chavez M, McMichael AJ, Hales S. Nonstationary influence of El Nino on the synchronous dengue epidemics in Thailand. PloS Med 2005; 2: e106. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  19. Ali M, Wagatsuma Y, Emch M, Breiman RF. Use of geographic information system for defining spatial risk for dengue transmission in Bangladesh: role for aedes albopictus in an urban outbreak. Am J Trop Med Hyg 2003; 69: 634–40. PubMed Abstract
  20. Almeida AS, Medronho RA, Valencia LIO. Spatial analysis of dengue and the socio economic context of the city of Rio de Janeiro (Southeastern Brazil). Rev Saude Publica 2009; 43.
  21. Ooi E-E, Gubler DJ. Dengue in Southeast Asia: epidemiological characteristics and strategic challenges in disease prevention. Cad Saude Publica 2008; 25(Suppl 1): S115–S24.
  22. Nishiura H. Mathematical and statistical analyses of the spread of dengue. Dengue Bull 2006; 30.
  23. Lu L, Lin H, Tian L, Yang W, Sun J, Liu Q, et al. Time series analysis of dengue fever and weather in Guangzhou, China. BMC Public Health 2009; 9: 395. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  24. Barbazan P, Guiserix M, Boonyuan W, Tuntaprasart W, Pontier D, Gonzalez J-P, et al. Modelling the effect of temperature on transmission of dengue. Med Vet Entomol 2010; 24: 66–73. PubMed Abstract | Publisher Full Text
  25. Wu PC, Guo HR, Lung SC, Lin CY, Su HJ. Weather as an effective predictor for occurrence of dengue fever in Taiwan. Acta Trop 2007; 103: 50–7. PubMed Abstract | Publisher Full Text
  26. Hsieh YH, Chen CWS. Turning points, reproduction number, and impact of climatological events for multi wave dengue outbreaks. Trop Med Int Health 2009; 14: 628–38. PubMed Abstract | Publisher Full Text

*Do Thi Thanh Toan
Biostatistics and Medical Informatics Department
Institute of Training for Preventive Medicine and
Public Health
Hanoi Medical University
Vietnam
Tel: +84 983 984 486
Email: toan.dothanh@maastrichtuniversity.nl

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Assessing the household financial burden associated with the chronic non-communicable diseases in a rural district of Vietnam

Hoang Van Minh* and Bach Xuan Tran

Department of Health Economics, Center for Health System Research, Institute for Preventive Medicine and Public Health, Hanoi Medical University, Hanoi, Vietnam

Abstract

Background: While there is accumulated evidence showing the rapid rise of the burden caused by non-communicable diseases (NCDs) in Vietnam, information on the extent to which households in the country suffer financial catastrophe or impoverishment caused by the diseases is still largely lacking. This paper aims to examine the self-reported prevalence of major chronic diseases among a population in rural Vietnam and to analyse the household financial burden associated with these diseases.

Methods: A cross-sectional survey of 800 randomly selected households was carried out in Vo Nhai District, Thai Nguyen Province, in 2010. Face-to-face interviews were conducted with key informants of selected households on diagnosed chronic NCDs, health care utilization and health expenditure of all household members. The World Health Organization's definitions of catastrophic expenditure and impoverishment were used. Both descriptive and analytical statistics were applied.

Results: The prevalence of chronic NCDs in households and individuals was 29.3 and 33.4%, respectively. The catastrophic health expenditure and impoverishment rates among the households who have at least one member with a chronic disease were 14.6 and 7.6%, respectively. These rates were significantly higher than the corresponding figures among the households whose members were free from the diseases (4.2 and 2.3%, respectively). The odds of experiencing catastrophic health expenditure and impoverishment among the household with NCD patients were 3.2 and 2.3 times greater than that of other households.

Conclusion: Findings from this study indicate that the epidemiological and household financial burdens caused by chronic diseases in Vietnam are now substantial and need immediate mitigation measures.

Keywords: chronic non-communicable diseases; out-of-pocket expenditure; financial burden; rural; Vietnam

Received: 1 June 2012; Revised: 15 September 2012; Accepted: 28 November 2012; Published: 20 December 2012

Glob Health Action 2012. © 2012 Hoang Van Minh and Bach Xuan Tran. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2012, 5: 18892 - http://dx.doi.org/10.3402/gha.v5i0.18892

The burden of chronic non-communicable diseases (NCDs) is increasing worldwide, especially in developing countries. Of the 57 million deaths that occurred globally in 2008, 36 million – accounting for almost two thirds – were due to chronic NCDs (mainly cardiovascular diseases, cancers, diabetes and chronic lung diseases) (1). Chronic NCDs not only cause premature death but also have major adverse effects on the quality of life of affected individuals and create large adverse economic effects on families, communities and societies (2, 3).

Vietnam is undergoing a rapid epidemiological transition that resulted in an increased number of chronic NCD cases. Chronic diseases have been shown to be major causes of morbidity and mortality in hospitals for the whole country. Hospital admissions due to chronic NCD diseases increased from 39% in 1986 to 66.2% in 2008 and chronic diseases deaths rose from 42% in 1986 to 63.3% in 2009 (4). Other community-based studies also proved that chronic NCDs have already been one of the leading causes of both morbidity (5) and mortality (6, 7) among Vietnamese populations.

Policy recommendations

The findings from this study indicate that the household financial burdens caused by chronic non-communicable diseases in Vietnam are substantial. The following points summarise the key policy recommendations:

  • More attention should be paid on prevention and control of chronic non-communicable diseases in Vietnam.
  • Interventions should be comprehensive and integrated, including both primary and secondary approaches, as well as policy-level involvements.
  • As health insurance was shown to have some impact on financial protection, expanding coverage of health insurance in all three dimensions (breadth, height and depth) should be a priority policy action.
  • Reforming provider payment methods could also be a good option to enhance the financial protection against chronic diseases in Vietnam.

The health system in Vietnam is a mixed public–private provider system, in which the public system still plays a key role in health care, especially in prevention, research and training. The public health care system in Vietnam is now organized into three levels (Ministry of Health, Provincial Health Services, District Health Facilities, and Commune Health Centres, including Village Health Workers). Vietnam is now using three main options to finance national health care expenditure, including the state budget, insurance contributions and direct out-of-pocket (OOP) payments by service users (8). Of the health financing sources, the state budget plays a critical role in protecting public health, accounting for approximately 25% of total health expenditure (9). National health insurance was introduced in Vietnam in 1992 and now has two national insurance schemes – a compulsory scheme and a voluntary scheme. The coverage of health insurance in Vietnam increased from 16% of the population in 2002 to 60% in 2010. However, health insurance contributions only accounted for 18% of total health expenditures in 2009 (10). The Government of Vietnam is committed to achieving full population health insurance coverage by 2020 (11). Another relatively large financial flow is household's direct OOP payments, partially as a result of the hospital fee policy introduced in 1989 following the Decision No. 45/HDBT dated 24 April 1989 by the Government of Vietnam, which allowed hospitals to recover costs through user fees. Direct OOP payments for health care refer to the expenditures directly made by households when they use services, primarily for the purchase of drugs, payment of hospital user fees, diagnostic service fees and other indirect expenses related to seeking medical care at state or private facilities (including self-medication) (8). In Vietnam, the OOP payments as a share of the total health expenditure have been always high, ranging from 50% to 70% (8).

While there is accumulated evidence showing the rapid rise of the burden caused by NCDs in Vietnam, information on the extent to which households in the country suffer financial catastrophe or impoverishment caused by the diseases is still largely lacking. This paper aims to examine the self-reported prevalence of major chronic diseases among a population in rural Vietnam and to analyse the household financial burden associated with the diseases.

Methods

Study design

This was a population-based cross-sectional study.

Study setting

The study was carried out in Vo Nhai District, Thai Nguyen province. This is a mountainous district located 90 km north from Hanoi capital. The district has approximately 63,000 people living in 14 communes and one township. According to the local health statistics in 2010, the annual income per capita of the local population was approximately US$ 500. The proportion of poor households, as classified by local authorities (based on national poverty line using per capita income, area of land and assets household possessed), was approximately 10%. In terms of health status, chronic diseases were the leading causes of death in the district. Up to now, there has been no intervention to prevent and control chronic diseases in this setting.

Sample size and sampling

A total of 800 households were randomly selected from the list of households in Vo Nhai provided by the district health centre. The same size was determined using the level of significance (α) of 0.05, relative precision (ɛ) of 0.3, the expected proportion of households incurring catastrophic health care payment (p) of 5.5% (finding from a study in Vietnam in 2011 (12)), and the anticipated non-response of 10%.

Data collection

Face-to-face interviews were conducted with key informant of the selected households (usually head of the household). The interviews were carried out by trained surveyors from Hanoi Medical University using a structured questionnaire. The questionnaire was developed by the research team containing questions on diagnosed chronic NCDs (cardiovascular diseases, cancers, diabetes and chronic lung diseases) among household members, health care utilization and expenditure (during the last 12 months for inpatient care, during the last 4 weeks for outpatient cares and self-treatments). Information on the socio-economic condition of the household was also collected. Spot-checks and re-checks on sample data were conducted by the research team for quality control.

Definitions

Respondents were asked if they had been told by a health worker that they had any of the following chronic conditions: cardiovascular, cancer, diabetes, chronic pulmonary diseases. The measures of catastrophic expenditure and impoverishment have been clearly described elsewhere (13, 14):

  1. OOP health payments refer to the payments made by households at the point they received health services. OOP payments are the net of insurance reimbursement.
  2. Household's consumption expenditure comprises both monetary and in-kind payment on all goods and services, and the money value of the consumption of homemade products.
  3. Household's capacity to pay is defined as effective income remaining after basic subsistence needs have been met. Effective income is taken to be the total consumption expenditure of the household. Some households may report food expenditure that is lower than subsistence spending.
  4. Household subsistence spending is the minimum requirement to maintain basic life in a society. A poverty line is used in the analysis as subsistence spending. Poverty line is defined as the food expenditure of the household whose food expenditure share of total household expenditure is at the 50th percentile in the country. In order to minimize measurement error, we use the average food expenditures of households whose food expenditure share of total household expenditure is within the 45th and 55th percentile of the total sample. Considering the economy scale of household consumption, the household equivalence scale is used rather than actual household size. The value of the parameter β has been estimated from previous studies based on 59 countries’ household survey data, and it equals 0.56.
  5. Catastrophic expenditure occurs when a household's total OOP health payments equal or exceed 40% of household's capacity to pay.
  6. Impoverishment: A non-poor household is impoverished by health payments when it becomes poor (when household expenditure is equal to or higher than subsistence spending but is lower than subsistence spending net of OOP health payments) after paying for health services

Data analysis

Data were analysed using Stata statistical software version 10. Both descriptive and analytical statistics were applied. Logistic regressions were used to identify the socio-economic correlates of the catastrophic and poverty impacts of household's out-of-pocket health expenditure. The dependent variables included catastrophic health expenditure and impoverishment. The independent variables are socio-economic indicators such as sex of household head, household size, number of old people in the household, number of children under 6 years of age in the household, number of household members with insurance card and economic status classified by local Authorities (based on national poverty line using per capita income, area of land and assets household possessed). A significance level of p<0.05 was used.

Ethical considerations

All human subjects in the study were asked for their informed consent before collecting data, and all had complete right to withdraw from the study at any time without having any threat.

Results

Of the 800 selected households, 768 households, consisting of 3,378 people (46.8% male and 53.2% female), responded to the survey (response rate of 96%). On average, each studied household had 4.4 people. The proportion of poor households, as classified by local Authorities, was 9.9%. The annual per capita income of the study population was VND 9.7 million (around US$ 500).

Overall, 29.3% of the studied households had at least one member who had one or more chronic NCDs of interest. At individual level, 33.4% of respondents reported having at least one chronic NCDs (35.5% among men and 31.6% among women). The prevalence of self-reported chronic pulmonary disease, cardiovascular disease (including hypertension and stroke), diabetes and cancers among the studied participants were 24.9, 16.3, 0.2 and 0.3%, respectively (Table 1).


Table 1.  Prevalence of self-reported chronic diseases among the study population
  Male (%) Female (%) Overall (%)
Any chronic disease 562 (35.5) 567 (31.6) 1129 (33.4)
Chronic pulmonary disease 349 (22.1) 496 (27.6) 841 (24.9)
Cardiovascular diseases 264 (16.7) 286 (15.9) 551 (16.3)
Diabetes 3 (0.2) 4 (0.2) 7 (0.2)
Cancer 2 (0.1) 9 (0.5) 10 (0.3)

As shown in Table 2, during the last 12 months, the rate of utilization of inpatient care services among households with and without NCD patients was 18.9 and 6.4%, respectively. The corresponding mean out-of-pocket payment for the inpatient care services was VND 5,760,000 and VND 2,472,000, respectively. During the last 4 weeks, the rate of utilization of outpatient care services among households with and without NCD patients was 52.1 and 32.3%, respectively. The corresponding mean out-of-pocket payment for the outpatient care services were VND 78,300 and VND 32,600, respectively. The differences in health service utilization rates and health payments between households with and without chronic NCD patients was statistically significant.


Table 2.  Pattern of health care utilization and out-of-pocket payment for health care
  Households where there was at least one member with a chronic disease Households without any person with chronic disease p
Inpatient care services during the last 12 months      
 Utilization: n (%) 43 (18.9) 35 (6.4) 0.00a
 Mean out-of-pocket payment: VND 5,760,000 2,472,000 0.00b
Outpatient care services during the last 4 weeks      
 Utilization: n (%) 117 (52.1) 175 (32.3) 0.00a
 Mean out-of-pocket payment: VND 78,300 32,600 0.00b
Note: aChi-squared test, bMann–Whitney test.

Table 3 presents the patterns of catastrophic health expenditure and impoverishment among the studied households. The proportion of catastrophic health expenditure among the households where there was at least one member with a chronic disease and among those without any member with a disease were 14.6 and 4.2%, respectively. The impoverishment rates among the households where there was at least one member with a chronic disease and among those without any member with a disease were 7.6 and 2.3%, respectively. The differences in both the catastrophic health expenditure and the impoverishment rates between the households of chronic disease patients and other households were statically significant.


Table 3.  Pattern of catastrophic health expenditure and impoverishment
  Households where there was at least one member with a chronic disease Households without any person with chronic disease p Chi-squared test
Catastrophic health expenditure: n (%) 33 (14.6) 23 (4.2) 0.00
Impoverishment: n (%) 17 (7.6) 12 (2.3) 0.00

Table 4 presents the results of logistic regression analysis of the correlates of catastrophic health expenditure. The significant correlates of the catastrophic health expenditure were: (1) Chronic disease: Households where there was at least one member with a chronic illness had a significantly higher rate of catastrophic expenditure (OR=3.2); (2) Household size: Having more people was significantly associated with a lower rate of catastrophic expenditure (OR=0.87); (3) Number of elderly people (aged 60 years and over) in the household: Having more elderly people was significantly associated with a higher proportion of catastrophic health expenditure (OR=1.2); (4) Economic status: Poor household were more likely to incur catastrophic expenditure for health (OR=2.6); and (5) Insurance status: Having at least one person with health insurance was associated with a slightly lower rate of catastrophic expenditure (OR=0.87).


Table 4.  Correlates of catastrophic health expenditure
  OR p
Having at least one member with a chronic illness    
  Yes 3.2 0.00
  No 1  
Sex of the household's head    
  Male 0.87 0.66
  Female 1  
Household size 0.79 0.00
Number of elderly people (aged 60 years and over) in the household    
  Yes 1.32 0.00
  No 1  
Number of children under 6 years in the household    
  Yes 1.1 0.07
  No 1  
Economic status    
  Poor 2.6 0.02
  Non-poor 1  
Having at least one person with health insurance card    
  Yes 0.87 0.04
  No 1  

Table 5 presents the results of logistic regression analysis of the correlates of impoverishment problem. The significant correlates of the impoverishment problem were: (1) Chronic disease: Households where there was at least one member with a chronic illness had a significantly higher rate of impoverishment (OR=2.3); 2) Household size: Having more people was significantly associated with a lower rate of impoverishment (OR=0.86); 3) Number of elderly people (aged 60 years and over) in the household: Having more elderly people was significantly associated with a higher proportion of impoverishment (OR=1.1); 4) Insurance status: Having at least one person with health insurance was associated with a slightly lower rate of impoverishment (OR=0.72).


Table 5.  Correlates of impoverishment
  OR p
Having at least one member with a chronic illness    
  Yes 2.3 0.00
  No 1  
Sex of the HH's head    
  Male 0.79 0.79
  Female 1  
Household size 0.86 0.00
Number of elderly people (aged 60 years and over) in the household    
  Yes 1.1 0.00
  No 1  
Number of children under 6 years in the household    
  Yes 1.3 0.12
  No 1  
Having at least one person with health insurance card    
  Yes 0.72 0.04
  No 1  

Discussion

A high prevalence of self-reported chronic NCDs in studied households and individuals (29.3 and 33.4%, respectively) observed in this study proves the fact that chronic NCDs have already affected a large population in this setting. A study in another rural area in the North of Vietnam in 2007 reported a higher individual prevalence of chronic disease (39%), because that also included chronic joint problems (5). Several other studies also showed that the burden of disease from chronic diseases in Vietnam was substantial (1518).

Of particular interest in this article is the household financial burden associated with chronic diseases. We have shown that when a household member has a chronic illness, he or she was forced to use more health services and, as a consequence, the household had to spend more of its OOP money on health care for that ill member. Many households in the study setting incurred a catastrophic level of health expenditure and/or were pushed into poverty because of health care payments. The catastrophic health expenditure and the impoverishment rates were higher among the households where there was at least one member with a chronic disease (14.6 and 7.6%, respectively) compared to the corresponding figures among the households who were free from the diseases (4.2 and 2.3%, respectively). The odds of catastrophic health expenditure and impoverishment among the household with NCD patients were 3.2 and 2.3 times greater than that for other households.

There have been very few studies on this area from Vietnam. A study by Hien et al. in northern Vietnam in 2004 found that 19% of rural dwellers with diabetes had to sell assets, using savings or borrowing from neighbours to pay for health care costs (19). A study by Thuan et al. in a rural district in Vietnam in 2006 revealed that household expenditure on the treatment of chronic disease illnesses were also considerable (20). Wagstaff et al. found that Vietnamese households have not been able to hold their food and non-food consumption constant in the face of income reductions and extra medical care spending because of chronic illness (21). A recent study by WHO, using the National Living Standard Survey data, reported that the catastrophic health expenditure and the impoverishment rates due to general health care expenditure of the whole country in 2008 were 5.5 and 3.9%, respectively (12).

The household financial burden from chronic diseases was also pronounced by international publications. A recent review by Saksena et al. documented findings from studies on the impact of OOP payments for treatment of NCDs from various developing countries such as Pakistan in 2004 (22), Burkinafaso in 2006 (23), Kenya in 2007 (24), China in 2009 (25), China in 2010 (26, 27), India in 2008 (28) and India in 2012 (29). The authors concluded that the households with chronic disease patients had to spend a substantial share of their income on care for these diseases, and many households experienced catastrophic health expenditure and impoverishment as a result of their health spending (30). A review by Engelgau et al. also confirmed that the financial risks from chronic diseases put on households in developing countries were significant (31).

In this study, we found that the financial burden on poor households was generally higher than the burden on richer households. This is similar to findings from previous studies in India in 2008 (28) and in China in 2009 (25). Engelgau et al. also demonstrated that the household financial burden from chronic diseases impacted more on the poor and vulnerable populations (31). We also found that households with more elderly people had a higher probability of encountering catastrophic payment. While Vietnam is undergoing demographic transition and experiencing rapid population ageing, there is still no specific policy on health for the elderly (32, 33). More attention on health care for elderly people in Vietnam is needed in the future.

Although financial protection is the most important aspect of health insurance coverage, we found little impact of health insurance on protecting people from catastrophic payment and impoverishment. Most of the studies on impacts of health insurance in Vietnam consistently found that insurance have only a modest effect on reducing OOP payments (3438). The modest impact of insurance on financial protection reflects the fact that insured patients did co-pay for a high share of costs for medical care, especially for chronic disease care services.

We need to note some limitations of this study. First, the cross-sectional nature of the data only allowed us to examine the short-term impacts of household direct OOP payments. Secondly, we were just able to observe the OOP payments for health, catastrophic health expenditure and impoverishment proportions among the households where there was at least one member with a chronic disease, and also figures among the households whose members were free from the diseases. We did not know whether or not the payments of the households with chronic disease patients were actually for the chronic disease care or for other health services.

Acknowledgements

This research was implemented as an activity at the Center for Health System Research, Hanoi Medical University, which is financially supported by the Rockefeller Foundation and The China Medical Board. We would like to thank the Sida project ‘Public health preparedness and response to critical global health issues in Vietnam and Sweden’ which is being implemented by Epidemiology and Global Health, Umeå University, Sweden & Hanoi Medical University, Vietnam for covering the publication fee of this paper.

Conflict of interest and funding

The authors have not received any funding or benefits from industry or elsewhere to conduct this study.

References

  1. WHO (2011). Global status report on noncommunicable diseases 2010. Geneva: World Health Organization.
  2. WHO (2005). Preventing chronic diseases – a vital investment. Geneva: World Health Organization.
  3. Strong K, Mathers C, Leeder S, Beaglehole R. Preventing chronic diseases: how many lives can we save? Lancet 2005; 366: 1578–82. [Crossref]
  4. Ministry of Health of Vietnam (2009). Health statistics year book. Hanoi: National Cancer Institute.
  5. Minh HV, Huong DL, Giang KB. Self-reported chronic diseases and associated sociodemographic status and lifestyle risk factors among rural Vietnamese adults. Scand J Public Health 2008; 36: 629–34. [Crossref]
  6. WHO (2002). Global Program on Evidence for Health Policy (EBD/GPE/EIP). Geneva: World Health Organization.
  7. Hoang VM, Dao LH, Wall S, Nguyen TK, Byass P. Cardiovascular disease mortality and its association with socioeconomic status: findings from a population-based cohort study in rural Vietnam, 1999–2003. Prev Chronic Dis 2006; 3: A89.
  8. Ministry of Health of Vietnam, Health Partnership group (2008). Join annual health review 2008: health financing in Vietnam. Hanoi: Minstry of Health of Vietnam.
  9. Ministry of Health of Vietnam (2010). National health accounts in Vietnam during 1998–2008 period. Hanoi: Ministry of Health of Vietnam.
  10. Ministry of Health of Vietnam (2011). National health accounts in Vietnam 2009. Hanoi: Ministry of Health of Vietnam.
  11. Ministry of Health of Vietnam, Health Partnership group (2012). Join annual health review 2011. Hanoi: Minstry of Health of Vietnam.
  12. WHO (2011). Financial burden of household out-of-pocket health expenditure in Vietnam: findings from the National Living Standard Survey 2002–2008. Hanoi: World Health Organization Office in Vietnam.
  13. Xu K, Evans DB, Kawabata K, Zeramdini R, Klavus J, Murray CJ. Household catastrophic health expenditure: a multicountry analysis. Lancet 2003; 362: 111–7. [Crossref]
  14. WHO (2005). Distribution of health payments and catastrophic expenditures methodology. Discussion Paper No. 2. Geneva: World Health Organization.
  15. Ministry of Health of Vietnam (2003). Vietnam National Health Survey 2001–2002. Hanoi: Ministry of Health Vietnam.
  16. Ministry of Health of Vietnam (2008). The pattern of cancer in Vietnam 2000. Hanoi: National Cancer Institute.
  17. Binh TV, Uoc HK, Cockram CS. The prevalence of diabetes and IGT in Vietnam. Results of a national survey conducted in 2002. 18th International Diabetes Federation Congress. Paris: Diabetes & Metabolism; 2002.
  18. Cockram CS, Van Binh T, Galea G. Diabetes prevention and control in Vietnam: a demonstration project in two provinces. Pract Diabetes Int 2006; 23: 361–4. [Crossref]
  19. Hien PT. Health care expenditure among diabetes patients. Hanoi: Institute for Health and Development; 2004.
  20. Thuan NTB, Lofgren C, Chuc NTK, Janlert U, Lindholm L. Household out-of-pocket payments for illness: evidence from Vietnam. BMC Public Health 2006; 6: 283. [Crossref]
  21. Wagstaff A. The economic consequences of health shocks: evidence from Vietnam. J Health Econ 2007; 26: 82–100. [Crossref]
  22. Mahmood N, Ali SM. The disease pattern and utilisation of health care services in Pakistan. Pak Dev Rev 2004; 42: 745–60.
  23. Su TT, Kouyate B, Flessa S. Catastrophic household expenditure for health care in a low-income society: a study from Nouna District, Burkina Faso. Bull World Health Organ 2006; 84: 21–7. [Crossref]
  24. Chuma J, Gilson L, Molyneux C. Treatment-seeking behaviour, cost burdens and coping strategies among rural and urban households in Coastal Kenya: an equity analysis. Trop Med Int Health 2007; 12: 673–86. [Crossref]
  25. Sun Q, Liu X, Meng Q, Tang S, Yu B, Tolhurst R. Evaluating the financial protection of patients with chronic disease by health insurance in rural China. Int J Equity Health 2009; 8: 42. [Crossref]
  26. Hao Y, Wu Q, Zhang Z, Gao L, Ning N, Jiao M, et al. The impact of different benefit packages of medical financial assistance scheme on health service utilization of poor population in rural China. BMC Health Serv Res 2010; 10: 170. [Crossref]
  27. Shi W, Chongsuvivatwong V, Geater A, Zhang J, Zhang H, Brombal D. The influence of the rural health security schemes on health utilization and household impoverishment in rural China: data from a household survey of western and central China. Int J Equity Health 2010; 9: 7. [Crossref]
  28. Dror DM, van Putten-Rademaker O, Koren R. Cost of illness: evidence from a study in five resource-poor locations in India. Indian J Med Res 2008; 127: 347–61.
  29. Engelgau MM, Karan A, Mahal A. The economic impact of non-communicable diseases on households in India. Global Health 2012; 8: 9. [Crossref]
  30. Saksena P, Xu K, Evans DB. Impact of out-of-pocket payments for treatment of noncommunicable diseases in developing countries: a review of literature. Discussion Paper No 2. Geneva: World Health Organization; 2011.
  31. Engelgau M, Rosenhouse S, El-Saharty S, Mahal A. The economic effect of noncommunicable diseases on households and nations: a review of existing evidence. J Health Commun 2011; 16(Suppl 2): 75–81. [Crossref]
  32. Giang TL. Health status and accessibility to health facilities of the elderly in Vietnam. Social Insur Rev 2008; 7: 37–43.
  33. Giang TL, Dfau WD. Demographic changes and the long-term pension finances in Vietnam: a stochastic actuarial assessment. J Popul Ageing 2009; 1: 125–51. [Crossref]
  34. Lieberman SS, Wagstaff A. Health financing and delivery in Vietnam: the short- and medium-term policy agenda. Hanoi: World Bank; 2008.
  35. Wagstaff A. Health insurance for the poor: initial impacts of Vietnam's health care fund for the poor. Policy Research Working Paper No. 4134. Volume Impact Evaluation Series No. 11. Washington, DC: World Bank; 2007.
  36. Wagstaff A, Pradhan M. Health insurance impacts on health and nonmedical consumption in a developing country. Policy Research Working Paper No. 3563. Washington, DC: World Bank; 2005.
  37. Wagstaff A, van Doorslaer E. Catastrophe and impoverishment in paying for health care: with applications to Vietnam 1993–1998. Health Econ 2003; 12: 921–34. [Crossref]
  38. O'Donnell O, Rannan-Eliya RP, Somanathan A, van Doorslaer E, Adhikari SR, Harbianto, et al. Who benefits from public spending on health care in Asia? EQUITAP Project Working Paper No. 3; 2005. Available from: http://www.equitap.org/publications/docs/EquitapWP3.pdf

*Hoang Van Minh
Department of Health Economics
Center for Health System Research
Institute for Preventive Medicine and Public Health
Hanoi Medical University
Hanoi, Vietnam
Email: hvminh71@yahoo.com

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Survival probability and prognostic factors for breast cancer patients in Vietnam

Nguyen H. Lan1,2, Wongsa Laohasiriwong3* and John F. Stewart4

1Graduate School, Khon Kaen University, Khon Kaen, Thailand; 2Hue College of Medicine and Pharmacy, Hue University, Hue, Vietnam; 3Faculty of Public Health, Khon Kaen University, Khon Kaen, Thailand; 4Department of Economics, University of North Carolina, Chapel Hill, NC, USA

Abstract

Background: Breast cancer is becoming a public health problem in Vietnam. The mortality to incidence ratio of the disease was ranked second among the most common cancers in women. This study estimates the survival probability at 1, 3, and 5 years following diagnosis and determines prognostic factors for breast cancer mortality in Vietnam.

Methods: A survival analysis was conducted based on retrospective data from Hue Central Hospital and the Cancer Registry in Ho Chi Minh City. Using the Kaplan-Meier method, the survival probability of patients with breast cancer was estimated at 1, 3, and 5 years following diagnosis. The covariates among prognostic factors for survival time were studied using an extended Cox proportion hazards model, including time-dependent predictors.

Results: Overall survival rates at 1, 3, and 5 years following diagnosis were 0.94, 0.83 and 0.74 respectively. Marital status, education level, stage at diagnosis, and hormone therapy were prognostic factors for mortality. For the stage at diagnosis, the relation to the risk of death for breast cancer was 1.32 (95% CI, 1.22–1.41). Married women faced a risk of death nearly 1.59 times higher than unmarried women (95% CI, 1.09–2.33). Women with higher levels of education and who received hormone therapy had approximately 10% (hazard ratio [HR]: 0.92; 95% CI, 0.89–0.96) and 80% (HR: 0.22; 95% CI, 0.12–0.41) risk reduction of death respectively, compared with those classified as illiterate and those without hormone therapy.

Conclusions: The 5-year survival probability of breast cancer was lower in Vietnam than in countries with similar distributions of the stage at diagnosis. Screening programs and related support policies should be developed to increase the life expectancy of women with breast cancer in Vietnam.

Keywords: breast cancer; survival; prognostic factors; Vietnam

Received: 19 May 2012; Revised: 8 December 2012; Accepted: 21 December 2012; Published: 17 January 2013

Glob Health Action 2013. © 2013 Nguyen H. Lan et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 18860 - http://dx.doi.org/10.3402/gha.v6i0.18860

 

Breast cancer is the most common cancer among women, with more than 1.1 million women newly diagnosed globally every year. Death from breast cancer accounts for 1.6% of female deaths every year worldwide (1). The mortality rates of breast cancer vary among different communities and countries. The 5-year overall survival rate tends to be lower in developing countries (2, 3). The relationship between survival time and stage at diagnosis is consistent in the literature (47). Detection at an earlier stage of the disease and better access to effective treatment have been recommended as solutions to improve the life expectancy of breast cancer patients (3, 8). Other prognostic factors, such as demography, socioeconomics, hormone receptors, and psychology, were found to have varying effects on breast cancer survival rates in different studies (914).

Policy recommendations from the study

  • An early detection strategy for breast cancer should be developed to improve life expectancy of women with breast cancer in Vietnam.
  • Breast cancer education and awareness should be supported hand in hand with early detection programs.
  • Financial support policies should be considered to promote access of women to appropriate diagnosis and treatment, particularly for the most economically disadvantaged.

The breast cancer incidence in Vietnam has increased steadily over the last decade from a crude rate of 13.8 per 100,000 women in 2000 to 28.1 per 100,000 women in 2010, with an estimated 12,533 breast cancer cases in the country (15). The mortality to incidence ratio of breast cancer was 0.44, ranked second in mortality of common cancers in Vietnamese women in 2002 (16). Previous studies in Vietnam estimated the overall 5-year survival rate to be 85.1%; this rate varied with the stage of breast cancer. The prognostic factors, the hormonal receptors, such as estrogen receptor (ER), progesterone receptor (PR), and hormone therapy, were reported inconsistently in these studies (17, 18). While previous studies investigated survival time only for specific patient groups such as premenopausal breast cancers with/without positive hormone receptors or patient groups who received supportive treatment with chemotherapy and/or hormones, the present study used a more general sample and included demographic factors not previously considered in Vietnam. The objectives of the study were to estimate the survival probability at 1, 3, and 5 years, following diagnosis of breast cancer patients, and to determine the effects of age, ethnicity, stage at diagnosis, marital status, educational level, type of primary treatment, and hormone therapy on mortality of breast cancer in Vietnam. The results of the study could help health policy makers engaged in planning for breast cancer interventions and can be used in models of cost-effectiveness analysis of interventions for breast cancer in the country.

Methods

Study settings

The study was conducted in Ho Chi Minh City and Thua Thien Hue province. Ho Chi Minh City in the South of Vietnam is the largest economic center in the country and the most crowed city, with a population of 7,196,100 in 2009 (19). The Ho Chi Minh Cancer Registry, a population-based cancer registry established in 1990, is one of the two cancer registries in Vietnam recognized by the International Association of Cancer Registries (IACR) (20, 21). Thua Thien Hue province is located in central Vietnam; its capital, Hue City, is famous as a cultural center of the country. The province had 1,087,600 people in 2009 (19). Hue Central Hospital is responsible for a cancer registry covering all provinces in the central region. It was established in 2000 and developed into a population-based cancer registry in 2008. These cancer registries do not include data on follow-up or death certificates of cancer patients (20). Breast cancer is the most common cancer among women in both study settings with the age standardized rate (ASR) of the disease estimated at 21/100,000 in Ho Chi Minh City and 11.9/100,000 in Thua Thien Hue province between 2004 and 2008 (22).

Data source and subjects

The morbidity data of the study came from the Ho Chi Minh Cancer Registry and the medical record reviews from Hue Central Hospital.

Inclusion criteria: all cases of female breast cancer with code of C 50 (ICD-10 version) (23) who met the following criteria were included in the study: age between 35 and 75 years; diagnosis with primary breast cancer in the period 2001–2006; resident of Ho Chi Minh City or Thua Thien Hue province; staging at diagnosis based on tumor/nodes/metastasis staging system (TNM) of the Union for International Cancer Control (UICC) (24).

A total of 1,584 cases of breast cancer were eligible for the study. This included 1,427 cases registered between 2003 and 2006 in Ho Chi Minh City. The remaining 157 cases were diagnosed between 2001 and 2006 in Thua Thien Hue province. Secondary data from the Cancer Registry in Ho Chi Minh City, and medical records in Hue Central Hospital provided information on age, ethnic group, date of diagnosis, stage at diagnosis, and the primary treatment following diagnosis. Those patients would be on the track in their community until December 31, 2010. The result of follow-up left 948 patients or their relatives (if the patients were deceased) available for direct interview. There was incomplete follow-up data for 636 patients (40.2%). The most common reason for incomplete follow-up was migration, particularly in Ho Chi Minh City. Population shifts are understandable in the biggest and the most crowded city in Vietnam. The data in Table 1 compare the characteristics between the two groups of patients, with and without loss to follow-up.


Table 1.  Characteristics of breast cancer patients for groups with and without loss of follow-up
Characteristics Group without loss of follow-up (N=948),% Group with loss of follow-up (N=636),% p*
Ethnic group      
  Majority (Kinh) 94.41 95.75 0.231
  Minority 5.59 4.25  
Age group      
  <40 14.24 13.99  
  40–49 43.25 36.62  
  50–59 26.05 29.72 0.436
  60–69 13.71 13.21  
  ≥70 2.74 3.46  
Stage at diagnosis      
  Stage 0 0.42 0.94  
  Stage I 10.76 11.01 0.094
  Stage II 61.18 57.08  
  Stage III 19.41 24.21  
  Stage IV 8.23 6.76  
The primary treatment      
  Surgery 78.69 72.96  
  Radiation 0.42 0.94 0.039
  Chemotherapy 10.55 11.32  
  Hormone therapy 0.11 0.31  
  No treatment 10.23 14.47  
*Pearson chi-square.

There was no statistically significant difference in most characteristics of breast cancer patients between these two groups (p>0.05). The exception was the primary treatment received. However, this difference was small (p=0.039). The group lost to follow-up was excluded from the survival analysis in the study.

Follow-up information was obtained from direct interviews of 948 patients or their relatives, using structured questionnaires. Information on marital status, educational level, hormone therapy during 5 years, patients status (alive or dead), date of death, and causes of death were collected for each patient.

Prognostic factors

We examined the effects of ethnic group, age, marital status, educational level, stage at diagnosis, type of primary treatment and hormone therapy on survival time of breast cancer patients.

Ethnicity in the sample was classified as majority (Kinh) or minority (primarily Chinese). The age variables were stratified into five groups: aged <40; 40–49; 50–59; 60–69; ≥70. Marital status was recorded as married or unmarried. Unmarried women included those who had never married before they were diagnosed with breast cancer. Educational level was categorized into five groups: illiterate; primary school; secondary school; high school and junior college, or higher. Stage at diagnosis (clinical or pathological) was analyzed as stage 0 (in situ); stage I, stage II (including stage IIA and stage IIB), stage III (including stage IIIA and stage IIIB), and stage IV. Because so few patients presented with stage 0, stage I and stage 0 were combined for the analysis. Primary treatment was classified into one of five categories based on the initial treatment received following diagnosis. The treatment categories were surgery, radiotherapy, chemotherapy, hormone therapy, and no treatment (treatment was not received or was unknown). Patients or their relatives’ recall of prescriptions for Tamoxifen or similar medication during the 5 years after primary treatment was used as an indicator of hormone treatment in this study.

Statistical analysis

The survival duration of each case was determined as the difference in time (months) between the date of initial diagnosis until the date of death (by breast cancer only), or the closing date of follow-up. Death by other causes was considered censored results in the study. Potential follow-up time ranged from 4 to 10 years. The overall survival of patients with breast cancer at 1, 3, and 5 years, following diagnosis was calculated by the Kaplan-Meier method. A Cox regression model was used to examine the effects of characteristics of patients and treatments on survival time. The following variables were included in the model as predictors: ethnic group, age group, marital status, education level, stage at diagnosis, primary treatment, and hormone therapy. A goodness-of-fit (GOF) test was conducted to assess the proportional hazard (PH) assumptions of the Cox model for given predictor variables (Table 2). The findings indicated that variable education level and variable stage at diagnosis did not satisfy PH assumptions (p<0.05). We therefore fitted the extended Cox model in which these variables were analyzed as time-dependent variables (25).


Table 2.  Goodness-of-fit test assessing proportional hazards assumption
Predictors rho* Chi-square test df** p
Ethnic group −0.06846 1.28 1 0.2571
Age group 0.03237 0.29 1 0.5871
Marital status −0.02643 0.19 1 0.6619
Education level −0.16793 7.48 1 0.0062
Stage at diagnosis −0.33341 27.11 1 0.0000
Primary treatment 0.09538 2.76 1 0.0967
Hormone therapy 0.11012 3.12 1 0.5871
Global test   38.11 7 0.0000
*The correlation coefficient between the residuals and time.
**Degree of freedom.

In addition, the possible interaction effects of education level, marital status, hormone therapy, and primary treatments with stage at diagnosis were considered in this full model.

The formula for the extended Cox model for time-dependent variables was applied as following:

where

h(t, Xi, Xj(t)) is Hazard function—expresses the hazard at time t for an individual with a given specification of set of time-independent predictor variables and time-dependent predictor variables.

ho(t) is Baseline hazard function.

Xi denotes the ith time-independent predictor variable.

Xj(t) denotes jth time-dependent predictor variable.

p1Xi is the entire collection of time-independent predictors (including variables, namely ethnic, married, treatment, hormone, age in the study).

p2Xj(t) is the entire collection of time-dependent predictors at time t (including variables stage(t), education(t), treatment*stage(t), married*stage(t), education*stage(t), hormone*stage(t) in the study).

ßi: regression coefficient for time-independent predictor variables.

dj: regression coefficient for time-dependent predictor variables.

Model assumes that hazard at the time t depends on the value of Xj (t) at the same time.

Then hazard ratio (HR) formula for extended Cox model is:

This formula describes the ratio of hazards at a particular time t that compares two specifications of time-independent predictors (X* and X) and those of time-dependent predictors at time t (X*(t) and X(t)). The coefficient δ represents the overall effect of the corresponding time-dependent variables, considering all times at which this variable has been measured in the study (25).

A reduced extended Cox model was then carried out by excluding interaction terms that were not statistically significant. In the final model, the prognostic factors that showed statistically significant effects were analyzed according to subgroups to determine the effect of each stratum on survival time for breast cancer patients. A p-value of less than 0.05 was considered statistically significant.

Results

Study subject characteristics are reported in Table 3. The mean age of patients at the time of diagnosis was 50 years. The most frequent age group was that from 40 to 49 years (43% of the patients). Most the women were married (83.1%). The education level of the study population was low, with 30.1 and 27.8% of women having completed primary school and secondary school, compared with 26.4 and 11.5% of those having attended high school and higher levels. Among the women, 4.2% were illiterate. The majority of the women had been diagnosed with breast cancer at stage II at primary diagnosis (61.2%). Late stage diagnosis (stage III and stage IV) was also considerable, making up 19.4 and 8.2% of the study population, respectively. Surgery was the most common treatment for patients following diagnosis with breast cancer (78.7%). Significantly, 10.2% of all the patients did not receive any type of primary treatment and 10.6% of cases were treated with chemotherapy after diagnosis. There were 74.6% of cases who reported receiving hormone therapy during 5 years following the primary treatment.


Table 3.  Characteristics of patients with breast cancer
Characteristics Number of patients Percent
Ethnic group    
  Majority group 894 94.3
  Minority group 54 5.7
Age at diagnosis    
  Mean (SD) 50 (9.2)  
Age group    
  <40 103 10.9
  40–49 408 43.0
  50–59 270 28.5
  60–69 131 13.8
  ≥70 36 3.8
Marital status    
  Married 788 83.1
  Unmarried 160 16.9
Education level    
  Illiterate 40 4.2
  Primary school 285 30.1
  Secondary school 264 27.8
  High school 250 26.4
  Higher 109 11.5
Stage at diagnosis    
  Stage 0 and I 106 11.2
  Stage II 580 61.2
  Stage III 184 19.4
  Stage IV 78 8.2
The first treatment    
  Surgery 746 78.7
  Radiotherapy 4 0.4
  Chemotherapy 100 10.6
  Hormone therapy 1 0.1
No treatment 97 10.2
Hormone therapy    
  Yes 707 74.6
  No 241 25.4
Total 948 100

The overall survival rate declined over time and was estimated to be 0.94, 0.83, and 0.74 at 1, 3, and 5 years, respectively, following diagnosis for breast cancer (Fig. 1).

Fig 1

Fig. 1.   Overall survival rate at 1, 3, and 5 years following diagnosis for breast cancer.

The findings in Table 4 show evidence of interaction between hormone therapy and stage at diagnosis on the survival time (p=0.023), but there was no significant interaction effect between primary treatment, marital status, education level, and stage on risk of death for breast cancer cases in this study (p>0.05). Estimates of the regression coefficients for time-dependent variables indicate that the death hazard decreases with education level and increases with the stage at diagnosis (Table 4). The results from the extended Cox model analysis, excluding insignificant interaction terms, revealed four characteristics that had prognostic value for survival probability for breast cancer, including marriage (p=0.016), education level (p<0.001), stage at diagnosis (p<0.001), and hormone therapy (p<0.001) (Fig. 2). Marital status had the strongest effect on risk of death, with HR of 1.59 (95% CI: 1.09 to 2.33), followed by stage at diagnosis (HR: 1.32; 95% CI: 1.22 to 1.41). A later stage of breast cancer and the status of ‘married’ were related to poor prognosis for survival probability. In contrast, a higher education level and hormone therapy reduced mortality, with HR of 0.92 (95% CI: 0.89 to 0.96) and HR of 0.22 (95% CI: 0.12 to 0.41), respectively.

Fig 2

Fig. 2.   The extended Cox model assessing the effect of characteristics of patients on mortality from breast cancer.


Table 4.  Extended Cox model, including patient characteristics and interaction terms
  Hazard ratio p 95% CI
Ethnic group 1.01 0.961 0.61–1.67
Marital status 2.41 0.093 0.86–6.74
Primary treatment 0.89 0.542 0.61–1.30
Hormone therapy 0.22 0.000 0.12–0.40
Age group 1.11 0.100 0.98–1.25
Education level* 0.86 0.034 0.76–0.99
Stage at diagnosis* 1.25 0.019 1.04–1.51
Primary treatment×stage** 1.02 0.235 0.99–1.06
Marital status×stage** 0.95 0.367 0.85–1.06
Education level×stage** 1.03 0.325 0.98–1.08
Hormone therapy×stage** 1.08 0.023 1.01–1.16
95% CI: 95% confidence interval for hazard ratio.
*Variables continuously vary with respect to time.
**Interaction terms defined as theproduct of hormone therapy and stage at diagnosis
Estimate of regression coefficient for variable education level is −0.14.
Estimate of regression coefficient for variable stage at diagnosis is 0.22.

A further analysis of these prognostic factors is shown in Table 5. The hazard ratio for stage IV was 2.27 times higher than that for stage I. Married women faced nearly 60% higher risk of death than unmarried women. Women with an education level of high school and higher and women who had received hormone therapy reduced risk of death by approximately 10 and 80%, respectively, compared with those classified as illiterate and those without hormone therapy.


Table 5.  Subgroup analysis of potential prognostic factors
Prognostic factors Hazard ratio p 95% CI
Marital status      
  Unmarried 1 (baseline)    
  Married 1.53 0.030 1.04–2.24
Hormone therapy      
  No 1 (baseline)    
  Yes 0.24 0.000 0.13–0.44
Education level*      
  Illiterate 1 (baseline)    
  Primary school 1.00 0.974 0.83–1.20
  Secondary school 0.95 0.571 0.78–1.14
  High school 0.79 0.020 0.64–0.96
  Higher 0.79 0.046 0.62–0.99
Stage at diagnosis*      
  Stage I 1 (baseline)    
  Stage II 1.24 0.060 0.99–1.56
  Stage III 1.64 0.000 1.29–2.09
  Stage IV 2.27 0.000 1.71–2.99
Hormone therapy×stage** 1.07 0.053 0.99–1.14
95% CI: 95% confidence interval for hazard ratio.
*Variables continuously vary with respect to time.
**Interaction terms defined as the product of hormone therapy and stage at diagnosis.

Discussion

The study showed similarities in the epidemiology of breast cancer in Vietnam and other Asian countries. The disease was most common in women aged between 40 and 49, while in the West incidence peaks among women aged between 60 and 70 years (13). The majority of women in the study population were newly diagnosed at stage II of breast cancer, whereas 60–70% of cases in developed countries were detected at earlier stages (stage I and stage 0) (1, 2). Consequently, surgery was the most common primary treatment in the study. Mastectomy or conserving surgery and/or axillary dissection are the favored treatments for breast cancer patients at early stages (stage I and stage II) in almost all countries (2). The survival probability of breast cancer in this study population decreased with time from initial diagnosis. Overall survival rates at 1, 3, and 5 years in our study were higher than those in a similar study from Chennai (known as previous Madras), India, in 1997, at 94; 83, and 74% versus 80, 58, and 48%, respectively (6). The 5-year overall survival of the study population was also higher than that of Uganda (56%) in 2008 (5). However, the rate revealed in the study was rather lower than those in other Asian countries, such as China (76.5%), Taiwan (78.37%), Japan (76.1–86.1%), and South Korea (83.5%) (2, 7), and even lower than previous results in North Vietnam (85.1%) (17, 18). This figure is still lower when compared with Western countries such as Sweden (89%), Canada (86%), and the United States (88%) (2).

Among prognostic factors, those who were diagnosed at a higher stage demonstrated a poor prognosis for survival time. This result is consistent with previous findings in many studies all over the world (47, 26). One reason for the lower survival rate in our study is the late stage at which breast cancer is detected compared to developed countries (2). Moreover, the absence of primary treatment for 10.2% of the study population compared with previous studies in Vietnam, in which all patients received standard treatment after diagnosis, may explain the poorer survival rate of breast cancer patients in this study. Lack of appropriate treatment could also be the reason for the higher mortality in our study compared to other Asian countries with the same stage distribution at diagnosis (35, 26). Indeed, although there have been many efforts in cancer control in Vietnam, these activities have only met 15% of demand. The country is facing a lack of infrastructure, equipment, and human resources in oncology. Programs for the early detection of breast cancer have not yet been implemented nationwide. Cancer awareness in the community is still low. These are the main reasons for presentation at hospital with late stages of cancer. In addition, most facilities with capacity to treat cancer are located in a few big cities; hence, most cancer patients must be referred to the central level (27). Our findings suggest that early detection of breast cancer along with the availability and accessibility of appropriate treatment should be recommended to improve life expectancy for women with the disease in Vietnam.

Being married appeared as the most unfavorable prognostic factor for survival probability for women with breast cancer in our study. A review by Falagas et al. (2007) showed that the effect of marital status on breast cancer outcome varied across studies (9). The benefits of emotional support, good lifestyle, and stable economics were likely protective factors on survival for married women with cancer in the previous studies (6, 9, 10, 28). However, other recent studies indicated that social support and social network had a more important role in reduced mortality for breast cancer, and that support received from social networks improved survival in unmarried women (29, 30). The marital status alone had no significant influence on longer survival of women with breast cancer (3133). In the context of Vietnam, married women, especially women with many children, often faced difficult economic circumstances. The financial burden of the treatment course for breast cancer could be a barrier to seeking care and to appropriate treatment compliance, more importantly, when patients must pay out of their pockets for health care services. This may contribute to the higher mortality of breast cancer among married women in this study. In this situation, universal coverage of health insurance should be promoted in Vietnam. Besides, the government budget to support vulnerable groups in health care expenditure should include cancer patients for whom the cost exceeds their ability to pay. This combination would encourage the patients in compliance with their long-term treatment and thus contribute to reducing deaths from cancers, including breast cancer.

The association between education level and survival rate of breast cancer was inconsistent across studies around the world (6, 12, 34). Generally, our study found a positive effect of education level on survival for breast cancer patients. The higher the level of education, the lower the risk of death due to breast cancer. This finding may reflect increasing knowledge and awareness of the disease; appropriate accessibility to health services; and good compliance to treatment and clinical follow-up by the higher education groups (6, 12). As mentioned earlier, the limited awareness and knowledge of cancer at the community level has been common in Vietnam (27). It is necessary to educate the public on breast health. The communication should be relevant to all subjects for the purposes of early detection of breast cancer, prevention of breast cancer, and prevention of death from the disease.

The final prognostic factor found in the study was hormone therapy within a 5-year treatment course. This therapy reduced the risk of death among breast cancer patients in the study population, in accordance with previous reports (35, 36). The effect of hormone therapy on survival improvement was often mentioned along with the influence of hormone receptors. An overview of randomized trials indicated that Tamoxifen was more effective in ER-positive than in ER-negative women (36). More importantly, Crowe et al. concluded that the presence of these hormonal factors impacted on longer survival for breast cancers (37). Prior studies in Vietnam found inconsistent effects of hormone receptors on prolonging survival time of women with breast cancer. However, all patients in these studies were taking hormone therapy during their treatment course (17, 18). It is difficult to identify the role of hormone therapy or hormone receptors as independent prognostic factors. A study by Caldarola et al. reported no significant relationship between ER or PR and improved survival (38). In our study, some of the patients had unknown hormonal receptors because the assay techniques used for ER/PR testing were not available in Hue Central Hospital before 2006. They were prescribed hormone therapy based on the experience of specialized doctors (personal communication, Dr. Nguyen Dinh Tung, Oncology Department, Hue Central Hospital). The greatest improvement in survival from breast cancer of hormonal therapy in our study (HR: 0.22) was strong enough to confirm this as an independent prognostic factor. Moreover, the presence of hormone therapy reduced the effect of stage on the risk of death from breast cancer (HR of interaction term: 1.08).

A number of limitations should be considered when interpreting the findings of this study. The high proportion of patients lost to follow-up (40.2%) and their exclusion from the survival analysis may have resulted in selection bias although their characteristics did not differ from those included in the analysis. The data were collected over the period 2001–2006 and do not reflect current utilization of advanced treatment methods and new medications for breast cancer treatment, which could affect the opportunity to improve survival probability in the study population. Moreover, diagnosis of breast cancer in the study population was defined as the time when patients presented at hospitals for primary treatment, so the lead-time bias could be a potential problem. The risk of recall bias can occur with information on education level and hormone therapy following primary treatment. Although the study population was representative for two big cities in Vietnam with different characteristics in socioeconomics and epidemiology of breast cancer, the generalizability of the study findings should be considered. The presence of high level hospitals might increase accessibility and availability of relevant treatment; the survival time estimates of our study could be better than those in other areas of the country where diagnosis and treatment techniques varied (15, 22). Despite these limitations, this is the first population-based survival analysis for breast cancer in Vietnam. Risk factors on mortality of breast cancer were identified. The results will contribute important information to cost-effectiveness analysis of interventions for breast cancer and will help health policy makers engaged in planning for programs to reduce breast cancer mortality. Among the risk factors, stage at diagnosis showed a strong relation to mortality on breast cancer. Developing early detection strategies for breast cancer to shift the stage at presentation to more favorable early stages is necessary to improve the life expectancy of women with breast cancer in Vietnam. Breast cancer education and awareness should be supported hand-in-hand with early detection programs. The poorest prognosis on survival time of married women with breast cancer suggested that financial support policies are needed to promote access of women, especially poor women, to appropriate diagnosis and treatment. Besides, the most important factor to achieve the goal of reducing breast cancer mortality is that infrastructure, equipment, and human resources for diagnosis and breast cancer treatment should be made available, accessible, and affordable to women in whom early detection strategies identify breast cancer. Overall, these recommendations are drawn from the evidence found in the study. Decisions about these policies should consider the available resources in the country and should also inform directions for future plans.

Conclusion

The survival probability of breast cancer patients in Vietnam decreased over time following diagnosis. Five-year survival rate was lower than that in other countries with similar distributions of stage at diagnosis. Married status, late stage at diagnosis, low education level, and lack of hormone therapy were prognostic factors for higher mortality of women with breast cancer. An early detection program with support policies could reduce breast cancer mortality in Vietnam.

Ethical approval

Ethical approval for the data collection was obtained from the University of Khon Kaen, Thailand (where the study was designed as part of a doctoral study program). In addition, approval for implementing the study at the sites was obtained from Health Services of Thua Thien Hue province and Ho Chi Minh City.

Acknowledgements

We would like to thank Prof. Pamela Wright, Director of Medical Committee Netherlands–Vietnam for her the editorial assistance. We would also like to thank Assoc. Prof. Bandit Thinkhamrop for his advice on statistical analysis. The research was funded by the Vietnam-Netherlands Project among Eight Medical Universities in Vietnam (Royal Netherlands Embassy in Hanoi) and the Graduate School of Khon Kaen University, Thailand.

Conflict of interest and funding

The authors have no potential conflict of interest. The research was funded by the Vietnam-Netherlands Project among eight Medical Universities in Vietnam, Royal Netherlands Embassy in Hanoi, and by the Graduate School of Khon Kaen University, Thailand.

References

  1. Peter B, Bernard L, eds. World cancer report 2008 [pdfs online]. WHO press: The International Agency for Research on Cancer (IARC). Available from: www.iarc.fr/en/publications/pdfs_online/wrc/2008/wcr_2008.pdf [cited 11 June 2011].
  2. Leong SP, Shen ZZ, Liu TJ, Agarwal G, Tajima T, Paik NS, et al. Is breast cancer the same disease in Asian and Western countries? World J Surg 2010; 34: 2308–24. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  3. Bray F, McCarron P, Parkin DM. The changing global patterns of female breast cancer incidence and mortality. Breast Cancer Res 2004; 6: 229–39. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  4. Meng L, Maskarinec G, Wilkens L. Ethnic differences and factors related to breast cancer survival in Hawaii. Int J Epidemiol 1997; 26: 1151–8. PubMed Abstract | Publisher Full Text
  5. Gakwaya A, Kigula-Mugambe JB, Kavuma A, Luwaga A, Fualal J, Jombwe J, et al. Cancer of the breast: 5-year survival in a tertiary hospital in Uganda. Br J Cancer 2008; 99: 63–7. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  6. Gajalakshmi CK, Shanta V, Swaminathan R, Sankaranarayanan R, Black RJ. A population-based survival study on female breast cancer in Madras, India. Br J Cancer 1997; 75: 771–5. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  7. Lee JH, Yim SH, Won YJ, Jung KW, Son BH, Lee HD, et al. Population-based breast cancer statistics in Korea during 1993–2002: incidence, mortality, and survival. J Korean Med Sci 2007; 22(Suppl): S11–6. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  8. Anderson BO, Yip CH, Smith RA, Shyyan R, Sener SF, Eniu A, et al. Guideline implementation for breast healthcare in low-income and middle-income countries: overview of the Breast health global initiative global summit 2007. Cancer 2008; 113(8 Suppl): 2221–43. PubMed Abstract | Publisher Full Text
  9. Falagas ME, Zarkadoulia EA, Ioannidou EN, Peppas G, Christodoulou C, Rafailidis PI. The effect of psychosocial factors on breast cancer outcome: a systematic review. Breast Cancer Res 2007; 9: R44. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  10. Kravdal H, Syse A. Changes over time in the effect of marital status on cancer survival. BMC Public Health 2011; 11: 804. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  11. Heck KE, Wagener DK, Schatzkin A, Devesa SS, Breen N. Socioeconomic status and breast cancer mortality, 1989 through 1993: an analysis of education data from death certificates. Am J Public Health 1997; 87: 1218–22. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  12. Hussain SK, Altieri A, Sundquist J, Hemminki K. Influence of education level on breast cancer risk and survival in Sweden between 1990 and 2004. Int J Cancer 2008; 122: 165–9. PubMed Abstract | Publisher Full Text
  13. Dunnwald LK, Rossing MA, Li CI. Hormone receptor status, tumor characteristics, and prognosis: a prospective cohort of breast cancer patients. Breast Cancer Res 2007; 9: R6. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  14. Fletcher AS, Erbas B, Kavanagh AM, Hart S, Rodger A, Gertig DM. Use of hormone replacement therapy (HRT) and survival following breast cancer diagnosis. Breast 2005; 14: 192–200. PubMed Abstract | Publisher Full Text
  15. Duc NB, et al. Epidemiology and program of control and prevention for cancer: preliminary report of national cancer project period 2008–2010. Viet J Oncol 2010: 21–31. Vietnamese.
  16. The Burden of Cancer in Asia. Pfizer facts; 2008. Available from: http://www.pfizer.com/files/products/cancersin_asia.pdf [cited 19 March 2012].
  17. Dieu B, Trang NTP, Binh ND. Premenopausal breast cancers with positive hormonal receptors: 5 year survival rate of surgical department of National Cancer hospital—Tam Hiep campus. Viet J Oncol 2010: 426–432. Vietnamese.
  18. Thang VH, Tu DA, Hoa NTT, Toan VQ, Hung TT, Huong TTTH. Analysis of predictive value of epidermal growth factors in breast cancer. Viet J Oncol 2010: 448–450. Vietnamese.
  19. General Statistics Office. Population of each region in Vietnam in 2010. Available from: http://www.gso.gov.vn/default.aspx?tabid=387&idmid=3&ItemID=11503 [cited 10 November 2011].
  20. Nguyen Dinh Tung. Cancer incidence in the population of Thua Thien Hue province, Vietnam, 2001–2009. J Sci, Hue University 2010; 27(61): 511–17.
  21. International Association of Cancer Registries IACR. Membership list. Available from: http://www.iacr.com.fr/asiact.htm [cited 2 May 2012].
  22. Duc NB, Thuan TV, Can DT, Dieu B, Nga NTH, Thang ND, et al. Situation of female breast cancer in some provinces and cities period 2001–2007. Viet J Oncol 2009: 5–11. Vietnamese.
  23. World Health Organization. International classification of diseases 10th revision. ICD-10 version; 2010. Available from: http://apps.who.int/classification/icd10/browse/2010/en#/II [cited 2 March 2012].
  24. Blogspot.com. TNM classification help: Breast tumours. Available from: http://cancerstaging.blogspot.com/2005/02/breast-tumours.html [cited 21 March 2012].
  25. David GK, Mitchel K. Survival analysis. A self-learning text. 2nd ed. New York: Springer Science and Business Media; 2005.
  26. Jack RH, Davies EA, Moller H. Breast cancer incidence, stage, treatment and survival in ethnic groups in South East England. Br J Cancer 2009; 100: 545–50. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  27. Vietnamscope. Situation: cancer situation [updated 21 September 2010]. Available from: http://www.vietnamscope.org/index.php?option=com_content&view=article&id=99&Itemid=139 [cited 25 May 2012].
  28. Osborne C, Ostir GV, Du X, Peek MK, Goodwin JS. The influence of marital status on the stage at diagnosis, treatment, and survival of older women with breast cancer [abstract]. Breast Cancer Res Treat 2005; 93: 41–7. PubMed Abstract | Publisher Full Text
  29. Kroenke CH, Kubzansky LD, Schernhammer ES, Holmes MD, Kawachi I. Social networks, social support, and survival after breast cancer diagnosis. J Clin Oncol 2006; 24: 1105–11. PubMed Abstract | Publisher Full Text
  30. Antonucci TC. An examination of sex differences in social support among older men and women. Sex Roles 1987; 17: 737–49. Publisher Full Text
  31. Laino C. Marital status doesn't affect breast cancer outcomes. Oncol Times 2005; 27: 53–5. Publisher Full Text
  32. Goodwin JS, Hunt WC, Key CR, Samet JM. The effect of marital status on stage, treatment, and survival of cancer patients. JAMA 1987; 258: 3125–30. PubMed Abstract | Publisher Full Text
  33. Goodwin PJ, Leszcz M, Ennis M, Koopmans J, Vincent L, Guther H, et al. The effect of group psychosocial support on survival in metastatic breast cancer. N Engl J Med 2001; 345: 1719–26. PubMed Abstract | Publisher Full Text
  34. Valanis B, Wirman J, Hertzberg VS. Social and biological factors in relation to survival among black vs. white women with breast cancer. Breast Cancer Res Treat 1987; 9: 135–43. PubMed Abstract | Publisher Full Text
  35. Sener SF, Winchester DJ, Winchester DP, Du H, Barrera E, Bilimoria M, et al. The effects of hormone replacement therapy on postmenopausal breast cancer biology and survival. Am J Surg 2009; 197: 403–7. PubMed Abstract | Publisher Full Text
  36. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG). Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of randomised trials. Lancet 2005; 365: 1687–717. Publisher Full Text
  37. Crowe JP, Jr., Gordon NH, Hubay CA, Shenk RR, Zollinger RM, Brumberg DJ, et al. Estrogen receptor determination and long term survival of patients with carcinoma of the breast. Surg Gynecol Obstet 1991; 173: 273–8. PubMed Abstract
  38. Caldarola L, Volterrani P, Caldarola B, Lai M, Jayme A, Gaglia P. The influence of hormone receptors and hormonal adjuvant therapy on disease-free survival in breast cancer: a multifactorial analysis. Eur J Cancer Clin Oncol 1986; 22: 151–5. PubMed Abstract | Publisher Full Text

*Wongsa Laohasiriwong
Faculty of Public Health
Khon Kaen University
Khon Kaen, 40002, Thailand
Tel: +66 897 121 455
Fax: +6643347058
Email: drwongsa@gmail.com

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Cost of treatment for breast cancer in central Vietnam

Nguyen Hoang Lan1,2, Wongsa Laohasiriwong3*, John Frederick Stewart4, Nguyen Dinh Tung5 and Peter C. Coyte6

1Graduate School, Khon Kaen University, Khon Kaen, Thailand; 2Hue College of Medicine and Pharmacy, Hue University, Hue city, Vietnam; 3Faculty of Public Health and Board Committee of Research and Training Center for Enhancing Quality of Life of Working Age People (REQW), Khon Kaen University, Khon Kaen, Thailand; 4Department of Economics, University of North Carolina, Chapel Hill, USA; 5Department of Oncology, Hue Central Hospital, Hue city, Vietnam; 6Institute of Health Policy, Management and Evaluation, University of Toronto, Toronto, Ontario, Canada

Abstract

Background: In recent years, cases of breast cancer have been on the rise in Vietnam. To date, there has been no study on the financial burden of the disease. This study estimates the direct medical cost of a 5-year treatment course for women with primary breast cancer in central Vietnam.

Methods: Retrospective patient-level data from medical records at the Hue Central Hospital between 2001 and 2006 were analyzed. Cost analysis was conducted from the health care payers’ perspective. Various direct medical cost categories were computed for a 5-year treatment course for patients with breast cancer. Costs, in US dollars, discounted at a 3% rate, were converted to 2010 after adjusting for inflation. For each cost category, the mean, standard deviation, median, and cost range were estimated. Median regression was used to investigate the relationship between costs and the stage, age at diagnosis, and the health insurance coverage of the patients.

Results: The total direct medical cost for a 5-year treatment course for breast cancer in central Vietnam was estimated at $975 per patient (range: $11.7–$3,955). The initial treatment cost, particularly the cost of chemotherapy, was found to account for the greatest proportion of total costs (64.9%). Among the patient characteristics studied, stage at diagnosis was significantly associated with total treatment costs. Patients at later stages of breast cancer did not differ significantly in their total costs from those at earlier stages however, but their survival time was much shorter. The absence of health insurance was the main factor limiting service uptake.

Conclusion: From the health care payers’ perspective, the Government subsidization of public hospital charges lowered the direct medical costs of a 5-year treatment course for primary breast cancer in central Vietnam. However, the long treatment course was significantly influenced by out-of-pocket payments for patients without health insurance.

Keywords: breast cancer; direct medical cost; health care payer; Vietnam

Received: 29 May 2012; Revised: 30 December 2012; Accepted: 3 January 2013; Published: 4 February 2013

Glob Health Action 2013. © 2013 Nguyen Hoang Lan et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 18872 - http://dx.doi.org/10.3402/gha.v6i0.18872

 

Breast cancer is the most common cancer among women worldwide (1). Advances in screening programs and treatment methods have improved the life expectancy of patients with breast cancer (2). From a societal perspective, the economic burden of this disease has been reported in several studies using available data in developed countries; however, the direct medical cost is thought to make the smallest contribution to total costs, accounting for 50% of indirect costs (morbidity and mortality) (3, 4). These medical costs, while a small proportion of overall costs, may overwhelm patients, particularly those with lower incomes. A study by Chu et al. in Taiwan found that, among major cancers, breast cancer was associated with the largest total lifetime medical costs at 5,046 million TWD (5). In the United States, Barron et al. (2008), using pooled administrative data for five US health plans, estimated costs of breast cancer treatment per patient per month at $2,896 or approximately at $34,752 per year (6). Medical costs were found to increase with the stage of the disease (3, 4, 6, 7). In 1996, Legorreta et al., using US medical records and claims data, determined that costs over a 4-year period for patients with stage III breast cancer averaged more than $60,000, whereas costs were lower in patients at stage 0, I, and IV at $19,000, $21,000, and approximately $40,000, respectively (8).

Policy recommendations from the study

  • Earlier diagnosis of breast cancer should be enabled through screening programs to increase treatment effectiveness and to save health care resources.
  • Universal health insurance coverage should be given more attention, especially since public hospital charges are expected to increase in the near future.
  • The Vietnamese government should have a policy to support cancer patients when the cost of their illness is expected to exceed their ability to pay (with or without health insurance).

Breast cancer has also become an important public health problem in Vietnam. The incidence rate increased from 13.8 per 100,000 women in 2000 to 28.1 per 100,000 women in 2010. In 2010, it was reported that there were 12,533 women with breast cancer in the country (9). In Vietnam, as in many other developing countries, breast cancer was characterized by late presentation, young patients, and low survival rates (1013). Recent studies in Vietnam have revealed that poor knowledge and awareness among the general public is a major contributor to those problems (13). However, the financial burden of treatment of breast cancer has not yet been considered as a contributing factor. The objective of this study is to provide estimates of the total direct medical costs for breast cancer treatment in central Vietnam. The findings can contribute to models of cost-effectiveness analysis of interventions for breast cancer and can support policy adaptations for better care of the women with this disease in Vietnam.

Methods

Study design

A retrospective study was designed to estimate the cost of treatment for women with breast cancer in central Vietnam. Medical records of patients with a code of C 50 (ICD-10 version) admitted to Hue Central Hospital (HCH) between 2001 and 2006 were searched to identify breast cancer patients presenting in those years (14). Data, from medical records and participant's recall, on the patients’ costs for medical care for breast cancer were collected for a period of 5 years following primary diagnosis. Calculation of expenditure for breast cancer treatment was based on actual patient-level cost data, excluding the costs for herbal treatment or unpaid family care, because it is difficult to control these costs, especially in the context of the many variations of herbal medicines in Vietnam. Unit costs during the period of study were provided from the financial department of the hospital. The direct medical cost of treatment for women with breast cancer was analyzed from the perspective of health care payers, including the cost borne by patients and health insurance providers. The payment amount or hospital fee included the cost of medications and materials used in clinical practice together with the user fees borne by patients. User fees are based on a decree on partial collection of public hospital fees as regulated by the Vietnamese government (1994) and the decree's revisions (15, 16). The direct non-medical costs (e.g. travel, accommodation, time) and indirect costs (e.g. lost income or premature death due to the disease) were not included in these calculations.

Data sources

The data were collected from two sources:

Primary data: Patients or their relatives (if patients were deceased) were interviewed directly using a structured questionnaire. Data on sociodemographic characteristics, the type of initial treatment received during hospitalization as well as during a 5-year follow-up period after the initial treatment, and compliance with the treatment regime for follow-up care were collected. For deceased patients, the date and cause of death were also noted.

Secondary data: The paper charts stored at Hue Central Hospital of patients with breast cancer were examined to obtain personal information (e.g. name, age, home address), date of admission, diagnosis and stage, treatment regimes, itemized invoices, and health insurance participation. Unit costs for treatments received over the study period were acquired from the hospital's finance department.

Study population

HCH was selected as the site for the study. This hospital is located in the city of Hue, the capital of the central coastal Thua Thien Hue province. HCH is one of the three largest hospitals under the management of the Ministry of Health in Vietnam. The hospital is a national general hospital and a leading referral hospital in the central region. The medical records of patients presenting with breast cancer at HCH were screened to identify those meeting the following criteria – inpatient admission to the hospital between January 2001 and January 2006, residents of Thua Thien Hue province, diagnosis of primary breast cancer identified in the paper charts by the occurrence of code C 50 (ICD-10 version), and evidence of stage of breast cancer according to the tumor/nodes/metastasis (TNM) staging system of the Union for International Cancer Control (UICC) (17); 160 patients were identified according to the criteria for inclusion and tracked until December 31, 2010, to determine their 5-year survival time. The time period from 2001 to 2006 was used to obtain a more comprehensive sample of patients at various stages of breast cancer from different age groups. This long time period was necessary because the incidence rate of the disease in Thua Thien Hue province was not high (9). The results of follow-up left 129 eligible patients for whom costs could be analyzed. The main reason for 31 cases being lost to follow-up was migration to another treatment site.

Diagnosis to define breast cancer

In HCH, between 2001 and 2006, major laboratory tests were often requested to define breast cancer, including breast ultrasound, hematogram, CA 15.3, tumor biopsy, and cytological tests. Some patients might also have had mammography and estrogen receptor (ER) tests, progesterone receptor (PR) tests, and Her 2-Neu tests from other health facilities in the province or in the country (personal communication, Dr. Nguyen Dinh Tung, Oncology Department, Hue Central Hospital). In this study, only those tests recorded in the patients’ medical records were included.

Treatment pattern for breast cancer between 2001 and 2006

During the study period, advanced medical equipment and new medications were in limited use in public hospitals in Vietnam. The most common guidelines used in Vietnam for the treatment of breast cancer are reported in Table 1 (personal communication, Dr. Nguyen Dinh Tung, Oncology Department, Hue Central Hospital).


Table 1.  Treatment patterns for breast cancer in Vietnam, 2001–2006
Stage of diagnosis Applied treatment
I Breast surgery, including mastectomy or breast-conserving surgery with/without axillary dissection. Eligible patients received hormone therapy.
II Mastectomy breast surgery with axillary dissection. Bilateral oophorectomy by surgery or radiation, if pre-menopause. Adjuvant radiation was supplemented with either external radiotherapy to the breast, chemotherapy, or both. Eligible patients also received hormone therapy.
III Chemotherapy followed by mastectomy with axillary dissection supplemented with adjuvant chemotherapy. External radiotherapy to the breast was also administered and eligible patients received hormone therapy.
IV Chemotherapy and/or radiation therapy were supplemented with hormone therapy for eligible patients.

Initial treatment

The first, or initial, treatment was implemented after the patient received a positive diagnosis for breast cancer. A range of methods was used, depending on the stage of the breast cancer and the characteristics of the patient. The most common treatment methods were surgery, radiation therapy, chemotherapy, and hormone therapy, either alone or in combination. Surgery involved either a complete mastectomy or breast-conserving surgery combined with axillary lymph node dissection. Bilateral oophorectomy might also have been performed. At the time of the study, radiation was delivered with a cobalt-60 unit. Different chemotherapy regimes were used, with the most common being FAC (a combination of cyclophosphamide, doxorubicin, and 5-fluorouracil), FEC 120 (cyclophosphamide, epirubicin, and 5-fluorouracil), and a combination of paclitaxel–doxorubicin. Tamoxifen was often used for hormone therapy. Before treatment began, patients were assessed by laboratory tests based on the proposed approach and regime. Specialists often requested tests, such as hematograms, chest X-rays, kidney or liver function tests (SGOT, SGPT), CA 15.3, and breast and abdominal ultrasound. The choice of tests varied considerably depending on the doctor and the characteristics of each patient. For patients on chemotherapy, the initial treatment often lasted for up to 9 months, but the duration was less for patients not on chemotherapy.

Follow-up care

Breast cancer treatment has a long course. Normally, patients are required to continue with follow-up care after completing an initial treatment so as to detect local recurrence or metastasis. This type of care was included as ‘follow-up care and supportive treatment’ in this study. During follow-up, outpatient appointments were scheduled every 3 months over the first 2 years and every 6 months in subsequent years. Physical examinations together with laboratory tests, such as hematograms, hepatic ultrasounds, chest X-rays, and CA 15.3, were performed at every out-patient visit. During this time period, most patients were prescribed tamoxifen daily (personal communication, Dr. Nguyen Dinh Tung, Oncology Department, Hue Central Hospital). Since 2006, prescribing tamoxifen has depended on the result of ER tests. In addition, selected patients with signs of recurring tumors received supportive treatment, which might consist of up to six cycles of chemotherapy and/or radiation therapy.

Cost analysis

Costs were divided into three categories: cost of diagnosis, cost of treatment, and cost of follow-up care. Costs of diagnosis comprised the total cost of laboratory tests that patients received to confirm the diagnosis of breast cancer. Treatment costs included surgery, chemotherapy, radiation therapy, hormone therapy and supportive medication, plus inpatient fees. Cost of initial treatment was specified as the combination of the cost of diagnosis and the cost of treatment, which was calculated on the basis of data collected from medical records of individual patients.

Cost of follow-up care included supportive treatment as well as fees for laboratory tests, out-patient visits, and, in some cases, the cost of a dose of tamoxifen. With our focus only on breast cancer, we assumed that all patients were administered the same tests on every outpatient visit. The compliance of patients with a 5-year period of follow-up care was defined as their conformation to out-patient visits and hormone therapy. A questionnaire that provided information about compliance with the course of treatment on the basis of repeated outpatient appointments and doses of medication was designed. These data were obtained through patient interviews (or interviews with their relatives if the patient was deceased) at the time of the study. We assumed that an affirmative response concerning regular outpatient visits and/or compliance with the tamoxifen regime meant that they were in complete compliance with the standard treatment course (until death or the stated end time). If the respondents said ‘sometimes’ or ‘partial compliance’, we set their care pathway to be 50% of the standard follow-up care. When they said ‘no’, patients were defined as non-compliant, and, accordingly, we set their follow-up care to zero. Costs were estimated based on unit cost of tests, outpatient fees, and price of tamoxifen over time. For supportive treatment, we used the data collected from medical records during the 5 years after diagnosis (if records were available). Costs were discounted at an annual rate of 3% as recommended by the World Health Organization (WHO) (18). These costs were then converted to 2010 figures on the basis of the annual inflation index in Vietnam (19). The cost analysis was performed using the following two methods:

  1. Cost analysis by category: For cost categories, aggregate 5-year cost and annual total cost, the mean, standard deviation, the median and cost range were estimated. Values of median for costs were compared with estimates of median regression in the further cost analysis. Costs were presented in US dollars for comparison purposes. The exchange rate used was that in effect on July 15, 2010 (1USD=18544 VND) (20).
  2. Cost analysis by key characteristics of patients: Because of non-normally distributed cost data (Shapiro–Wilk test, p-values <0.001), a quantile regression model was used to analyze the relationship between characteristics of patients and treatment costs for breast cancer. First, key characteristics of the study population, such as stage at diagnosis, health insurance coverage, and age group, were incorporated in a median regression model to determine factors affecting the 5-year total cost for breast cancer. From this analysis, variables with p-value <0.05 were analyzed further in a median regression to estimate the difference in median of cost categories according to their groups. The median difference among groups was presented along with p value and their 95% CI. Differences among groups were considered to be statistically significant when the p value was ≤0.05.

Sensitivity analysis

According to statistics from the Ministry of Health, user fees accounted for 60–70% of all hospital revenues in 2006, the rest were from the government budget and other sources (21). Sensitivity analysis, which added 30–40% to unit costs, presented costs for breast cancer treatment with the government budget supplement.

Results

Study subject characteristics are reported in Table 2. The mean age of patients at the time of diagnosis was 51 years (range 33–75 years). The most frequent age group was that from 40 to 49 years (36.4% of the patients). The study population was evenly divided between urban (51.9%) and rural (48.1%) residences. Slightly more than half of the study population had health insurance (52.7%). At primary diagnosis, the majority of the women had been diagnosed with stage II breast cancer (56.6%). Late-stage diagnosis (stage III and IV) was also common, accounting for 27.1% and 9.3% of the study population, respectively.


Table 2.  Characteristics of patients with breast cancer in Hue Central Hospital
Characteristics Number of patients Percentage
Age at diagnosis (years)    
  Mean (SD) 51 (9.5)  
  Range 33–75  
  <40 16 12.4
  40–49 47 36.4
  50–59 42 32.6
  60–69 17 13.2
  ≥70 7 5.4
Residence    
  Urban 67 51.9
  Rural 62 48.1
Health insurance coverage    
  Yes 68 52.7
  No 61 47.3
Stage of breast cancer at diagnosis    
  I 9 7.0
  II 73 56.6
  III 35 27.1
  IV 12 9.3

More patients with health insurance reported complete or partial compliance than did patients without insurance. The proportion of patients dropping out of treatment was larger among patients without health insurance than among those with health insurance (26.2% vs. 5.9%) (Table 3).


Table 3.  Compliance with breast cancer treatment in relation to health insurance
  Health insurance coverage (%)  
Compliance Yes (95% CI) No (95% CI) Total (95% CI)
Complete compliance 76.5 (64.8–85.2) 68.9 (56.1–79.3) 72.9 (64.4–79.9)
Partial compliance 17.6 (10.2–28.7) 4.9 (1.6–14.4) 11.6 (7.1–18.5)
No compliance 5.9 (2.2–14.8) 26.2 (16.6–38.8) 15.5 (10.2–22.9)
95% CI: 95% confidence interval; p=0.0016 (Pearson's chi-square test).

Figure 1 presents estimates of survival probabilities of up to 5 years for patients with breast cancer by stage at diagnosis. The survival rate was the lowest for late-stage breast cancer, with 43% of patients at stage III and no cases at stage IV surviving as long as 5 years following diagnosis. Patients at stage I and II at the time of primary diagnosis had higher survival rates after 5 years at 78% and 73%, respectively (log-rank test showed p-value <0.001).

Fig 1

Fig. 1.   Kaplan–Meier estimates of 5-year survival probability by stage of breast cancer.

Table 4 displays the cost of the different components of treatment, following the primary diagnosis. Women with breast cancer faced a mean cost estimated at $632.85 per patient over the first 9 months of treatment, but the range was very wide ($11.70–$3955.40). The highest average cost incurred was for chemotherapy, at $476.48 per patient. The cost for surgery was also considerable, at $82.35 per patient regardless of whether the method was a complete mastectomy or breast conservation. The lowest treatment cost was for hormone therapy, at only $4.25. Costs for follow-up care over a 5-year period after primary diagnosis included supportive treatment and other follow-up care as described earlier. The mean total cost for follow-up was estimated at $356.24 per patient, with the greatest proportion of costs for follow-up care ($342.18). Aggregated costs over the 5-year treatment course for breast cancer were on average $975.01 per patient but with a wide range ($11.70 to $3955.40). The annual average cost during the 5 years of treatment was an average of $195 per patient.


Table 4.  Cost estimation per category of breast cancer treatment
  Costs (US dollars)
Category Mean SD Median Range
Diagnosis 16.02 4.93 16.00 6.9–33.20
Treatment        
  Surgery 82.35 31.82 83.00 0.0–235.70
  Chemotherapy 476.48 752.81 379.40 0.0–3772.7
  Radiation therapy 22.87 37.57 9.00 0.0–300.60
  Hormone therapy 4.25 15.20 0.00 0.0–109.60
  Other (supportive medication) 4.50 14.34 0.00 0.0–93.10
Inpatient fee 26.38 12.17 24.60 0.0–88.10
Initial treatment cost 632.85 754.54 509.00 11.7–3955.4
Costs for follow-up care        
  Supportive treatment 14.06 75.08 0.00 0.0–634.10
  Cost for follow-up care 342.18 259.25 409.20 0.0–901.70
Total cost for follow-up care 356.24 260.09 429.40 0.0–901.70
5-Year total cost for treatment course 975.01 730.79 844.20 11.7–3955.4
Annual treatment costs 195.00 146.16 168.80 2.3–791.10
SD, standard deviation; costs are adjusted for inflation to the year 2010; discount rate=3%; exchange rate in July 2010: 1 USD=18,544 VND.

Figure 2 shows that the stage at diagnosis was significant in terms of the 5-year total cost for breast cancer treatment (p=0.001), but there were no significant differences in median total costs related to patient age (p=0.329) or whether or not they had health insurance (p=0.468).

Fig 2
Fig. 2.   The relationship between key characteristics of patients and the total 5-year cost of the course of treatment for breast cancer.

As shown in Table 5, further analysis of the relationship between stage at diagnosis and different cost categories revealed that costs increased with stage at diagnosis for the initial treatment period. The median costs were $128.70, $368.80, $684.10, and $537.90 for stage I, II, III, and IV, respectively. The difference in initial treatment cost among stages was statistically significant (p-values <0.05). By contrast, cost analysis for follow-up care showed that patients with earlier stages at diagnosis faced higher costs because they survived for a longer time period (p-values <0.001). The patients with stage II incurred the highest median cost ($516.50), followed by those with stage I ($409.20) for follow-up care. These costs were lower at stage III ($218). Median cost was estimated at $0 for follow-up care in patients at stage IV because 50% of patients at this stage survived less than a year after diagnosis (Figure 1). However, median costs for aggregated 5-year total cost and annual treatment cost revealed that patients at stage II incurred $333.20 and $66.70 higher costs than those with stage I for 5-year total cost or annual treatment cost, respectively (p-value=0.009) while variance in treatment costs for late stages (stage III and IV) of breast cancer were not statistically significant from stage I (p-values >0.05).


Table 5.  Variance in cost of breast cancer treatment according to stage at diagnosis
Cost category Median Median difference P 95% CI
Initial treatment cost        
  Stage I 128.7 0.0 0.218 −76.86–334.26
  Stage II 368.8 240.1 0.032 20.92–459.28
  Stage III 684.1 519.4 <0.001 286.74–752.06
  Stage IV 537.9 482.2 0.001 208.45–334.26
Follow-up cost        
  Stage I 409.2 0.0 <0.001 383.91–434.49
  Stage II 516.5 107.3 <0.001 80.42–134.18
  Stage III 218.4 −109.8 <0.001 −219.43–− 162.17
  Stage IV 0.0 −409.2 <0.001 −436.21–− 382.19
Aggregated 5-year total cost        
  Stage I 568.6 0.0 <0.001 334.74–802.46
  Stage II 901.8 333.2 0.009 83.86–582.54
  Stage III 816.1 247.5 0.067 −17.19–512.19
  Stage IV 603.4 42.30 0.789 −269.13–353.73
Annual treatment cost        
  Stage I 113.7 0.0 <0.001 66.97–160.43
  Stage II 180.4 66.7 0.009 16.88–116.52
  Stage III 163.2 49.5 0.066 −3.39–102.39
  Stage IV 120.7 8.5 0.787 −53.73–70.73
Cost unit: US dollars.

Data related to government subsidy and other sources included in the estimates of treatment cost for breast cancer at public hospitals are presented in Table 6. When these funding sources were included, corresponding mean and median total costs of 5-year treatment and mean and median annual treatment cost were 40% to nearly 70% higher (corresponding to the support of government of 30% and 40%, respectively).


Table 6.  Estimated costs, including the government subsidy and other sources
  Cost category
  5-Year total cost Annual treatment cost
Coverage of government budget and other sources Mean Median Mean Median
30% 1392.9 1206 278.57 241.14
40% 1625.0 1407 325.00 281.33
Baseline 975.01 844.20 195.00 168.80
Cost unit: US dollars.

Discussion

The results of this study showed that breast cancer was common among young women in central Vietnam during the study period. This is the general profile of breast cancer in developing countries. In developed countries, the majority of breast cancer patients are postmenopausal, 60–70 years old (1, 10). The low coverage of health insurance among the study population was reflective of the study period for Vietnam (22).

The majority of the women in the study population were diagnosed at stage II breast cancer. During the study period, increases in household income due to economic growth and improvements in diagnostic methods for breast cancer (such as the use of ultrasound) provided opportunities for Vietnamese women to contact health facilities and to have their disease detected at an earlier stage than was likely in the past (13).

The costs presented in the study were adjusted to the year 2010 by the growth in the consumer price index and were annually discounted at 3%. The mean total cost of a 5-year course of treatment was estimated at $975.01, with a wide range ($11.70–$3955.40). The compliance with treatment and the type of initial treatment influenced this finding. Some patients refused treatment following diagnosis or did not complete their course of treatment. The majority of those patients that did not complete their treatment course were those not covered by health insurance (Table 3). Establishing a policy of universal health insurance coverage in Vietnam would positively impact the current lack of affordable access to appropriate treatment for chronic diseases such as breast cancer. Nevertheless, the costs determined by this study were much lower than those reported for developed countries. In France, for example, the mean medical cost for a 5-year treatment period for breast cancer was $10,744 (23). Groot et al. estimated the 10-year total cost of treatment per patient with breast cancer based on data retrieved from the WHO-CHOICE database in Africa, Asia, and the Americas in 2000. They reported lower estimates of $602, $356, and $8,530 for Africa, Asia, and North America, respectively (24). These comparisons are similar to a review by Radice et al. in which the cost of breast cancer treatment in developing regions was considered less than or equal to 5% of that in the developed world (3). Comparisons among the wide range of cost estimates for breast cancer treatment and generalizations drawn from economic studies on the disease are made difficult by the different characteristics and patient populations of each country. The diversified unit costs for resource use in different countries could explain the different findings. For instance, in the United Kingdom in 2007, the cost of breast cancer surgery ranged from £1,261 for conservative surgery to £2,073 for a mastectomy, compared to the average cost of $82.35 (about £55 in 2010) for breast cancer surgery that we found in our study (25). In Vietnam, public hospital charges did not measure the full cost of health care resource usage. Unit costs included the price of medications and materials used in the course of treatment but only a portion of those resources that were subsidized by government policy, such as the hospital facility and clinical staff (15, 16). Sensitivity analysis showed that including the government subsidy increased cost estimates by 40 to 70% (Table 6). However, even if user fees and government subsidies were combined, hospital charges were still underestimated. Remuneration of health staff and capital depreciation have not been adequately estimated (26). Unit costs in Vietnamese public hospitals and hospital fees are, therefore, lower than the real cost of the resources used. In a cost analysis of health services in Vietnam, Flessa et al. (2004) determined that the unit cost of an operation such as breast cancer surgery at a central hospital was $175.89, double the cost of our findings (27). In addition, at the time of our study, advanced treatment guidelines were not yet available in Vietnam.

Initial treatment costs were found to be the most expensive component of total costs, accounting for $632.85; these costs represented 65% of the total cost, compared to 63–73.3% in previous studies (6, 7, 23). Chemotherapy costs made up the highest proportion of the initial treatment-attributable costs. In 2005, Oestreicher et al. estimated the cost of chemotherapy for US women with early-stage breast carcinoma to be $23,019, which is 50 times greater than our estimate of $476.48 for Vietnam in 2010 (28). The factors that may contribute to high chemotherapy costs are the types of chemotherapy agents used and the cost of supportive care agents (2). The variety of chemotherapy regimens could explain the wide range of estimated costs for initial treatment as well as the total 5-year treatment course. The regime with paclitaxel–doxorubicin was found to be the most expensive treatment option for chemotherapy over the study period. Many studies have compared the cost-effectiveness of alternative chemotherapy regimes for the treatment of breast cancer. For example, the analysis reported by Mittmann (2010) showed that a protocol with docetaxel offered improved life expectancy but at a higher cost compared with fluorouracin–adriamycin–cyclophosphamid (FAC) (29). According to the experience of the oncologists in HCH (personal communication), the use of expensive chemotherapy regimens depended on the patients’ ability to pay for them. Research on the economic evaluation of different breast cancer chemotherapy regimes should be conducted in the Vietnamese context. The result will help health care providers as well as patients in choosing an affordable and effective treatment method. The high proportion of chemotherapy costs is also a reason as to why initial treatment costs were more for patients diagnosed at stage II and higher; chemotherapy is recommended for most of those patients (see Table 1). In fact, the costs of chemotherapy in our study exceeded the total cost of follow-up care over the 5 years after diagnosis ($476.48 vs. $356.24). Because follow-up treatment for breast cancer in the years after the initial treatment was relatively simple, as described in the method section, the related costs were estimated to be small and relatively stable, as was found in other studies (24, 30). For patients diagnosed with stage I breast cancer, the initial treatment costs were very low, but the follow-up care accounted for a higher proportion of the total cost of treatment. The opposite was true for the patients diagnosed with stage IV. For late-stage breast cancers, the treatment was ineffective. Our study revealed that patients diagnosed at a late stage incurred the same costs as those diagnosed at an early stage but had lower survival times. Early detection of breast cancer may not only increase life expectancy but could also result in resource savings for health care (2, 3). Presently, a pilot screening program for breast cancer has been introduced in some regions of Vietnam. An economic evaluation is necessary before the program will be available nationwide.

Because of underestimation of charges in public hospitals, the annual direct medical cost for breast cancer treatment in this study amounted to about 18% of gross national income (GNI) per capita in Vietnam in 2010 ($195 vs. $1,100) (31). A review by Pisu et al. revealed that out-of-pocket costs for direct medical care were a substantial burden for low-income breast cancer survivors, whose expenses for the disease within 1 year after diagnosis amounted to 75% of their total annual income, compared with only 8% for breast cancer survivors in the highest income group (32). Out-of-pocket costs are the main obstacle to medical treatment, especially in case of diseases with a long natural history (such as breast cancer). Indeed, our study found that a higher number of patients in the group without health insurance coverage dropped out of their treatment regime. Universal health insurance coverage is not yet a reality in Vietnam but should be given more attention, especially since public hospital charges are expected to increase in the near future. The government should have a policy to support cancer patients for whom the cost of illness exceeds their ability to pay or even to co-pay for health insurance. In addition, if a network of primary health care were to be established throughout the country, alternatives such as home care and community care should be promoted to provide health care services to patients who require long-term care, following an initial hospital stay (such as breast cancer patients). The shift to home care settings may improve the compliance with treatment and reduce out-of-pocket costs for patients in Vietnam, where the access to health facilities for cancer treatment has been limited (33).

A number of factors should be considered when interpreting the findings of this study. The data were collected over the period 2001 to 2006 and do not reflect current utilization of advanced treatment methods and new medications for breast cancer treatment. The analysis was limited to costs of primary breast cancer cases, excluding recurrent cases. Both of these factors could lead to an underestimation of the costs. The exclusion of the governmental subsidy and other resources in our cost estimates meant that our estimates did not represent the ‘complete’ resource costs incurred for the treatment of breast cancer in Vietnam, although the estimates do reflect the full costs borne by health care payers. Many changes in the socioeconomic structure of Vietnam occurred during the study period and continue to the present. These changes in the socioeconomic environment might limit the ability to generalize from these study results. In addition, precise data were not available for much of the follow-up period for care; these costs were mainly estimated based on the patient's (or relative's) recall of at least 5 years and therefore are subject to potential bias. Although efforts were made to enroll a suitable number of cases in the analysis, the low incidence rate of breast cancer in Thua Thien Hue province combined with limitations of medical record preservation before 2008 resulted in a small sample size. This affected the opportunity to identify significant differences in cost comparisons among various groups of patients. The estimated costs for breast cancer treatment might not be representative of other main public hospitals in Hanoi and Ho Chi Minh City or of private hospitals in Vietnam, where unit costs may differ from those in our study (15, 16), thereby limiting the ability to generalize our study findings. Despite these limitations, the cost estimates in this article provide the first piece of evidence regarding the cost of breast cancer treatment in Vietnam. These findings reflect the financial burden on health care payers at public hospitals. They will contribute important information to cost-effectiveness analysis of interventions for breast cancer and will help decision-makers engaged in health system planning and resource allocation.

Conclusion

The direct medical costs of a 5-year course of treatment for primary breast cancer in central Vietnam are much lower than in developed countries. The exclusion of government subsidies and other resources lowered the total costs included in our analysis. However, the long treatment course significantly influenced out-of-pocket payments by patients without health insurance. Having health insurance increased patients’ compliance with treatment because the ability to pay played a major role in treatment uptake. The initial treatment, especially chemotherapy, accounted for the largest part of total costs though the range in costs was wide. There is no significant difference in 5-year total cost with regard to age at diagnosis, health insurance coverage, and between early- and late-stage breast cancer patients in the study. Patients diagnosed with late-stage breast cancer incurred higher costs for initial treatment than those diagnosed at early stages, while their survival time was shorter. Facing these challenges, early detection of breast cancer through screening programs, access to relevant treatment, and an increase in health insurance coverage along with other financial supports to chronic patients should be implemented to improve access to care and the prognosis of breast cancer patients in Vietnam.

Acknowledgments

We thank Prof. Pamela Wright, Director of Medical Committee Netherland Vietnam for editorial assistance. We also thank Assoc. Prof. Bandit Thinkhamrop for his advice on statistical analysis.

Ethical approval

Ethical approval for primary and secondary data collection was obtained from the University of Khon Kaen, Thailand (where the study was designed as part of a doctoral study program). In addition, approvals for implementing the study at the various study sites were obtained from the Provincial Health Service of Thua Thien Hue province.

Conflict of interest and funding

The authors have no potential conflict of interest. The research was funded by the Vietnam–Netherland Project at Hue College of Medicine and Pharmacy, Vietnam, and Graduate School of Khon Kaen University, Thailand.

References

  1. Peter B, Bernard L., editors. World Cancer Report 2008 [online]. WHO press. The International Agency for Research on Cancer (IARC). Available from: www.iarc.fr/en/publications/pdfs_online/wrc/2008/wcr_2008.pdf. [cited 17 June 2011].
  2. Anderson BO, Yip CH, Smith RA, Shyyan R, Sener SF, Eniu A, et al. Guideline implementation for breast healthcare in low-income and middle-income countries. Overview of the breast health global initiative global summit 2007. Cancer 2008; 113(Suppl. 8): 2221–43. PubMed Abstract | Publisher Full Text
  3. Radice D, Redaelli A. Breast cancer management: quality-of-life and cost considerations [abstract]. Pharmacoeconomics 2003; 21: 383–96. PubMed Abstract | Publisher Full Text
  4. Pallis A, Tsiantou V, Simou E, Maniadakis N. Pharmacoeconomic considerations in the treatment of breast cancer. Clinicoecon Outcomes Res 2010; 2: 47–61. PubMed Abstract | PubMed Central Full Text
  5. Chu PC, Hwang JS, Wang JD, Chang YY. Estimation of the financial burden to the national health insurance for patients with major cancers in Taiwan [abstract]. J Formos Med Assoc 2008; 107: 54–63. PubMed Abstract | Publisher Full Text
  6. Barron JJ, Quimbo R, Nikam PT, Amonkar MM. Assessing the economic burden of breast cancer in a US managed care population [abstract]. Breast Cancer Res Treat 2008; 109: 367–77. PubMed Abstract | Publisher Full Text
  7. Will BP, Berthelot JM, Le Petit C, Tomiak EM, Verma S, Evans WK. Estimates of the lifetime costs of breast cancer treatment in Canada [abtract]. Eur J Cancer 2000; 36: 724–35. PubMed Abstract | Publisher Full Text
  8. Legorreta AP, Brooks RJ, Leibowitz AN, Solin LJ. Cost of breast cancer treatment. A 4-year longitudinal study. Arch Intern Med 1996; 156: 2197–201. Publisher Full Text
  9. Duc NB. Epidemiology and program of control and prevention for cancer: preliminary report of National Cancer Project period 2008–2010. Vietnamese J Oncol 2010; (1): 21–6. Vietnamese.
  10. Anyanwu SN. Temporal trends in breast cancer presentation in the third world. J Exp Clin Cancer Res 2008; 27: 17. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  11. World Health Organization (2010). Breast cancer: prevention and control. Available from: http://www.who.int/cancer/detection/breastcancer/en/index.html [cited 12 September 2010].
  12. Ngoan le T, Lua NT, Hang LT. Cancer mortality pattern in Viet Nam. Asian Pac J Cancer Prev 2007; 8: 535–8. PubMed Abstract
  13. Anh PT, Duc NB. The situation with cancer control in Vietnam. Jpn J Clin Oncol 2002; 32(Suppl. 1): S92–7. PubMed Abstract | Publisher Full Text
  14. World Health Organization (2012). International classification of diseases 10th revision. ICD-10 version: 2010. Available from: http://apps.who.int/classification/icd10/browse/2010/en [cited 16 January 2012].
  15. Regulation of Vietnamese Government (1994). Partially collection of hospital fee. Decree No 95/CP. Available from: http://www.hspi.org.vn/vcl/Thong-tu-Lien-Bo-y-te---Tai-chinh---Lao-dong-TBXH--Ban-Vat-gia-Chinh-phu-so-14TTLB-ngay-30-thang-9-nam-1995-Huong-dan-thuc-hien-viec-thu-mot-phan-vien-phi-t96–1037.html [cited 27 August 1994].
  16. Regulation of Vietnamese Government. Revised some points of interministerial circular number 14/TTLB dated 30 September 1995. Circular No 11/TTLB dated 19 September 1997.
  17. TNM Classification Help (2012). Breast tumours. Available from: http://cancerstaging.blogspot.com/2005/02/breast-tumours.html [cited 2012 March 21].
  18. Tan-Torres T, Baltussen R, Adam T, Hutubessy R, Acharya A, Evans BD, et al., editors. Making choices in health: WHO guide to cost-effectiveness analysis. Geneva: World Health Organization, 2003.
  19. Vietnamese General Statistics Office (2011). Statistics of consumer price index by year. Available from: http://www.gso.gv.vn. [cited 20 November 2011].
  20. Ministry of Finance. Rate of Exchange in July 2010. Available from: http://www.mof.gov.vn/portal/page/portal/mof_vn/1641775?p_folder_id=3817325&p_recurrent_news_id=14280946 [cited 20 February 2012].
  21. Ministry of Health (2006). Report of health care activities in 2006. Summation Conference for Hospital Activities in Vietnam, Hanoi: Department of Therapy, Ministry of Health.
  22. Long TG. Vietnam: Social health insurance: Current issues and policy recommendations. The ILO's Asia-Pacific Regional High-Level Meeting on Socially-Inclusive Strategies to Extend Social Security Coverage. Bangkok: the International Labor Organization (ILO); 2008.
  23. Lilliu H, Stevens D, Brun C, Morel J, Pen CL, Bonastre J, et al. Cost of treatment and follow up of breast cancer: a retrospective assessment in a comprehensive cancer centre. Bull Cancer 2002; 89: 635–42. PubMed Abstract
  24. Groot MT, Baltussen R, Uyl-de Groot CA, Anderson BO, Hortobagyi GN. Costs and health effects of breast cancer interventions in epidemiologically different regions of Africa, North America, and Asia. Breast J 2006; 12(Suppl. 1): S81–90. PubMed Abstract | Publisher Full Text
  25. Karnon J, Kerr GR, Jack W, Papo NL, Cameron DA. Health care costs for the treatment of breast cancer recurrent events: estimates from a UK-based patient-level analysis. Breast J Can 2007; 97: 479–85.
  26. Health Services of Ho Chi Minh City (2006). The implementation of policy on health socialization in Ho Chi Minh City. Document No 4422/SYT-QLDVYT. Available from: http://www.medinet.hochiminhcity.gov.vn/data/news/2006/8/3887/thxhh.htm [cited 2012 May 25].
  27. Flessa S, Dung NT. Costing of services of Vietnamese hospitals: identifying costs in one central, two provincial and two district hospitals using a standard methodology. Int J Health Plann Manage 2004; 19: 63–77. PubMed Abstract | Publisher Full Text
  28. Oestreicher N, Ramsey SD, McCune JS, Linden HM, Veenstra DL. The cost of adjuvant chemotherapy in patients with early-stage breast carcinoma. Cancer 2005; 104: 2054–62. PubMed Abstract | Publisher Full Text
  29. Mittmann N, Verma S, Koo M, Alloul K, Trudeau M. Cost effectiveness of TAC versus FAC in adjuvant treatment of node-positive breast cancer. Curr Oncol 2010; 17: 7–16. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  30. Warren JL, Brown ML, Fay MP, Schussler N, Potosky AL, Riley GF. Costs of treatment for elderly women with early-stage breast cancer in fee-for-service settings. J Clin Oncol 2002; 20: 307–16. PubMed Abstract | Publisher Full Text
  31. World Bank (2012). World development indicators database 1 July 2011. Available from: http://siteresources.worldbank.org/DATASTATISTICS/Resources/GNIPC.pdf [cited 2012 February 11].
  32. Pisu M, Azuero A, McNees P, Burkhardt J, Benz R, Meneses K. The out of pocket cost of breast cancer survivors: a review. J Cancer Surviv 2010; 4: 202–9. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  33. Vietnamscope. Situation: cancer situation [updated September 21, 2010]; Available from: http://www.vietnamscope.org/index.php?option=com_content&view=article&id=99&Itemid=139 [cited May 25 2012].

*Wongsa Laohasiriwong
Faculty of Public Health
Khon Kaen University
Khon Kaen, 40002
Thailand
Tel: +66 897 121 455
Fax: +66 43347058
Email: drwongsa@gmail.com

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Knowledge of the health consequences of tobacco smoking: a cross-sectional survey of Vietnamese adults

Dao Thi Minh An1*, Hoang Van Minh1, Le Thi Huong1, Kim Bao Giang1, Le Thi Thanh Xuan1, Phan Thi Hai2, Pham Quynh Nga3 and Jason Hsia4

1Department of Epidemiology, Institute for Preventive Medicine and Public Health, Hanoi Medical University, Hanoi, Vietnam; 2The Vietnam Steering Committee on Smoking and Health, Hanoi, Vietnam; 3World Health Organization Office in Viet Nam, Hanoi, Vietnam; 4Center for Disease Control and Prevention, Atlanta, USA

Abstract

Background: Although substantial efforts have been made to curtail smoking in Vietnam, the 2010 Global Adult Tobacco Survey (GATS) revealed that the proportion of male adults currently smoking remains high at 47.4%.

Objectives: To determine the level of, and characteristics associated with, knowledge of the health consequences of smoking among Vietnamese adults.

Design: GATS 2010 was designed to survey a nationally representative sample of Vietnamese men and women aged 15 and older drawn from 11,142 households using a two-stage sampling design. Descriptive statistics were calculated and multivariate logistic regression was used to examine associations between postulated exposure factors (age, education, access to information, ethnic group etc.) and knowledge on health risks.

Results: General knowledge on the health risks of active smoking (AS) and exposure to second hand smoke (SHS) was good (90% and 83%, respectively). However, knowledge on specific diseases related to tobacco smoking (stroke, heart attack, and lung cancer) appeared to be lower (51.5%). Non-smokers had a significantly higher likelihood of demonstrating better knowledge on health risks related to AS (OR 1.6) and SHS (OR 1.7) than smokers. Adults with secondary education, college education or above also had significantly higher levels knowledge of AS/SHS health risks than those with primary education (AS: ORs 1.6, 1.7, and 1.9, respectively, and SHS: ORs 2.4, 3.9, and 5.7 respectively). Increasing age was positively associated with knowledge of the health consequences of SHS, and access to information was significantly associated with knowledge of AS/SHS health risks (ORs 2.3 and 1.9 respectively). Otherwise, non-Kinh ethnic groups had significantly less knowledge on health risks of AS/SHS than Kinh ethnic groups.

Conclusions: It may be necessary to target tobacco prevention programs to specific subgroups including current smokers, adults with low education, non-Kinh ethnics in order to increase their knowledge on health risks of smoking. Comprehensive messages and/or images about specific diseases related to AS/SHS should be conveyed using of different channels and modes specific to local cultures to increase knowledge on smoking health consequences for general population.

Keywords: knowledge; smoking; health consequences; global adult tobacco survey; Vietnam

Received: 7 May 2012; Revised: 11 December 2012; Accepted: 17 December 2012; Published: 31 January 2013

Glob Health Action 2013. © 2013 Dao Thi Minh An et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 18707 - http://dx.doi.org/10.3402/gha.v6i0.18707

 

Over the past 10 years, Vietnam had made substantial efforts on tobacco control. In 2005, Vietnam ratified the Framework Convention on Tobacco Control (FCTC), and in August 2009, the Prime Minister's Decision No. 1315/QĐ-TTg approved a plan to implement FCTC. The Ministry of Health, Ministry of Education & Training, and Ministry of Transport also issued official directives for the implementation of a smoke-free policy. At the same time, levels of smoking remain high in Vietnam. In 2001–02, 69.1% of men smoked cigarettes; 23.2% of men smoked water pipe tobacco. In addition, 63% of households had at least one smoker, and 71% of children under age 5 lived in households with at least one smoker (1). In 2003, nearly 60% of school-age youth reported being exposed to secondhand smoke at home (2). The prevalence of cigarette smoking among students was 3.3% overall, while that among male students was 5.9% (3). The 2010 Global Adult Tobacco Survey (GATS) in Vietnam showed that smoking prevalence among adult males was 47.4%, and in a survey conducted in hospitals and public places evidence revealed that smoking was common (4, 5).

Policy Recommendations

  • IEC activities in Vietnam should be designed specifically to focus on specific target populations, given their smoking status and demographic characteristics.
  • Information conveyed to smokers should be comprehensive, including not only general messages of smoking-health harms but also specific health risks of active, secondhand smoking, and disease burdens of smoking based on the real data of specific diseases related to smoking which Vietnamese people suffered from.
  • To improve accessibility to information of health consequences of smoking, messages that smoking is harmful should be conveyed appropriately, frequently, and efficiently to different targeted populations by different channels/modes of communication which are mostly preferred or mostly accessible by each targeted population in the Vietnamese setting.

Knowledge is defined by the Oxford English Dictionary as expertise and skills acquired by a person through experience or education; while perception is the process by which humans interpret and organize sensation to produce a meaningful experience of the world (6). It is now well established that if people's perceptions of the commonality and acceptability of a behavior can be adjusted, their inclination to engage in that behavior may be influenced. For example, the more common and acceptable young people think smoking is among their immediate peers, their family group and society as a whole, the more likely they are to take up the habit (7). Conversely, smoking uptake can potentially be reduced if pro-smoking norms are challenged and anti-smoking norms are strengthened. Gerards Hastings in his book titled ‘Social Marketing – Why should the devil have all the best’ wrote that ‘Normative education or de-normalization programs, therefore, can correct erroneous perceptions of the prevalence and acceptability of drug and alcohol use and establish conservative group norms that are postulated to operate through lowering expectations about prevalence and acceptability of use and the reduced availability of substances in peer-oriented social settings’.

Therefore, Article 4 of FCTC of the World Health Organization (WHO) stated that every person should be informed of the health consequences, addictive nature, and mortal threat posed by tobacco consumption and exposure to tobacco smoke. As a result, several countries have attempted to educate their populations about the health consequences of smoking (8, 9). The Vietnam Steering Committee on Smoking and Health (VINACOSH) has made significant efforts to convey messages of tobacco control to Vietnamese population during the past 10 years by several Information, Education, and Communication (IEC) activities in cooperation with the International Development Establishment (IDE). These activities consisted of celebrity television interviews; tobacco control communication activities in combination with a trans – Viet bicycle riding tour from Hanoi to Ho Chi Minh City; hosting training IEC workshops on designing a smokers’ behavior influencing project, called ‘Keep tobacco smoke away from your wife and kids’ in Hai Phong City; and improving public awareness on the negative effects of tobacco on human health and the economy. VINACOSH has been cooperating with IDE in developing tobacco and secondhand smoke control IEC materials, applying social marketing approaches in designing IEC materials, and selecting relevant communication channels (10). Although many efforts have been made under the IEC programs in Vietnam, there is still a need for additional research to understand the relationship of knowledge of smoking-health risks and smoking behavior in Vietnam. The purpose of this study was to identify the level of knowledge of health consequences of tobacco smoking and its associated factors to inform IEC programs on reducing tobacco smoking.

Methods

Selection and description of participants

Data used in this article were obtained from the 2010 GATS Vietnam, which was part of the multi-countries household survey of the ongoing Global Tobacco Surveillance System (GTSS) (11). The GATS in Vietnam was designed as a nationally representative survey of all non-institutionalized men and women aged 15 and older, with their primary residence in Vietnam.

A total of 11,142 households were selected using a two-phase sampling design analogous to a three-stage stratified cluster sampling. In 2009, the General Statistics Office (GSO) in Vietnam conducted a population and housing census. Meanwhile, GSO prepared a 15% master sample to serve as a future national survey-sampling frame. The 15% master sample contains a subset of enumeration areas (EAs) that consist of 15% of the population in Vietnam stratified into three groups. The first group consists of 132 districts, towns, or cities of provinces. The second group consists of 294 plain and coastal districts. The third group consists of 256 mountainous and island districts. The GATS sample was drawn from the 15% master sample after further stratification of the three groups into urban and rural areas (six strata in total).

At the first stage of sampling, the primary sampling unit (PSU) was EAs. The sampling frame was a list of the EAs, from the 15% master sample, with the number of households as well as identifiable information, administered by the GSO Vietnam in 2009 from the census. For each of the six strata, the designated number of EAs was selected. A selection probability proportional to size (PPS) sampling method was used, where the size was the selection probability of an EA using PPS sampling from the entire target population divided by the selection probability of an EA for the master sample.

At the second stage of sampling, 18 households from the selected urban EAs and 16 households from the selected rural EAs were chosen using simple systematic random sampling. One eligible household member from each selected household was then randomly chosen for an interview.

Note that the current design and the design where EAs were sampled directly universally were analogous. The selection probability of an eligible individual was calculated as a product of selection probability for each stage. The sampling base weight for an eligible individual was the inverse of the selection probability shown above.

Data collection

Handheld computers (iPAQ) were used for collecting data. Interviewers and supervisors from GSO conducted fieldwork, under the co-supervision of the WHO in Vietnam and Hanoi Medical University. The fieldwork lasted from March 22, 2010, to May 13, 2010, in all 63 provinces of Vietnam. The interviewers and supervisors were experienced, trained in using computers and handheld (iPAQ) devices, and had previous experience working with local authorities, which is key to minimizing non-response rates. A case file containing addresses and names of the households assigned to the interviewer was preloaded in the iPAQ prior to the fieldwork. The data collectors went to the residences of the respondents and met the head of the household to acquire general information about the number of eligible individuals in the household. This number was entered into the handheld computer and one individual was automatically selected to be interviewed. All responses were entered in the iPAQ.

Variables and definition used in this article

Dependent variables

  1. Knowledge of specific diseases of active smoking defined by one who answered ‘yes’ to all four situations (smoking causes serious illness, stroke, heart attack, and lung cancer)
  2. Knowledge of health consequences of secondhand smoke defined by one answered ‘yes’ to question ‘breathing other people's smoke cause serious illness in non-smokers’.

Independent variables

  1. Smoking status: current smokers and current non-smokers
  2. Demographic variables: age; sex; education, quintile of household income which is based on the current study; area (urban/rural); region (ecological area), ethnic group
  3. Access to positive information channel in the last 30 days (Answer ‘yes’ to one of the following: information about health consequences of smoking, encouragement to quit, or health warnings on cigarette packages)
  4. Access to negative information channel in the last 30 days (Answer ‘yes’ to one of the following: cigarette advertisement through media, cigarette advertisement events, or cigarette promotion).

Statistics

Descriptive and statistical analyses with percentages and 95% confidence interval (CI) were calculated using Stata 10 software (Stata Corporation). The relationship between demographic variables (sex, areas, education, age group, quintiles of income, regions), smoking status, and levels of access to information and knowledge of health consequences were conducted. Multivariate logistic regression modeling was performed to identify what variables associated with knowledge of health consequences of active smoking and passive smoking in which variables of demographic characteristics, smoking status, rule of ‘no smoking at home’ and levels of access to information were screened for bivariate association and then all entered into the model as the independent factors. Backward elimination was used to remove ones that were not statistically significant (p>0.05). The odds ratio (OR) with 95% CI was used. Sampling weights were used in all of the computations.

Results

The household response rate was 96.9% and was a little higher in rural areas than that in urban areas (97.5 and 96.5%, respectively). The individual response rate was 95.7% and was also a little higher in rural areas than that in urban areas (96.3 and 95.0%, respectively). Overall, 0.6% of the households and 0.6% of the selected individuals refused to respond to the survey. The total response rate was 92.7% (93.9% in rural areas and 91.7% in urban sites).

Among 9,925 completed interviews of adults aged 15 and above, two-thirds were living in rural areas. People aged 25–44 made up the largest proportion (41.9%). The educational level that predominated was secondary school (52.5%), followed by primary or less (26%), while college degree or above was only 7.2% of the total. The main occupation of the study population was farmers (49.6%), followed by service/sales (19.2%) and production/driving (12.9%). By ethnicity, 84.5% were Kinh people, and the remaining 15.5% belonged to other ethnic groups. By marital status, 67.7% of the total was married, 26.2% were single, and the remaining 6.2% were separated/divorced/widowed (Table 1).


Table 1.  Distribution of adults ≥15 years by selected demographic characteristics – Viet Nam GATS, 2010
  Weighted    
Demographic characteristics Percentage (95% CI*) Number of adults (in thousands) Unweighted number of adults
Overall 100 64,321 9,925
Gender      
  Male 48.6 (47.3–49.9) 31,259 4,356
   Female 51.4 (50.1–52.7) 33,063 5,569

     
Age (years)      
  15–24 25.9 (24.6–27.2) 16,637 1,656
  25–44 41.9 (40.6–43.2) 26,944 4,251
  45–64 23.4 (22.4–24.5) 15,065 2,886
  65+ 8.8 (8.2–9.5) 5,675 1,132

     
Residence      
  Urban 30.7 (30.0–31.4) 19,725 4,958
   Rural 69.3 (68.6–70.0) 44,596 4,967

     
Education level      
  Primary or less 26.0 (24.2–27.8) 12,377 2,034
  Lower secondary 52.5 (50.8–54.3) 25,031 3,981
  Upper secondary 14.3 (13.1–15.5) 6,794 1,023
  College or above 7.2 (6.6–7.9) 3,447 1,227
*95% confidence interval.
N=9,925 individuals from 656 PSUs of 6 strata.

Generally, the percentage of adults who agreed that active smoking and exposure to secondhand smoke causes serious illness were at high levels (93.4 and 83.8%, respectively) but differed by demographic characteristics. Regarding knowledge of harmful health effects of active and passive smoking, adults living in urban areas were more knowledgeable than those living in rural areas (97.1 vs 92.1% and 90.8 vs 81.3%, respectively); Kinh ethnic had greater knowledge than non-Kinh ethnic (96.8 vs 84.3% and 88.2 vs 72.0%, respectively). The respondents who had higher income and education were more likely to have better knowledge than those who had not. There were no differences for knowledge of health damage by sex, age group, and ecological region (Table 2).


Table 2.  Knowledge of health consequences of tobacco smoking by demographic characteristics
  Active smoking causes serious illness Secondhand smoking causes serious illness
Demographic characteristics Percentage (95% CI*) Percentage (95% CI*)
Over all 93.4 (91.0–95.2) 83.8 (81.3–86.1)
Sex    
  Male 93.7 (91.9–95.1) 84.3 (81.8–86.4)
  Female 93.2 (89.9–95.4) 83.4 (80.3–86.1)

   
Residence    
  Urban 97.1 (96.4–97.6) 90.8 (89.7–91.8)
  Rural 92.1 (88.8–94.5) 81.3 (78.0–84.2)

   
Ethnic group    
  Kinh 96.8 (96.1–97.3) 88.2 (87.0–89.2)
  Other 84.3 (76.6–89.8) 72.0 (64.7–78.2)

   
Ecological regions    
  Red River Delta 97.4 (96.6–98.0) 91.5 (90.2–92.7)
  Northern midland and mountain 96.1 (94.6–97.2) 86.2 (82.9–88.9)
  North Central area and Central coastal 97.1 (95.0–98.3) 92.6 (90.0–94.5)
  Central highlands 98.0 (94.5–99.3) 93.0 (84.6–97.0)
  South East 95.1 (93.8–96.2) 82.5 (80.0–84.8)
  Mekong River Delta 88.6 (82.1–92.9) 78.9 (72.5–84.1)

   
Age groups    
  15–24 93.9 (90.6–96.0) 89.7 (86.1–92.4)
  25–34 93.7 (89.0–96.5) 85.0 (79.6–89.1)
  35–44 94.0 (89.3–96.7) 84.6 (81.0–87.6)
  45–54 95.6 (94.1–96.7) 87.0 (84.8–88.9)
  55–64 93.5 (90.1–95.8) 79.1 (75.1–82.6)
  >64 87.1 (83.7–89.9) 70.0 (66.1–73.6)

   
Incomes    
  Quintile 1 85.3 (78.8–90.1) 69.6 (64.1–74.6)
  Quintile 2 96.0 (94.5–97.1) 86.7 (84.1–89.0)
  Quintile 3 96.9 (95.6–97.8) 88.9 (86.5–90.8)
  Quintile 4 97.9 (97.0–98.6) 92.1 (90.6–93.4)
  Quintile 5 97.7 (96.8–98.3) 94.8 (93.5–95.8)

   
Education    
  Primary or less 84.5 (79.0–88.7) 64.6 (60.3–68.7)
  Lower secondary 96.8 (95.4–97.8) 88.7 (86.6–90.5)
  Upper secondary 98.2 (96.9–99.0) 95.0 (93.1–96.3)
  College and/or above 99.1 (98.4–99.5) 97.1 (95.8–97.9)
*95% confidence interval.
N=9,919 individuals from 656 PSUs of 6 strata.

However, only 51.5% of interviewees answered correctly to all three specific health consequences (stroke, heart attack, and lung cancer). The most common health consequence was lung cancer (95.8%), while strokes and heart attacks were found to be much lower (67.6 and 60.9%, respectively). Of interest, current smokers displayed significantly lower knowledge of health risks of active smoking than current non-smokers, for example, smoking causes serious illness (83.3 vs 95.1%), stroke (59.4 vs 70.3%), heart attack (54.2 vs. 63.1%), lung cancer (93.0 vs 96.7%), all three main consequences (43.1 vs 54.3%), and secondhand smoke (77.3 vs 86.0%) (Table 3).


Table 3.  Knowledge of health consequences of tobacco smoking by smoking status
Knowledge of health consequences of tobacco smoking Current smokers Current non-smokers Total
Smoking causes Percentage (95% CI*) Percentage (95% CI*) Percentage (95% CI*)
  Serious illness 88.3 (82.8–92.2) 95.1 (93.6–96.3) 93.4 (91.0–95.2)
  Stroke 59.4 (54.0–64.6) 70.3 (67.8–72.7) 67.6 (64.9–70.3)
  Heart attack 54.2 (49.9–58.5) 63.1 (60.8–65.4) 60.9 (58.5–63.4)
  Lung cancer 93.0 (89.9–95.2) 96.7 (95.8–97.3) 95.8 (94.6–96.7)
  Stroke, heart attack, and lung cancer 43.1 (38.0–48.4) 54.3 (52.0–56.6) 51.5 (48.8–54.2)
Breathing other people's smoke cause serious illness in non-smokers 77.3 (71.3–82.4) 86.0 (84.2–87.7) 83.8 (81.3–86.1)
*95% confidence interval.
N=9,919 individuals from 656 PSUs of 6 strata.

There were significant differences in the knowledge of health consequences for those who have access to positive information and those who did not, with those having access to information having more knowledge of health consequences of active smoking than those who did not, for example, knowledge of: serious illness 96.2 vs 76.3%, stroke 71.8 vs 41.1%, heart attack 64.5 vs 37.9%, lung cancer 97.7 vs 83.4%, and all three main health consequences 55.5 vs 27.0%. This relationship held for individuals having access to information about secondhand smoke exposure; 87.9% of individuals with access knew that breathing other people's smoke can cause serious illness in non-smokers, while among adults without access only 59.0% knew about the consequences. However, as demonstrated in Table 4, there were not many other differences between the groups.


Table 4.  Knowledge of health consequences of tobacco smoking by different channels of accessing to information
  Access to positive information** Access to negative information***
Knowledge of health consequences of tobacco smoking Percentage (95% CI*) Percentage (95% CI*)
Smoking causes No Yes No Yes
  Serious illness 76.3 (67.5–83.3) 96.2 (94.9–97.2) 93.2 (90.9–94.9) 94.7 (90.1–97.2)
  Stroke 41.1 (34.5–48.0) 71.8 (69.7–73.7) 66.5 (63.5–69.3) 75.1 (70.0–79.7)
  Heart attack 37.9 (32.0–44.3) 64.5 (62.5–66.4) 59.6 (56.9–62.2) 69.6 (64.7–74.0)
  Lung cancer 83.4 (77.7–87.9) 97.7 (97.1–98.1) 95.4 (94.1–96.5) 97.8 (96.6–98.6)
  Stroke, heart attack, and lung cancer 27.0 (21.9–32.7) 55.5 (53.4–57.6) 50.1 (47.3–52.9) 60.0 (55.1–64.7)
Breathing other people's smoke cause serious illness in non-smokers 59.0 (52.0–65.6) 87.9 (86.0–89.6) 83.4 (80.9–85.6) 86.6 (82.2–90.0)
*95% confidence interval.
**Access to positive information channel in the last 30 days (information about health consequences of smoking or encouragement to quit; health warnings on cigarette packages).
***Access to negative information channel in the last 30 days (cigarette advertisement through media, cigarette advertisement events, cigarette promotion).
N=9,919 individuals from 656 PSUs of 6 strata.

Education level was reported only among respondents aged 25 years with the assumption that at age of 25 and above, people have completed their education and have acquired knowledge and attitudes about tobacco use. Two models were constructed: (1) Model a: for all of the study subjects (all aged 15 years and above) education was excluded and (2) Model b: for those aged 25 years and above and education was included as an independent variable. Model a had similar results as model b. In this article, model b was presented in Table 5 of the results section while model a was presented in Table 6 in the appendix section. Multivariate logistic regression analysis indicated that after adjusting for demographic characteristics, accessibility to information, rule of ‘no smoking at home’, and smoking status, predictors of knowledge of health consequences of active smoking are education, ethnicity, access to information, and smoking status. Adults at lower secondary, upper secondary, and college or above were more likely to have significantly better knowledge of health consequences of active smoking than those at primary school (OR: 1.6, 1.7, and 1.9, respectively). It was also the case of knowledge on health consequences of secondhand smoke (OR: 2.4, 3.9, and 5.7, respectively). Adults belonging to non-Kinh ethnic had significantly lower knowledge of active and passive smoking-health risks than Kinh ethnic (OR: 0.7 and 0.4, respectively). This model also indicated that accessing positive information had significant association with knowledge of both active and passive smoking-health risks (OR: 2.3 and 1.9, respectively). Noticeably, current non-smokers have significantly better knowledge of active and passive smoking-health risks than current smokers (OR: 1.6 and 1.7, respectively). Increasing age was positively related to knowledge of the health consequences of secondhand smoke (Table 5).


Table 5.  Logistic regression analysis for factors associated with knowledge of health consequences of smoking (model b)
    Dependent variables  
    Knowledge on health risks of active smoking Knowledge of health risks of secondhand smoking
Independent variables Sub-categories OR** 95% CI* OR** 95% CI*
Gender Male 1      
  Female 0.9 0.8–1.1 1 0.7–1.4

         
Age group 25–34 1      
  35–44 0.9 0.7–1.2 2 1.5–2.7
  45–54 1 0.9–1.3 1.9 1.5–2.4
  55–64 1.2 1–1.5 2.2 1.6–2.9
  65 and above 1.3 1–1.6 1.3 1–1.7

         
Education Primary 1      
  Lower secondary 1.6 1.3–1.9 2.4 1.9–3
  Upper secondary 1.7 1.3–2.2 3.9 2.6–5.8
  College or above 1.9 1.4–2.5 5.7 3.7–8.6

         
Income Quintile 5 1      
  Quintile 1 0.9 0.7–1.1 0.5 0.3–0.6
  Quintile 2 0.9 0.7–1.1 0.8 0.6–1.1
  Quintile 3 0.8 0.7–1 0.7 0.5–1
  Quintile 4 0.9 0.8–1.1 0.9 0.6–1.2

         
Ethnic Kinh 1      
  Non-Kinh ethnic 0.7 0.5–0.8 0.4 0.3–0.6

         
Access to positive information No 1      
  Yes 2.3 2–2.6 1.9 1.6–2.3

         
Access to negative information No 1      
  Yes 0.7 0.4–1.1 0.7 0.2–1.9

         
Area Urban 1      
  Rural 1.1 1–1.3 1.1 0.9–1.3

         
Region Red River Delta 1      
  Northern midland and mountain 1.2 0.9–1.5 1.9 1.2–3
  North Central area and Central coastal 1.2 1–1.4 1.1 0.8–1.5
  Central highlands 1.2 0.8–1.7 1.2 0.7–2.1
  South East 1.1 0.9–1.4 0.8 0.6–1.1
  Mekong River Delta 0.9 0.7–1.1 0.5 0.4–0.7

         
Smoking status Current smokers 1      
  Current non-smoker 1.6 1.3–1.9 1.7 1.1–2.5
*95% confidence interval, **odds ratio.
N=8,265 individuals from 656 PSUs of 6 strata.

Discussion

This study found that although there was a high proportion among adults answering that active and secondhand smoking can cause serious illness (Table 1), only 51.5% of them understood that smoking can cause all three specific diseases (stroke, heart attack, and lung cancer) which were scientifically documented to have close relationships with smoking (Table 2) (12, 13). The finding that the risk of lung cancer was most frequently reported is consistent with other findings about the causes of disease reported by adults, even though heart disease is the number one killer of smokers (14).


The difference in knowledge between current smokers and current non-smokers was also studied elsewhere. Yang et al. found in the 2010 GATS China that current smokers were aware of fewer health effects of smoking than current non-smokers, respectively. For individual health effects, only 68% of current smokers agreed that smoking causes lung cancer in smokers while among current non-smokers, the percentage is more than 90%. In addition, only 36% of current smokers agreed that smoking causes coronary heart diseases while among current non-smokers the percentage is over 50% (15). The difference in knowledge of health risk between smokers and non-smokers is similar to patterns observed in China and Western countries, where smokers systematically underestimated their personal risks from smoking, presumably in an attempt to minimize cognitive dissonance from smoking and shield themselves from worry (1618). Regarding differences in knowledge between Kinh ethnic and non-Kinh ethnic, the 2002 Vietnamese national health survey indicated that non-Kinh ethnic groups are people living in rural areas with lower levels of education than those living in urban areas (19), and a World Bank survey indicated that 90% of the poor in Vietnam live in the rural areas (20). ‘This has resulted in significant educational challenges’, as said by the Vice General Director of the Department of Sports, Entertainment and Economic Information at Viet Nam Television (VTV). In addition, in a very recent household survey in Vietnam conducted in 2011, it indicated that people at low education levels were more likely to smoke (21). This is where the IEC can play an important role in preventing smoking. Chee Ruey Hsieh in his study on knowledge of health risks in anti-smoking campaigns found that anti-smoking campaigns had a significantly positive effect on the public health-related knowledge (22). The Centers for Disease Control (CDC)'s best practice guidelines suggested that public education is an integral part in the efforts to both prevent initiation of tobacco use and to encourage tobacco cessation (23). This current study supports the importance of the ability to access information in both descriptive and multivariate analyses. Those accessing information of the health harms of active and secondhand smoking were 2.3 and 1.9 times, respectively, more likely to have more knowledge than those who did not. Returning to the first major model for communication in 1949 by Claude Elwood Shannon and Warren Weaver, the process of communication was broken down into eight discrete components, that is, information source, message, transmitter, signal, channel, noise, receiver, and destination. The current study has identified three out of eight components of this model that should be carefully considered when developing and carrying out an IEC program for tobacco control conducted in Vietnam.

First, by indicating that an understanding about specific health risks related to tobacco smoking among Vietnamese adults, especially among current smokers and non-Kinh ethnic, may still not be specific, this study can help inform IEC programs designed to prevent tobacco smoking; messages should be designed to be scientifically credible, comprehensive, and consistent for the nation as a whole. Second, by indicating that current smokers and non-Kinh ethnic groups have lower levels of knowledge than other groups, it may be necessary to target messages to individual population subgroups. Third, by indicating that access to positive information is predictive of knowledge, this current study highlights the importance of coverage of an IEC program. This issue is specially concerned in Vietnamese context because the GATS 2010 indicated that, percentage of population accessing to mass media was still very low (32.8%) and that accessing to health warnings on cigarette packets among current smokers was only 14% and among general population was only 12.7% (24).

Therefore, in terms of policy implication, it is necessary to develop a national IEC program for preventing smoking tobacco which would be designed for different target groups of adults, including a general one, smokers, and non-Kinh ethnic groups in which clear and comprehensive messages/images about the health harm of tobacco smoking is conveyed appropriately and efficiently by different channels/modes to local-specific cultures.

Conclusion

The 2010 GATS in Vietnam showed that adults’ knowledge of specific diseases related to tobacco smoking was still vague as reflected in only 51.5% adults knowing that smoking can cause all three diseases of stroke, heart attack, and lung cancer. Regarding knowledge of health harms of active and passive smoking, current non-smokers were 1.6–1.7 times likely to have better knowledge than current smokers, respectively; non-Kinh ethnic groups were less likely to have knowledge (OR=0.7 and 0.4, respectively) than Kinh ethnic group smokers. Accessing positive information had a close association with knowledge of smoking-health risks (OR=2.3 and 1.9, respectively, with p<0.001). The more education adults had, the better knowledge of health consequences of tobacco smoking they got. Increasing age was positively related to knowledge of the health consequences of secondhand smoke.

Acknowledgements

Since the beginning of the 2-year survey, we have received very close and valuable technical assistance from CDC in the development of questionnaires, sample design, data analysis, as well as standardized GATS methodology and protocols in a series of manuals and guidelines. Our appreciation goes out to CDC for these valuable contributions.

We also acknowledge and highly appreciate the strong commitment, leadership, and support from Ministry of Health and Ministry of Planning and Investment for completing this survey. Excellent cooperation from the Vietnam Steering Committee on Smoking and Health, the GSO, and Hanoi Medical University has contributed to the success of this project.

Collaboration and support from related governmental and non-governmental organizations and tobacco control experts are also highly appreciated.

Our sincere thanks go to the World Health Organization, from Headquarters to Regional Offices to Country levels, for facilitating GATS implementation, providing technical and management assistance, financial support, and coordinating national and international partners.

The most grateful acknowledgment goes to the hard work of field supervisors, field interviewers, and all respondents. Without their contributions, our work would never have been possible.

Conflict of interest and funding

The authors have not received any funding or benefits from industry or elsewhere to conduct this study.

References

  1. Guindon G, Hien NTT, Emily M (2010). Tobacco taxation in Vietnam. Report of the International Union Against Tuberculosis and Lung Disease. Paris, ISBN 2-914365-65-9. p. 3.
  2. CDC (2003). Global youth health survey. Atlanta: Centers for Disease Control and Prevention.
  3. Centers for Disease Control and Prevention (2007). Global youth tobacco survey (GYTS) 2007. Atlanta: Centers for Disease Control and Prevention.
  4. An DTM. Establishing smoke-free environments in 9 hospitals in Vietnam. APACT 2010. Handbook and Conference Program. FCTC in the Asia Pacific: Change, Challenge and Progress. APACT 2010-Sydney, Australia, 2010, p. 78.
  5. Hang PT, Trang DT, Huy NV. Measuring SHS exposure in public places in Hanoi, Vietnam in support of policy development and implementation. APACT 2010. Handbook and Conference Program. FCTC in the Asia Pacific: Change, Challenge and Progress. APACT 2010-Sydney, Australia, 2010, pp. 101–2.
  6. Rosenstock IM, Strecher VJ, Becker MH. Social learning theory and the health belief model. Health Educ Q 1988; 15: 175–83. PubMed Abstract | Publisher Full Text
  7. Ukwayi JK, Eja OF, Chibuzo C. Unwanede, peer pressure and tobacco smoking among undergraduate students of the University of Calabar, Cross River State. High Educ Stud 2012; 2: 92.
  8. Kennedy D. Knowledge of the causal association between smoking and blindness in Canada, United States, United Kingdom, and Australia: findings from the ITC four country survey. APACT 2010. Handbook and Conference Program. FCTC in the Asia Pacific: Change, Challenge and Progress. APACT 2010-Sydney, Australia, 2010, p. 141.
  9. Otani J. Awareness of college students on health impact of smoking and impact of health education, measured by the Kano Test for Social Nicotine Dependence (KTSND): comparative survey of Japan and Central Asian countries. APACT 2010. Handbook and Conference Program. FCTC in the Asia Pacific: Change, 2010. Challenge and Progress. APACT 2010-Sydney, Australia, p. 163.
  10. Vinacosh, International Cooperation. Vietnam Steering Committee on Smoking and Health; 2012. Available from: http://www.vinacosh.gov.vn [cited 15 January 2012].
  11. Warren CW, Lee J, Lea V, Goding A, O'Hara B, Carlberg M, et al. Evolution of the global tobacco surveillance system (GTSS) 1998–2008. Glob Health Promot 2009; 16(2 Suppl): 4–37. PubMed Abstract | Publisher Full Text
  12. CDC. State-specific smoking-attributable mortality and years of potential life lost – United State, 2000–2004. MMWR Morb Mortal Wkly Rep 2009; 58: 91.
  13. U.S. Department of Health and Human Services (2010). How tobacco smoke causes disease. Report of the Surgeon General – Executive Summary. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention.
  14. U.S. Department of Health and Human Services (2004). The health consequences of smoking a report of the surgeon general. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health.
  15. Yang J, Hammond D, Driezen P, Fong GT, Jiang Y. Health knowledge and perception of risks among Chinese smokers and non-smokers: findings from the Wave 1 ITC China Survey. Tob Control 2010; 19(Suppl 2): 18–23. Publisher Full Text
  16. Weinstein N. Public understanding of the illnesses caused by smoking. Nicotine Tob Res 2004; 6: 349–55. PubMed Abstract | Publisher Full Text
  17. Slovic P. Smoking: risk, perception and policy. Thousand Oaks, CA: Sage; 2001.
  18. Weinstein N. Accuracy of smokers’ risk perception. Ann Behav Med 1998; 20: 135–40. PubMed Abstract | Publisher Full Text
  19. Committee for Population. Demographic and Health Survey 2002. Calverton, MD: Committee for Population, Family and Children and ORC Macro; 2003, p. 11.
  20. The Word Bank. Rural development & agriculture in Vietnam. Available from: http://web.worldbank.org/WBSITE [cited 15 September 2011].
  21. Cuong NV. Demographic and social-economic determinants of smoking behaviors: evidence from Vietnam. Econ Bull 2012; 32: 2300–12.
  22. Hsieh CR, Yenb LL, Liu JT, Lin CJ. Smoking, health knowledge, and anti-smoking campaigns: an empirical study in Taiwan. J Health Econ 1996; 15: 87–104. PubMed Abstract | Publisher Full Text
  23. CDC. Tips from former smokers. Video former smoker campaign; 2012. Available from: http://www.cdc.gov/tobacco/campaign/tips/resources/videos/ [cited 22 May 2011].
  24. Ministry of Health (2010). Global adult tobacco survey in Vietnam. Ministry of Health.

*Dao Thi Minh An
Department of Epidemiology
Institute for Preventive Medicine and Public Health
Hanoi Medical University
Hanoi, Vietnam
Email: daothiminhan@yahoo.com

Appendix

Table 6.  Logistic regression: factors associated with knowledge of health consequences of smoking (model a)
    Dependent variables  
    Knowledge on health risks of active smoking Knowledge of health risks of secondhand smoking
Independent variables Sub-categories OR** 95% CI* OR** 95% CI*
Gender Male 1      
  Female 0.8 0.7–1 0.8 0.5–1.1

         
Age group 25–34 1      
  35–44 0.8 0.6–1 3 2.2–4.1
  45–54 0.8 0.7–1.1 2.6 2–3.4
  55–64 1 0.8–1.2 3.1 2.3–4.1
  65 and above 0.7 0.5–0.9 1.7 1.2–2.2

         
Education Primary 1      
  Lower secondary 1.6 1.3–1.9 2.4 1.9–3
  Upper secondary 1.7 1.3–2.2 3.9 2.6–5.8
  College or above 1.9 1.4–2.5 5.7 3.7–8.6

         
Income Quintile 5 1      
  Quintile 1 0.7 0.6–0.9 0.3 0.2–0.4
  Quintile 2 0.8 0.6–1 0.5 0.4–0.7
  Quintile 3 0.8 0.6–0.9 0.5 0.4–0.7
  Quintile 4 0.9 0.7–1 0.7 0.5–1

         
Ethnic Kinh 1      
  Non-Kinh ethnic 0.6 0.5–0.7 0.3 0.3–0.5

         
Access to positive information No 1      
  Yes 2.4 2.1–2.8 2.2 1.8–2.6

         
Access to negative information No 1      
  Yes 0.6 0.4–1.1 0.6 0.2–1.8

         
Area Urban 1      
  Rural 1.1 0.9–1.2 1 0.8–1.2

         
Region Red River Delta 1      
  Northern midland and mountain 1.2 0.9–1.5 1.8 1.1–2.9
  North Central area and Central coastal 1.2 1–1.4 1.1 0.8–1.5
  Central highlands 1.1 0.7–1.6 1 0.6–1.8
  South East 1 0.8–1.3 0.6 0.5–0.9
  Mekong River Delta 0.8 0.6–1 0.4 0.3–0.6

         
Regulation of ‘no smoking at home’ No 1      
  Yes 1 0.8–1.3 0.8 0.6–1.2

         
Smoking status Current smokers 1      
  Current non-smokers 1.6 1.3–2 1.8 1.2–2.7
*95% confidence interval, **odds ratio.
N=9,919 individuals from 656 PSUs of 6 strata.

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Alcohol consumption and household expenditure on alcohol in a rural district in Vietnam

Kim Bao Giang1*, Hoang Van Minh1 and Peter Allebeck2

1Institute for Preventive Medicine and Public Health, Hanoi Medical University, Hanoi, Vietnam; 2Social Medicine, Public Health Sciences, Karolinska Institute, Solna, Sweden

Abstract

Introduction: Alcohol use and alcohol-related problems are on the rise in low- and middle-income countries. Expenditure on alcohol is an important problem for families and communities and needs to be assessed.

Aim: This study examines level of alcohol consumption and expenditure on alcohol in a district in Vietnam.

Methods: A cross-sectional survey was conducted in a rural district in northern Vietnam. Multi-stage sampling was employed to randomly select participants from 20 communities and a town in the same district. One thousand five hundred and sixty-four adults (765 males and 799 females) aged 18–60 years were interviewed. Information about alcohol use as well as expenditure on alcohol consumption four weeks prior to the interview was gathered. Non-parametric tests and log-linear regression were employed to compare expenditure on alcohol consumption across socioeconomic groups.

Results: The prevalence of alcohol use one month prior to interview was 35% (66% among men and 5% among women). The median alcohol consumption among those who reported use of alcohol in the week prior to the interview was 7.9 standard drinks. Excessive drinking (more than 14 standard drinks per week for men and more than seven standard drinks per week for women) occurred among 35% of those who used alcohol. Median expenditure for alcohol consumption during one month by those who drank alcohol was USD 3.5, accounting for 4.6% of household food expenditure, 2.7% of total household expenditure, and 1.8% of household income. The differences in alcohol consumption and expenditure between sexes and between socioeconomic groups are also presented.

Conclusion: Our study confirms that alcohol consumption and alcohol-related problems are common among men in Vietnam. The share of alcohol expenditure in total household expenditure is substantial, especially among poor households. This should be considered an important public health issue, which needs to be taken into account in the alcohol policy debate.

Keywords: alcohol consumption; alcohol problems; alcohol expenditure; Vietnam

Received: 10 June 2012; Revised: 4 January 2013; Accepted: 4 January 2013; Published: 28 January 2013

Glob Health Action 2013. © 2013 Kim Bao Giang et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 18937 - http://dx.doi.org/10.3402/gha.v6i0.18937

 

Alcohol has a causal relationship with more than 60 medical conditions, and overall it causes 4% of total global burden of diseases (1, 2). Social and economic burden of alcohol consumption has also been well documented (25). The cost of alcohol problem has been estimated to be 1% of the gross domestic product in Canada and Australia (6), and 4% in New Zealand (7). In the United States, more than 20 cost components were included in a model to estimate economic cost of alcohol abuse, which consisted of health care cost, lost earning due to premature deaths, alcohol-related illness, crime and victim, and so on. The economic cost of alcohol abuse in the United States has increased by 3.8% every year (8). The social cost of alcohol consists of different types of alcohol-related costs, such as health service resources used, social work services, criminal justice, emergency services, economic cost, and human cost. The social cost of alcohol in Europe is estimated to be from 1 to 3% (9). Meanwhile, the alcohol-related tax revenue covers only a small part of the economic cost of alcohol abuse (4). Alcohol consumption also has a significant share in total household or consumer expenditure in many countries. It amounts to 5.8% of all consumer expenditure and 5.2% of household consumption expenditure in the United Kingdom (9).

Policy recommendations

In this study, we found that alcohol use and alcohol problems are very common among men in Vietnam. The share of alcohol expenditure in household expenditure is substantial, especially among poor households. Following are key policy recommendations:

  • Alcohol use among men should be considered an important public health issue in Vietnam. Health communication and prevention programs should be conducted to raise community awareness and to minimize negative impacts of alcohol use in Vietnam.
  • Development of alcohol policies and strategies should be given priority. Measures to reduce availability of, and accessibility to alcohol are necessary to decrease alcohol abuse, especially among men in Vietnam.

The level of alcohol consumption is reported to have decreased in developed countries, but it has been increasing in developing countries (2). In Vietnam, alcohol is widely available in the market and easily accessible (10). Alcohol drinking among men is traditionally and socially encouraged. Vietnam is a country with a high prevalence of alcohol use and alcohol-related problems (1113). Among men, the prevalence of alcohol use during the 1-week period has been estimated to be 46% (12). A study that used the Alcohol Use Disorder Identification Test (AUDIT) – an instrument developed by the World Health Organization (WHO) to detect alcohol problems in Vietnam found the prevalence of alcohol-related problems to be 25.5% (13).

As the results of the increase of production in connection with economic growth in Vietnam, alcohol consumption has increased during the past decades (2). Thus, more economic and social costs would be attributed to alcohol consumption. In 2006, drunk driving caused 6.8% of traffic accidents (14). Alcohol use and alcohol-related problems are thus increasing public health concerns in Vietnam, as well as in many other low- and middle-income countries. For an informed debate on alcohol policy, it is important to take into account not only the well-known harmful effects on health but also other effects. Excessive expenditure on alcohol can be detrimental to families and communities. This study examines alcohol consumption and expenditure on alcohol consumption in a district in Vietnam.

Methods

Setting

This study was conducted in Thanh Oai district, Ha Noi, Vietnam, a typical rural district in northern Vietnam. The district had 20 communities, a town with a total population of 204,729 people, and a land area of 129.6 km2. The main economic activities were agriculture (47%), small scale industries and construction work (25%), and trade (27%). Around 3.7% of the households in this district were classified as poor according to official statistics.

Design

A cross-sectional study was conducted in all communities and in the town of the district.

Sample size and sampling

Sample size was estimated using the WHO formula for sample size to estimate prevalence of alcohol problems (p =0.25 is the prevalence of alcohol problems in previous studies (13, 15); relative precision = 0.1; significance level = 0.05). The required sample size was 1,153. This sample size is also adequate to compare expenditure on alcohol consumption in different socioeconomic groups. A multi-stage sampling was applied to select study subjects in this district. First, all 20 communities and a town in the district were selected. Second, Proportional Population to Size sampling technique was applied to estimate number of households to be interviewed from each community. To prevent non-response cases, in each community, we increased the sample size by 10–20 households based on its population size (3, 8% of total required sample size). The actual sample size was 1,600 households. Systematic sampling was used to randomly select households from each community. Finally, in each selected household, a person in the age range of 16–60 was randomly selected to be interviewed using Kish table. This is also called the Kish grid, a method using a pre-assigned table of random numbers to find the persons to be interviewed in a household (16). Eleven persons who were too sick or incapable of participating were excluded. Twenty-five persons could not be reached because of absence from home during the survey. Interviews were completed with 1,564 persons.

Key variables

Alcohol consumption

Consumption of any kinds of alcoholic beverage. Beer at 4%, wine at 12%, 30% and 40% of pure alcohol are among the most common used beverage in this rural Vietnam. Information on alcohol consumption during 1 week and frequency of alcohol consumption during 4 weeks and 12 months prior to the interview was gathered.

Alcohol user

Those persons who consumed any kind of alcoholic beverage during 12 months prior to the interview.

Standard drink

The WHO definition of a standard drink was used to assess the volume of alcohol consumed during the week prior to the interview. A standard drink refers to the amount of about 12.6 g pure alcohol, which equals to 330 ml of 5% beer, 40 ml of 40% liquor, or 130 ml of 12% wine, and so on (17).

Alcohol consumption per person per year

Volume of alcohol consumption during a 1-week period, multiplied up to 1 year (4.333 weeks per month and 12 months per year).

Binge drinking

Consumption of more than five standard drinks in one sitting.

Excessive drinking

Intake of more than 14 standard drinks a week for men (more than two standard drinks a day) and more than seven standard drinks a week for women (more than one standard drink a day) (18).

Alcohol problem

An AUDIT score of 8 or higher was recognized as a social and health problem caused by alcohol. The AUDIT was developed by the WHO, and it has been used in many settings. The AUDIT consists of 10 questions, which have a possible maximum score of 40. The questions 1–8 may score 0–4 points, and question 9 and 10 may score 0, 2, or 4 points. The optimal cutoff level was found to be 8, and this has been proved to be reliable and valid to diagnose alcohol problems in many countries (17), including Vietnam (13, 15).

Household income quintile

The total income of a household during 1 year prior to data collection. The total household income from different sources and economic activities of all household members was assessed getting information from the head of the household.

Household expenditure during 1 month

Expenditure on different beforehand categories, including expenditure on 1) food and drinks; 2) electricity, telephone, and fuels; 3) education; 4) health care; and 5) others.

Expenditure on alcohol consumption

The entire money that an alcohol user spent on alcohol during a month. We did not take into account alcohol consumption, wherein the respondents did not have to pay.

Share of total household expenditure on alcohol expenditure

Expenditure on alcohol reported during 1 month prior to the interview divided by total expenditure of the household during the said month.

Share of household monthly food expenditure on alcohol expenditures

Expenditure on alcohol divided by the total expenditure on food and drink in the household in 1 month period.

Proportion of alcohol expenditure in total household monthly income

Expenditure on alcohol reported during 1 month prior to the interview divided by the household monthly income.

Data collection

A questionnaire was developed to collect the information as mentioned above, including background information of study subjects, drinking frequency, average drinking volume, and expenditure on alcohol. The Vietnamese version of the AUDIT was integrated in the questionnaire to help identify cases with alcohol problems. The questionnaire was tested and revised to make sure all questions were understandable by participants. Face-to-face interviews were performed by 20 health staff from communal health stations in Thanh Oai. Interviewers were carefully trained to understand the purpose of the study, how to ask questions, and how to fill in the interview forms. Four staff members from district health centers and four research staff from the Hanoi Medical University were recruited for field supervision. Daily field supervision was done to ensure the quality of data collection.

Data management and data analysis

Data were entered using the Epidata software. Check file was used to ensure logical information and reduce errors during data entry. STATA 10 was used to perform all analyses. Chi-square test was used to compare the prevalence of alcohol use and alcohol problems across groups. Non-parametric tests and linear regression were employed to compare expenditure for alcohol consumption across socioeconomic groups. To perform the linear regression model, logarithmic transformation was performed on four variables, including alcohol consumption, alcohol expenditure, share of alcohol expenditure in total household expenditure, and proportion of alcohol expenditure in household monthly income. In the log-linear regression models, number of household members, age group, marital status, education, occupation, and income quintiles were included as independent variables.

Ethical clearance

This study was approved by the Ethical Committee of the Hanoi Medical University. Oral informed consent was gathered for each interview during data collection. Study subjects were informed about the study objectives, confidentiality of their individual information, and their right to refuse or stop the interview at any stage of the study.

Results

Table 1 presents sociodemographic characteristics of the study subjects. Among 1,564 participants, 48.9% were men and 51.1% were women. Half of them (50.7%) were in the age range of 25–44 years, 74.7% were married, 72.8% of them had secondary level of education or higher, 44.6% of them were farmers, and only 16.2% of them were working in public offices or were students. Men and women had similar distribution of age and marital status. The proportion of farmers among women was higher than that among men (48.3% vs. 40.7%, p =0.0013), while the proportion of men who had at least high school level of education was higher than that of women (32.3% vs. 20.9%, p <0.0001).


Table 1.  Background information of study subjects
    Males (N =765) Females (N =799) Overall
Variables   n % n % N %
Age group 18–24 168 22.0 175 21.9 343 21.9
  25–44 386 50.5 401 50.2 787 50.3
  45–60 211 27.6 223 27.9 434 27.7
Marital status Other 196 25.6 199 24.9 395 25.3
  Married 569 74.4 600 75.1 1,169 74.7
Education level Primary and lower 181 23.6 245 30.7 426 27.3
  Secondary 337 44.1 387 48.4 724 46.3
  High school and upper* 247 32.3 167 20.9 314 26.5
Occupation Farmer 311 40.7 386 48.3 697 44.6
  Government staff and student 126 16.4 128 16 254 16.2
  Worker 195 25.5 117 14.6 312 19.9
  Others 132 17.2 166 20.7 298 19.1
Income quintile Poorest* 132 17.3 182 22.8 314 20.1
  Second poorest 153 20.0 165 20.7 318 20.3
  Middle 167 21.8 158 19.8 325 20.8
  Second richest 148 19.4 151 18.9 299 19.1
  Richest 165 21.6 143 17.9 308 19.7
*The difference between two sexes is statistically significant by Chi-square test.

Alcohol use and alcohol problem were much more common among men. The prevalence of alcohol use during the 12-month period was 49.6% (82.9% among men and 17.8% among women). The prevalence of alcohol use during 1 week prior to the interview was 34.7% (65.5% among men and 5.3% among women, p <0.0001). The prevalence of excessive drinking was 24.4% among all men and was 0.4% among all women (p =0.0001) in the sample. The prevalence of excessive drinking was 35% among all alcohol users (37.3% among male alcohol users and 7.1% among female alcohol users; p <0.0001). The prevalence of binge drinking was 7% (14% among men and only 0.3% among women, p <0.0001). Alcohol problem was only recorded among males (24.1%) (Table 2).


Table 2.  Alcohol use and alcohol problem
    Males (N =765) Females (N =799) Overall
Variables   n % n % N %
Alcohol consumption Last 12 months* 634 82.9 142 17.8 776 49.6
  Last week* 501 65.5 42 5.3 543 34.7
Drinking frequency None 131 17.1 657 82.2 788 50.4
  Monthly or less 159 20.8 128 16.0 287 18.4
  2–4 times a month 171 22.4 9 1.1 180 11.5
  2–3 times a week 133 17.4 0 0.0 133 8.5
  4 or more times a week 171 22.4 5 0.6 176 11.3
Average drinking level per sitting 1–2 drinks 571 74.6 796 99.6 1,367 87.4
  3–4 drinks 144 18.8 3 0.4 147 9.4
  5–6 drinks 39 5.1 0 0.0 39 2.5
  ≥7 drinks 11 1.5 0 0.0 11 0.7
Excessive drinking Among drinkers# 187 37.3 3 7.1 190 35.0
  Among study sample 187 24.4 3 0.4 190 12.2
Binge drinking Never 658 86.0 797 99.7 1,455 93.0
  Less than monthly, %* 76 9.9 2 0.3 78 5.0
  Monthly 27 3.5 0 0.0 27 1.7
  Weekly or daily 4 0.6 0 0.0 4 0.2
Alcohol problem AUDIT score ≥8* 184 24.1 0 0.0 184 11.8
*The difference between two sexes is statistically significant by Chi-square test.
#The prevalence is only calculated for alcohol users during 12 months prior to the interviews (Male alcohol users: 501; female alcohol users: 44).

The median of alcohol consumption during 1-week period prior to the interview among alcohol users was 7.9 standard drinks (8.9 standard drinks for males and 1.4 standard drinks for females) (Table 3). Non-parametric test comparing medians found that men significantly consumed more alcohol than did women (Table 3). Calculated as volume of alcohol per year, the average was estimated to 4.73 liters per person per year (9.37 liters per male and 0.33 liters per female).


Table 3.  Number of standard drinks consumed by alcohol users during 1-week period
  Alcohol users
Indicators Male Female Overall
Quartile at 25% 3.0 0.8 3.0
Quartile at 50% (median) 8.9 1.4 7.9
Quartile at 75% 21.3 2.0 20.7

The mean expenditure on alcohol during 1 month among alcohol users was USD 6.5. More than 50% of the alcohol users spent at least USD 3.5, and 25% of alcohol users spent at least USD 7.1 for alcohol within a month.

On average, the monthly alcohol expenditure of an alcohol user accounted for 4.6% of monthly household food and drink expenditure, 2.7% of total household monthly expenditure, and 1.8% of household monthly income. Of the alcohol users, 25% spent at least 10% of the household food expenditure, 6% of the total household expenditure, or 4% of total household income on alcohol a month. Descriptive analysis shows that alcohol users in the middle income group had spent least money on alcohol (Table 4).


Table 4.  Alcohol expenditure (in USD), share of alcohol expenditure in household food expenditure (%), in household expenditure (%), in household income among alcohol users during 1 month period
Variables Income quintile Median IQR25 IQR75
Alcohol expenditure per month (USD) Poorest 2.8 0.9 4.7
  Second poorest 3.5 1.9 5.9
  Third richest 2.4 1.2 4.7
  Second richest 3.5 2.1 7.1
  Richest 4.7 1.9 7.5
  Overall 3.5 1.6 7.1
Share of household monthly food expenditure on alcohol expenditure (%) Poorest 5.7 1.8 13.3
  Second poorest 5.7 2.7 8.6
  Middle 3.6 1.9 8.0
  Second richest 4.8 2.7 10.0
  Richest 4.4 2.0 12.0
  Overall 4.6 2.2 10.0
Share of monthly household expenditure on alcohol expenditure (%) Poorest 2.8 1.2 8.3
  Second poorest 3.7 1.9 5.7
  Third richest 2.1 1.1 4.8
  Second richest 2.9 1.6 6.1
  Richest 2.6 1.3 6.0
  Overall 2.7 1.3 6.0
Share of monthly household income on alcohol expenditure (%) Poorest 4.7 2.1 11.5
  Second poorest 3.1 1.8 5.2
  Third richest 1.7 0.8 3.5
  Second richest 1.6 0.9 3.0
  Richest 1.1 0.5 2.2
  Overall 1.8 0.9 4.0
SD, Standard deviation; IRQ, Interquartile.

In the log-linear regression models, female alcohol users consumed significantly less alcohol than male. The married persons consumed significantly less alcohol than did unmarried persons. Those who had alcohol problems consumed significantly more alcohol than others. Those who had high school or higher education spent significantly less on alcohol, and had a lower share of household monthly food expenditure, household monthly expenditure, and household income each month on alcohol expenditure compared to those who had primary educational level or less.

Alcohol users in the highest income group spent significantly more money on alcohol than did the lower income groups. However, there was a clear and significant tendency that the better off groups had a lower share of household income on alcohol expenditure. There was no significant difference between income quintiles with regard to the share of household food expenditure and total household expenditure on alcohol expenditure.

Persons with alcohol problem spent more on alcohol, had larger share of alcohol expenditure in household food expenditure, total household expenditure, and household income. Those who were in the age range of 45–60 years had a lower share of household food expenditure and total household expenditure spent on alcohol during a month. No difference between occupation groups was found (Table 5).


Table 5.  Intercepts and slopes for each variable, resulting from linear regression models for logarithm of monthly expenditure on alcohol consumption (USD), share of monthly household food expenditure, and household expenditure and household income on alcohol expenditure (%)
        Share on alcohol expenditure by (n =422)
Variable   Alcohol consumption (n =528)# Alcohol expenditures (n =422) Household monthly food expenditure Monthly household expenditure Monthly household income
Number of household members   −0.03 −0.02 −0.12* −0.11* −0.03
Gender Male          
  Female −1.15* −0.13 −0.19 −0.12 −0.18
Age group 18–24          
  25–44 −0.20 −0.25 −0.36 −0.35 −0.26
  45–60 0.05 −0.32 −0.56* −0.58* −0.41
Marital status Other          
  Married −0.65* −0.24 −0.28 −0.21 −0.18
Education level Primary and lower          
  Secondary 0.11 −0.17 −0.26 −0.25 −0.17
  High school and upper 0.01 −0.32* −0.55* −0.49* −0.36*
Occupation Farmer          
  Government staff and student −0.19 0.14 0.19 0.08 0.16
  Worker 0.17 0.22 0.17 0.24 0.23
  Others −0.04 0.08 0.04 0.08 0.08
Income quintile Poorest          
  Second poorest 0.07 0.20 0.02 0.05 −0.47*
  Middle −0.16 0.01 −0.24 −0.30 −1.03*
  Second richest 0.23 0.23 −0.05 −0.04 −1.09*
  Richest 0.36 0.40* −0.04 −0.12 −1.54*
Alcohol problem No          
  Yes 0.99* 0.71* 0.69* 0.68* 0.71*
Intercept   1.893 10.89 2.46 1.91 1.82
R2   0.218 0.15 0.13 0.14 0.28
#Only estimated among those who consumed alcohol during 1-week period.
Only estimated among those drinkers who had to pay for alcohol during 1-week period.
*Statistical significant (p <0.05).

Discussion

This study was conducted in a rural district in northern Vietnam. The population structure, educational status, economic activities, and economic and cultural status of this district are the same as several other districts in Vietnam. The findings are likely to reflect the situation regarding alcohol consumption and expenditure in rural Vietnam, although there are local variations.

We found that the 1-week prevalence of alcohol consumption was 34.7%, and the 12-month prevalence was 49.6%. The median of alcohol consumption among alcohol users was 7.9 standard drinks per week, and alcohol consumption estimated for a person per year was 4.73 liters of pure alcohol. This is slightly higher than what was recently reported by the WHO (3.8 liters per capita per year for persons aged 15 year old and above in Vietnam). Although Vietnam is not yet among the list of countries with high alcohol consumption levels, it is important to note that the level of alcohol consumption in Vietnam has increased at a rapid pace, by about 50% in the 5-year period between 2000 and 2005 (19). Although results from different studies in different time periods may not be comparable, together with other reports it seems clear that the prevalence of alcohol use has increased. For instance, in 2002, the National Health Survey found the 1-week prevalence of alcohol use to be 46% among men and 2% among women (12). In 2005, a study conducted in seven provinces reported 1-week prevalence of alcohol use to be 64% among men and 1% among women (11). Another study in rural Vietnam reported 12-month prevalence of alcohol use among male adults to be 87.3% and 10.2% among female adults (13). Alcohol drinking is reported to be on the rise in developing countries (2), including Vietnam (19).

The prevalence of alcohol problem was 24.1% among men, while none was found among women. This is supported by other studies reporting a similar distribution of alcohol problem among men and women (11, 13). The prevalence of excessive drinking in our study was 35% (37.3% among men and 7.1% among women). The prevalence of excessive drinking among men in our study is in line with the report of a study in seven provinces in Vietnam that 36% of men consumed more than three standard drinks per day. (11). Our study again indicates clear gender differences with regard to alcohol use, alcohol consumption, and alcohol problem (1113). The prevalence of alcohol use among women in our study is slightly higher than reported in earlier studies from Vietnam (1113). This suggests that alcohol use among women could have increased in Vietnam during the past few years. This can be seen as the result of socioeconomic development that contributes to close the gap between men and women, as has been observed in several other countries (20, 21).

Although we only took into accoun alcohol expenditure by a drinker from each household, we found that alcohol expenditure accounted for 4.6% of household food expenditure and 2.7% of total household expenditure. The share of alcohol in the household expenditure in our setting is slightly lower than what was reported in a country with higher consumption, such as the United Kingdom (19). In the United Kingdom, alcohol expenditure accounted for 5.2% in 2006 (9) and 3.4% in 2010 (3) of household expenditure.

In our setting, median of monthly expenditure on alcohol consumption of each individual alcohol user was USD 3.5. This is already about 30% of per capita income per month of a poor household in Vietnam, where the poverty line set by the government for the period 2006–2010 was VND 200,000 per capita per month (equal to USD 11.8) (22). Using the results of our study to crudely estimate total household expenditure on alcohol consumption in Vietnam, total expenditure on alcohol consumption among adults aged 18–60 years would be around USD 1.4 billion each year (about 60% of the 86 million population is in the age range of 18–60 (23), and the prevalence of alcohol consumption in that age group is 34.7%). These are crude estimations and may underestimate the consumption and expenditure in urban areas. Our figures nevertheless indicate an important burden on the economy of the household and that of the community.

It may seem obvious that those who had alcohol problems spent significantly more on alcohol and had larger share of alcohol in household expenditure and household income. We found that the highest income group spent significantly more on alcohol, but alcohol expenditure among the lowest income group accounted for a larger share of household income, thus increasing their already vulnerable situation. Also, Yen et al. (24) in a study from China found income, education, and other household characteristics to be significant determinants of household expenditure on alcohol. For a poor household, a large amount of income used for alcohol means less money available for food, health, and education. Thus, alcohol consumption could be one of the factors that put poor households at high risk of falling into a poverty trap.

Although Vietnam has already some policies, such as excise tax on alcohol products, mandatory business license to do alcohol trade, ban on alcohol sales in specific places, regulation on alcohol advertisement, ceiling on level of alcohol concentration in blood and breath of drivers, aimed at decreasing both alcohol supply and demand, these are far from being well implemented. Findings from this and other studies point to the need of improving and better enforcing current alcohol policy (11).

Limitation

This study was conducted in rural Vietnam, so generalization of the results to other regions should be made with caution. In the urban settings, women are expected to have higher prevalence of alcohol use and consume more alcohol because in the urban areas women have more social activities and are more equal to men than women in the rura areas. Household income and expenditure could have been underestimated because these data were collected from the head of the households. Alcohol expenditure in the urban settings would be higher than in the rural settings because people seem to use more industrial alcohol products that are more expensive than home-made products often used in the rural settings (25), thus increasing the share of household expenditure and income on alcohol. Although in a household it is common for a person to buy alcohol to share with other household members, our study may underestimate alcohol expenditure because only money spent to buy alcohol by an alcohol user from each household was considered. Nevertheless, our results indicate alcohol expenditure accounted for a large part of household expenditure and income.

One may argue that alcohol abuse has been seen as a negative behavior, prompting the user to probably underreport the level of alcohol consumption. Recall bias may also be a problem when recall is requested. However, self-report methods are widely used, and they can be reliable and valid (26, 27). In our study, participants were requested to report volume of alcohol consumption only for a recall period of 1-week. Furthermore, we clearly explained the purpose of the study as well as the confidentiality of the private information. Interviews were conducted in a comfortable and private environment.

Conclusion

Our study confirms that alcohol consumption and related problems are common among men in Vietnam. The share of alcohol expenditure in total household expenditure is substantial, especially among poor households, and it should be considered an important public health issue, which needs to be taken into account in the alcohol policy debate.

Conflict of interest and funding

The authors have not received any funding or benefits from industry or elsewhere to conduct this study.

References

  1. Room R, Babor T, Rehm J. Alcohol and public health. Lancet 2005; 365: 519–30. PubMed Abstract
  2. World Health Organisation (2004). Global status report on alcohol 2004. Geneva: Department of Mental Health and Substance Abuse.
  3. Collis J, Grayson A, Johal S. Econometric analysis of alcohol consumption in the UK. London: HM Revenue & Customs; 2010.
  4. Roeber J. The human and economic cost of alcohol abuse in New Mexico, 2006. New Mexico Epidemiol 2009; 2009: 1–3.
  5. Minnesota Department of Health (2011). The human and cost of alcohol use in Minnesota. Minnesota: Minnesota Department of Health.
  6. International Center for Alcohol Policies (1999). Estimating costs associated with alcohol abuse: towards a patterns approach. Report No. ICAP Report 7.
  7. Cohen A, Kleiman K, Saraceno B. World mental health casebook. Introduction. New York: Kluwer Academic Press; 2002.
  8. Harwood H. Updating estimates of the economic costs of alcohol abuse in the United States: estimates, updates methods, and data: report prepared by The Lewin Group for the National Institute on Alcohol Abuse and Alcoholism; 2000.
  9. Institute of Alcohol Studies (2008). Economic cost and benefits. Cambridge: Institute of Alcohol Studies.
  10. General Statistics Office, Ministry of Health, World Health Organisation, United Nations Children's Fund (2003). Survey assessment of Vietnamese youth.
  11. Health Strategy and Policy Institute. Situation assessment of alcohol abuse in Vietnam (in Vietnamese: danh gia tinh hinh am dung ruou bia tai Viet Nam). Hanoi: Medical Publishing House; 2006.
  12. Ministry of Health and General Statistic Office. National health survey, 2001–02. Hanoi: Hanoi Medical Publishing House; 2003.
  13. Giang KB, Allebeck P, Spak F, Van Minh H, Dzung TV. Alcohol use and alcohol consumption-related problems in rural Vietnam: an epidemiological survey using AUDIT. Subst Use Misuse 2008; 43: 481–95. PubMed Abstract | Publisher Full Text
  14. Tuan NN. Alcohol related traffic accident and its prevention. Bureau for Road and Railway Traffic Security Ministry of Public Security; 2009.
  15. Giang KB, Spak F, Dzung TV, Allebeck P. The use of audit to 518 assess level of alcohol problems in rural Vietnam. Alcohol Alcohol 2005; 40: 578–83. PubMed Abstract | Publisher Full Text
  16. Kish L. Survey sampling. New York, NY: John Wiley; 1965.
  17. Babor TF, John CH, Saunders J, Monteiro MG. AUDIT – The Alcohol Use Disorders Identification Test – guidelines for use in primary care. Geneva, Switzerland: World Health Organisation, Department of Mental Health and Substance Dependence; 2001.
  18. US Department of Agriculture. Nutrition and your health: dietary guidelines for Americans. Washington, DC: US Department of Health and Human Services; 2000.
  19. WHO (2011). Global status report on alcohol and health. Geneva: World Health Organization.
  20. Simons-Morton BG, Farhat T, ter Bogt TFM, Hublet A, Kuntsche E, Gabhainn SN, et al. Gender specific trends in alcohol use: cross-cultural comparisons from 1998 to 2006 in 24 countries and regions. Int J Public Health 2009; 54(Suppl 2): 199–208. DOI: 10.1007/s00038-009-5411-y. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  21. Wilsnack RW, Vogeltanz ND, Wilsnack SC, Harris TR. Gender differences in alcohol consumption and adverse drinking consequences: cross-cultural patterns. Addiction 2002; 95: 251–65. Publisher Full Text
  22. Prime Minister. Decision No 170/2005/QD-TTg referring the issue of poverty line for the period 2006–2010. In: Vietnamese Government, editor; July 08, 2005.
  23. General Statistics Office. Survey on population changes and family planning 1/4/2005: main findings. Hanoi: Statistics Publishing House; 2006.
  24. Yen ST, Jensen HH. Determinants of household expenditures on alcohol. Ames, IA: Center for Agriculture and Rural Development; 1995.
  25. Giang KB. Alcohol use in three provinces from three geographical regions in Viet Nam. Hanoi: Hanoi Medical University; 2012 (unpublished source).
  26. Del Boca FK, Darkes J. The validity of self-reports of alcohol consumption: state of the science and challenges for research. Addiction 2003; 98(Suppl 2): 1–12. PubMed Abstract | Publisher Full Text
  27. National Institute on Alcohol Abuse and Alcoholism. Alcohol consumption measures. In: Allen JP, Wilson VB, eds. Assessing alcohol problems: a guide for clinicians and researchers. 2nd ed. Bethesda, MD: National Institute on Alcohol Abuse and Alcoholism; 2003, pp. 75–100.

*Kim Bao Giang
Institute for Preventive Medicine and Public Health
The Hanoi Medical University
No. 1 Ton That Tung Street, Dong Da district
Hanoi, Vietnam
Tel: +84-913026483
Email: kbgiangvn@yahoo.com and kimbaogiang@hmu.edu.vn

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Alcohol-related harm among university students in Hanoi, Vietnam

Pham Bich Diep1,2*, Ronald A. Knibbe2, Kim Bao Giang1 and Nanne De Vries2

1Institute of training of Preventive Medicine and Public Health, Hanoi Medical University, Hanoi, Vietnam; 2Department of Health Promotion, CAPHRI School for Public Health and Primary Care, Maastricht University, Maastricht, The Netherlands

Abstract

Introduction and Aim: This study examines the prevalence of and risk factors for alcohol-related harm and types of harm among medical students from Hanoi Medical University (Vietnam). Risk factors include aspects of drinking patterns and relevant socio-demographic variables.

Study Design and Methods: A cross-sectional study involving 1st to 6th year students (N=1216; response rate 96.5%). Of these, 210 students from each academic year were randomly selected from a sampling frame covering all students from each academic year. Data were collected using a questionnaire distributed in class by researchers. Drinkers completed 23 questions on alcohol-related harm categorized into: 1) ‘negative influence on daily activities’; 2) ‘social conflict’; 3) ‘loss of control, acute consequences, and withdrawal’; 4) ‘mental health conditions’; and 5) ‘physical and medical health problems’. Logistic and Poisson regression models were used to identify the predictors of alcohol-related harm and the amount of harm, respectively.

Results: The prevalence of alcohol use associated with at least one or more of the five types of harm was higher in men (81.8%) than in women (60.4%). In female and male students, the most common harm category was ‘loss of control, acute consequences, and withdrawal’ (51.8 and 75.6%, respectively), followed by ‘negative influence on daily activities’ (29.4 and 55.8%, respectively). Age, living away from home, and average number of standard drinks per occasion among male drinkers, and age and frequency of drinking per week among female drinkers were associated with alcohol-related harm.

Conclusions: These data suggest that alcohol-related harm represents a serious public health problem among young educated individuals in Vietnam. The risk factors indicate that prevention should be aimed at aspects of drinking patterns and specific subpopulations defined by gender, age, and (for men only) type of living situation.

Keywords: female students; male students; alcohol-related harm; type of harm; drinking patterns; Vietnam

Received: 28 May 2012; Revised: 10 January 2013; Accepted: 10 January 2013; Published: 1 February 2013

Glob Health Action 2013. © 2013 Pham Bich Diep et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 18857 - http://dx.doi.org/10.3402/gha.v6i0.18857

Alcohol is the third leading cause of burden of disease worldwide. It is reported that out of all deaths among young people aged 15–29 years, 9% is related to alcohol (1). Harmful drinking patterns, such as binge drinking (mostly defined as drinking ≥6 glasses on one occasion), have increased among young adults and adolescents (2). Harmful drinking is also popular among college students (3, 4). A study among Vietnamese medical students indicated that the prevalence of alcohol consumption (65.5%) and alcohol-related problems (12.5%) is relatively high (5). However, data on the negative consequences of harmful drinking in this population are scarce.

Regarding alcohol-related harm, the prevalence of negative consequences of alcohol use is high in both male and female students in developed countries such as New Zealand and the USA (4, 6). After drinking, male and female students often experience blackouts, unintended/unprotected sexual activity, academic impairment, short and long-term physical illness, and poor mental health, as well as anti-social risk behavior, fights, and interpersonal violence (711). A study among young Australian students suggested that alcohol-related harm has increased dramatically in recent years (12); however, studies on alcohol-related harm among students in developing countries are still scarce. A study in Thailand also indicated that the prevalence of hangover, nausea, and vomiting among adolescent drinkers is high (46.9%) (13); however, information on alcohol-related harm among students in Vietnam is lacking. Two studies among adolescents/young adults in Vietnam examined the association between alcohol use and sexual behavior only; the results show a strong link between alcohol consumption and engaging in sexual behavior among both males and females (14, 15).

Policy recommendations

The findings from this study indicate that it is a particular concern to develop alcohol policies to reduce harmful use of alcohol among students. The following points summarize the key policy recommendations:

  • More attention should be paid to reduce the harmful use of alcohol among students.
  • Intervention should aim at risk factors, including aspects of drinking patterns and specific subpopulations defined by gender, age and (for men only) type of living situation.
  • Education should be emphasized aiming at increasing awareness of students about the harmful use of alcohol.

Drinking patterns, in addition to quantity, are strongly related to harm. The volume of consumption (mostly expressed as the number of standard drinks per week or month) is an important determinant of alcohol-related harm (16, 17). Nevertheless, the same volume can still conceal very different drinking patterns. The aspect of drinking patterns most widely examined is that of binge drinking. Other measures for incidental high consumption, such as the greatest number of glasses consumed on one occasion and frequency of drunkenness, are also related to an increased risk for harm. A national survey of students at 140 campuses in the USA showed that frequent binge drinkers and infrequent binge drinkers were 25 and 5 times, respectively, more likely to have experienced at least five harms compared to non-binge drinkers (3). Notably, students who drink both heavily and frequently experienced negative consequences almost three times as often as those who drink heavily but less often (18).

Socio-demographic factors (such as age and gender) are also associated with alcohol-related harm. A study in the United Kingdom showed that older adolescents are more likely to report alcohol-related violence and alcohol-related regretted sex (19). Many studies in developed countries (e.g. USA, New Zealand, Australia, Sweden, and Germany) and in developing countries (e.g. Thailand and China) report a gender difference in drinking patterns that influence harms (13, 2027). Similarly, a study among students in Vietnam reported that men were 14.3 times more likely to have alcohol problems compared with women (5).

The present study among Vietnamese students addresses three related questions: 1) What is the prevalence of alcohol-related harm in this group? 2) How are socio-demographic variables and drinking patterns associated with alcohol-related harm? 3) Do socio-demographic and drinking pattern variables explain the variation in the number of alcohol-related harms that students report?

Methods

Setting

A cross-sectional study was conducted between November 2008 and January 2009 at Hanoi Medical University (the oldest and largest medical university in Hanoi, Vietnam).

Sample size and sampling

Within each academic year, the World Health Organization sample size calculation was applied to calculate a sample size, assuming a 45% prevalence of alcohol-related harms among drinkers (with a precision of ±0.2 and a 95% confidence level). Since we cannot select drinkers, the sample size required is calculated based on the prevalence of alcohol use among students, that is, 65% (5); therefore, a sample size of 180 students per academic year was needed. This number was increased by 5% to account for losses and by an additional 10% to control for confounding, yielding a sample size of 207 students for each academic year (rounded up to 210 students per academic year). The total sample size of 1260 students was also sufficient to achieve 90% power to detect an absolute difference of 15% in the proportions of having alcohol-related harm among male and female students (level of significance of 5%, and non-response rate and confounding control of 15%). Subsequently, 210 students per academic year were randomly selected from the register of medical students for each academic year (provided by the Department of Training and Education at Hanoi Medical University). At this stage, a total of 1260 university students from the 1st to 6th study years were selected. Finally, 1216 students (96.5%) participated in the study; 44 students (3.5%) declined to participate. The age and sex distribution of the non-respondents did not differ from that of the respondents.

Data collection

A letter explaining the aims, assurance of confidentiality, and specification of the date, time, and place to fill in the questionnaire was delivered to the selected students by their class monitors. The investigators and research assistants were trained before data collection.

Each time point of data collection involved a maximum of 30 students, one investigator, and two research assistants. The structured questionnaire and pictures of the most common beverages in Vietnam (with their ethanol levels, and corresponding units for a standard drink) were distributed to the students by the research assistants. The investigators explained the definition of a standard drink (SD) used for this study (see ‘Measures’ below) and instructed students on how to fill in the questionnaire. In addition, the students were assured of their right to withdraw from the study at any time and for any reason. If a student had any query related to the questions, the investigator provided clarification. After the questionnaire was completed, it was handed to the investigator.

Questionnaire

The questionnaires were newly developed to measure alcohol-related harms based on literature and experts’ opinions. First, the questionnaire was pre-tested among 20 students at Hanoi Medical University to ensure that they clearly understood the meaning of all of the questions.

The questionnaire was divided into two parts.

The first part included questions on demographics (age, gender, type of living situation, and academic year level) and on drinking patterns. For example: ‘How often did you drink at least one full SD of alcohol in the previous 12 months?’ and ‘How often did you drink at least four SDs (for females) or five SDs (for males) per occasion in the previous 12 months’.

Responses were made on a 7-point scale: 0=never (recoded to 0); 1=almost daily (recoded to 7); 2=3–4 days per week (recoded to 3.5); 3=1–2 days per week (recoded to 1.5); 4=1–2 days per month (recoded to 0.375); 5=once per month (recoded to 0.25); and 6=less than once per month (recoded to 0.125). Midpoints of categories were used for the recoding (28, 29). To gain more insight into drinking patterns, questions were also asked about ‘How many SDs did you on average consume per occasion?’ and ‘What is the highest number of SDs you have ever consumed in the previous 12 months?’. These types of answers were coded as the actual numbers of SDs consumed.

The second part of the questionnaire included 23 possible alcohol-related harms, categorized into five main types of harm (see Appendix); these were developed from the literature and based on the opinions of experts. For the present study, a set of items within each of the five types of harm was tested for internal consistency (Cronbach's alpha). The five types of harm are ‘negative influence on daily activities’ (Cronbach's alpha=0.78); ‘social conflict’ (Cronbach's alpha=0.67); ‘loss of control, acute consequences, and withdrawal’ (Cronbach's alpha=0.70); ‘mental health condition and physical illness’ (Cronbach's alpha=0.51); and ‘medical health problems’ (Cronbach's alpha=0.53).

In the first four types of harm, respondents were asked to rate the number of harms experienced during the previous 12 months on a 4-point scale (0=never; 1=one time; 2=two times; and 3=at least three times). For the fifth type (‘medical health problems’), respondents were asked whether they had experienced these harms (0=no and 1=yes) during the previous 12 months. The response to each type of harm is the sum of the positive answers to each of the items indicating that type of harm. For the logistic regression, the sum score was computed by recoding each type of harm into two categories: 0=never and 1=yes. In this way, the sum score indicates the variety of different harms experienced by the respondent.

Measures

For the present study a standard drink (SD)=1 can of beer (330 ml at 5%)=1 glass of wine (140 ml at 12%)=1 shot of spirit (40 ml at 40%)=12.6 g of pure alcohol.

Abstainers are students who reported not to drink at least one full SD of alcohol in the previous 12 months. Drinkers are students who reported to drink at least one full SD of alcohol in the previous 12 months. Binge drinkers are students who reported to drink at least 4 SDs (for females) or 5 SDs (for males) per occasion in the previous 12 months.

Data analysis

Analyses were performed with SPSS for Windows (version 15) and STATA (version 10). Cronbach's alpha was calculated to establish the internal consistency of the scales. Descriptive statistics were used to detect differences between male and female students. All other analyses were performed for males and females separately.

Descriptive statistics were used to estimate the frequency and prevalence of alcohol-related harm. Intercorrelations between potential predictors in the multivariate analysis were low (r<0.3), except those between the ‘maximum number of SD consumed’ and the ‘average number of SD consumed’ (r=0.53). Logistic regression was used to compare the drinkers without alcohol-related harm and those with at least one type of harm. Independent variables were entered in two steps: 1) age and type of living situation; and 2) drinking pattern variables. This allowed us to assess the predictive ability of the drinking pattern variables while controlling for the effects of variables in step 1. In turn, the dependent variables representing the five types of harms were entered separately into the model. Results were presented as odds ratios (OR) with 95% confidence intervals (95% CI).

Poisson regression analyses were then conducted to investigate relationships between age, type of living situation, drinking pattern variables, and number of harms. The number of harms is calculated by summing up how many of the five types of harms the students scored positively (score 0–5). In all multivariate analyses, unweighted data were used.

Results

The sample population (N=1216) included a similar number of female (n=606; mean age 20.8 years) and male students (n=610; mean age 20.6 years). Regarding living situation, more male than female students lived in a rented house, whereas more female than male students lived in a dormitory or with a family. Male students were twice as likely to be drinkers than female students (Table 1).


Table 1.  Socio-demographic and drinking behavior characteristics of 1216 medical students in the survey at Hanoi Medical University
Variables Females (n=606) Males (n=610) p Total (N=1216)
Age in years: mean (range) 20.8 (17–26) 20.6 (18–28) >0.05 20.7 (17–28)
Academic year (%)        
  First year 13.5 20.7 <0.001 17.1
  Second year 14.5 19.0   16.8
  Third year 17.2 16.1   16.5
  Fourth year 18.0 15.1   16.5
  Fifth year 19.0 13.9   16.4
  Sixth year 17.8 15.2   16.5
Type of living situation (%)        
  Dormitory 41.7 30.5 <0.001 36.0
  Rented house 24.0 43.2   33.6
  With family 34.3 26.4   30.3
Drinking behavior (%)        
  No 62.3 22.8 <0.001 42.5
  Yes 37.7 77.2   57.5
Sample size varies slightly for each category because of missing values.

Occurrence of harms

Among drinkers, during the previous 12 months male students were significantly more likely to experience harm (81.8%) than female students (60.4%). For male students, the median number of experienced harms was 2, compared with 1 in female students (Table 2).


Table 2.  Drinking patterns and number of harms by gender in a sample of 699 drinkers in the survey at Hanoi Medical University
  Females (n=227) Males (n=466)   Total (N=699)
Variables Median [IQR] Median [IQR] p Median [IQR]
1. Drinking pattern        
  1.1. Frequency of drinking 0.125 [0.125; 0.125] 0.25 [0.125; 0.375] <0.001 0.125 [0.125; 0.25]
  1.2. Frequency of binge drinking 0 [0; 0] 0 [0; 0.125] <0.001 0 [0; 0.125]
  1.3. Average number of standard drinks per occasion 1 [1; 2] 3 [1; 5] <0.001 2 [1; 3]
  1.4. Maximum number of standard drinks consumed 2 [1; 3] 4 [2; 8] <0.001 3 [2; 6]
2. Occurrence of number of harms        
  2.1. Number of types of harms 1 [0; 2] 2 [1; 3] <0.01 1 [0; 2.5]
  2.2. Frequency of number of harms (%)        
   0 harm 39.6 18.2 <0.001 25.1
   1 harm 27.8 24.1   25.4
   2 harms 21.6 26.0   24.5
   3 harms 7.5 21.5   16.9
   4 harms 2.2 8.8   6.6
   5 harms 1.3 1.5   1.4
Sample size varies slightly for each category because of missing values.

The most common type of harm among men and women were ‘loss of control, acute consequences, and withdrawal’ and ‘negative influence on daily activities’. The least common harm among men and women was ‘social conflict’ (Table 3).


Table 3.  Prevalence of type of harm by gender in a sample of 699 drinkers in the survey at Hanoi Medical University
Variables Females (n=227) Males (n=466) p Total (N=699)
Negative influence on daily activities* (%) 29.4 55.8 <0.001 47.1
Social conflict* (%) 3.1 11.7 <0.001 8.9
Loss of control, acute consequences, withdrawal* (%) 51.8 75.6 <0.001 67.8
Mental health condition and physical illness* (%) 11.9 26.6 <0.001 21.8
Medical health problems* (%) 10.1 13.4 >0.05 12.3
Sample size varies slightly for each category because of missing values.
*Percentage of students scoring 1 or more of these items.

Association between specific factors and alcohol-related harm

Among female drinkers, very few significant relations were found (Table 4). Only the frequency of drinking per week and age were predictors of a mental health condition and physical illness. Female students who drank more frequently per week were about six times more likely to have experienced mental health conditions and physical problems. The older female students experienced mental health conditions and physical problems less often (Table 4).


Table 4.  Alcohol-related harm among 228 female drinkers at Hanoi Medical University by socio-demographics and drinking behavior
  Negative influence on daily activities Social conflict Loss of control Mental health condition and physical problem Medical health problems
Model OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Model 1                    
Age 0.9 0.8–1.1 0.6 0.4–1.0 1.0 0.9–1.2 0.7** 0.6–0.9 0.9 0.7–1.1
Living situation                    
  Dormitory vs. with family 1.3 0.7–2.6 na na 1.0 0.5–1.7 1.5 0.5–4.1 2.1 0.7–6.3
  Rent house vs. with family 1.0 0.5–2.3 na na 1.1 0.5–2.3 1.0 0.3–3.8 1.1 0.3–4.6
Model 2                    
Age 0.9 0.8–1.1 0.7* 0.5–0.9 1.0 0.5–2.4 0.7*** 0.5–0.8 0.9 0.7–1.1
Living situation                    
  Dormitory vs. with family 1.4 0.7–2.9 na na 1.0 0.5–1.2 2.0 0.6–6.8 1.7 0.5–5.7
  Rented house vs. with family 1.0 0.4–2.3 na na 1.0 0.5–1.9 1.2 0.3–4.8 0.8 0.2–3.5
Frequency of drinking per week 0.7 0.1–3.2 na na 1.1 0.5–2.5 6.2* 1.2–33.6 na na
Frequency of binge drinking per week 1.9 0.9–4.1 na na 0.0 0.0–3.4 0.6 0.2–1.4 na na
Maximum number of SD consumed 1.1 1.0–1.3 1.1 0.9–1.6 1.1 1.0–1.3 1.1 1.0–1.3 1.1 0.9–1.5
Average number of SD consumed per occasion 1.1 1.0–1.3 0.4 0.0–2.6 1.3 0.9–1.9 1.0 0.9–1.3 0.9 0.7–1.2
OR=odds ratio; CI=confidence interval; SD=standard drinks; *p<0.05; **p<0.01; ***p<0.001.

Among male drinkers, far more significant relations were found. Predictors of different types of harms are the average number of SD consumed per occasion, age, and type of living situation. Among drinking variables, only ‘average number of SD consumed per occasion’ was associated with a negative influence on ‘daily activities’ and ‘medical health problems’. Male students living in a dormitory or in a rented house were more likely to have experienced a ‘negative influence on daily activities’, ‘mental health condition’, and ‘physical problems’ than those living with a family. Only those living in a dormitory were more likely to experience ‘loss of control, acute consequences, and withdrawal’. The older the male students, the more likely they were to have experienced a ‘negative influence on daily activities’, ‘social conflict’, ‘loss of control, acute consequence, and withdrawal’, ‘mental health condition’, and ‘physical/medical problems’ (Table 5). These findings suggest that the relations between socio-demographic variables and harms were not explained by the drinking pattern variables.


Table 5.  Alcohol-related harm among 470 male drinkers at Hanoi Medical University by socio-demographic and drinking behavior
  Negative influence on daily activities Social conflict Loss of control Mental health condition and physical problem Medical health problems
Model OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Model 1                    
Age 1.20*** 1.1–1.3 1.2** 1.1–1.4 1.2* 1.0–1.3 1.1* 1.0–1.3 1.1 1.0–1.3
Living situation                    
  Dormitory vs. with family 2.2** 1.3–3.7 1.2 0.6–2.7 2.4** 1.4–4.4 2.0* 1.1–3.6 0.9 0.4–1.9
  Rented house vs. with family 1.9* 1.1–3.0 1.0 0.5–2.2 1.8* 1.0–3.0 1.9* 1.1–3.5 1.6 0.8–3.2
Model 2                    
Age 1.2** 1.0–1.3 1.2** 1.1–1.4 1.1* 1.0–1.3 1.1* 1.0–1.2 1.1 1.0–1.3
Living situation                    
  Dormitory vs. with family 2.1** 1.2–3.8 1.3 0.4–1.5 2.3** 1.3–4.2 2.0* 1.0–3.6 0.9 0.4–1.9
  Rented house vs. with family 1.8* 1.1–3.0 1.1 0.3–1.7 1.7* 1.0–3.0 2.0* 1.1–3.6 1.6 0.8–3.1
Frequency of drinking per week 1.0 0.7–1.4 1.3 1.0–1.8 1.0 0.6–1.7 1.2 0.9–1.5 1.1 0.8–1.6
Frequency of binge drinking per week 4.3 0.1–248.1 1.3 0.9–1.7 2.1 0.1–29.4 1.2 0.9–1.8 1.1 0.8–1.6
Maximum number of SD consumed 1.0 0.9–1.1 1.0 1.0–1.0 1.0 0.9–1.1 1.0 1.0–1.1 1.0 1.0–1.1
Average number of SD consumed per occasion 1.1* 1.1–1.2 1.0 1.0–1.1 1.1 0.9–1.3 1.0 1.0–1.1 1.1 1.0–1.1
OR=odds ratio; CI=confidence interval; SD=standard drinks; *p<0.05; **p<0.01; ***p<0.001.

Association between specific factors and number of harms

All drinkers who did not experience any harm and experienced at least one harm were included in the analysis. Table 6 shows that age was a significant predictor of the number of harms among female drinkers while age, living away from home, and average number of standard drinks per occasion are significant predictors of the number of harms among male drinkers.


Table 6.  Poisson regression model for number of harms among 228 female and 470 male drinkers at Hanoi Medical University by socio-demographic and drinking behavior
  Female students Male students
Model Coef 95% CI Coef 95% CI
Model 1        
Age −0.074* −0.15–− 0.01 0.07*** 0.03–0.10
Living situation        
  Dormitory vs. with family 0.16 −0.14–0.46 0.27** 0.08–0.45
  Rented house vs. with family 0.09 −0.25–0.45 0.23 0.05–0.41
Model 2        
Age −0.09* −0.02–− 0.02 0.06** 0.03–0.09
Living situation        
  Dormitory vs. with family 0.21 −0.11–0.52 0.25** 0.06–0.44
  Rented house vs. with family 0.04 −0.33–0.41 0.23* 0.05–0.41
Frequency of drinking per week 0.16 −0.05–0.82 0.06 −0.04–0.15
Frequency of binge drinking per week 0.02 −0.19–0.23 0.08 0.01–0.17
Maximum number of SD consumed 0.04 −0.00–0.09 0.01 −0.00–0.02
Average number of SD consumed per occasion 0.05* −0.01–0.11 0.02* 0.01–0.03
CI=confidence interval; SD=standard drinks; *p<0.05; **p<0.01; ***p<0.001.

Discussion

Our findings suggest that alcohol-related harms are common among medical students in Vietnam. The results indicate that female and male students have a similar experience regarding the most common types of harm (‘loss of control, acute consequences, and withdrawal’ and ‘negative influence on daily activities’) and less common types of harm (‘social conflict’) but differ in the prevalence of alcohol-related harms and factors (age, type of living situation, and drinking pattern) that influence harms. Medical students may not differ strongly in terms of drinking from students at other universities in Vietnam. Also, the influence of age and living situation on alcohol-related harm is probably similar for students from other universities.

Our findings are similar to those from other studies worldwide; for example, in New Zealand, the most common negative consequences of alcohol among students are hangover (55%), blackouts (33%), and vomiting (21%) (4), and among students in Australia being sick (12.8%), hangovers (12.3%), and being unable to remember what happened after drinking (10.4%) (21). A study among adolescents in Thailand also indicated that the negative consequences were nausea and vomiting (46.9%), being criticized by someone (38.8%), hangover (37.8%), driving a car or motorcycle after drinking (35.4%), and missing class (32.8%) (13).

In the present study, the prevalence of harm is higher compared with other studies. However, it is difficult to compare the prevalence of alcohol-related harm between studies due to the different measures used. For example, in many studies, each type of harm generally includes only one item, whereas in our study each type of harm included 2–7 items of harm. This means that students in the present study who experienced 1 out of 2–7 items of harm were considered to have alcohol-related harm, leading to a higher prevalence of harm. A Swedish study used a similar measure of collecting data over 12 months and categorized harms into five types (10). This resulted in a prevalence of 43% of them having at least one harm, which is lower than that in Vietnam. Moreover, compared with our students, the prevalence of each type of harm in the Swedish study was also lower, that is, 1) physical health (25.1%); 2) financial situation (12.9%); 3) study or work life (5.5%); 4) family life, marriage, or relationship (1.5%); and 5) friendships or social life (0.9%) (10). The higher negative consequences in our study might be explained by the strength of the alcohol beverages. Medical students in Vietnam who reported drinking home-brewed wine (≥30 proof) or traditional medicinal wine were more likely to develop alcohol problems (5). Other reasons such as coping style, motives and expectancies of drinking, drinking context, and/or drinking location are associated with alcohol-related harms (30) and should be included in future research.

Our results generally confirm that men and women differ in factors that influence alcohol-related harm. In both male and female students, age is a predictor of specific types of harm and of the number of harms. For male students, the higher the age the more harm they experienced. This finding is similar to a study in the United States, showing that being a male student and being older is more likely to lead to alcohol-impaired driving (31). In contrast to men, the older the female student the less harm they experienced. An explanation might be that older female students experienced harm when they were younger and, therefore, they try to minimize such harm when they are older. A study among psychology students in the United States reported that young women were more influenced in their future drinking decisions by their most negative experiences (6).

Type of living situation was the strongest predictor of specific types of harms among male students only. Living with a family seems to be a protective factor, perhaps due to parental control. Our finding is supported by a study among Australian students in which unsupervised drinkers were almost seven times more likely to experience alcohol-related harm than supervised drinkers (21). A study among adolescents in the United States showed that teenagers who are not allowed to drink in school by their parents drink less alcohol during weekends, have a lower frequency of drinking, and tend to experience fewer negative consequences than those who are allowed to drink (20). Another study showed that living in a residence hall led to more opportunities for social activities (such as parties and drinking games), which can lead to alcohol-related harm (32). Thus, intervention programs should take this into account and focus on male students who do not live with a family.

The present study also shows a considerable gender difference in drinking behavior. Male students drank more frequently, engaged in more binge drink, and experienced more alcohol-related harms than female students; this finding is consistent with other studies among students in the United States and Australia (9, 21). The general finding from different countries is that men drink more frequently and in larger quantities per occasion than women, which causes more harm, especially in low-income countries (2). However, in the present study, the gender difference in drinking behavior seems greater than that in other studies, but is consistent with an earlier study among medical students in Vietnam (5). Consistent with other studies, the reported average number of SDs consumed per occasion by men was predictive of specific types of harm. Among women, the reported frequency of drinking was a predictor of their mental health condition and physical problems; this was not found in other studies. It should be noted that the 95% CI (1.2–33.6) is very large for this variable, probably due to the relatively small number of women. This result should be replicated before considering the frequency of drinking to be a definite risk factor for alcohol-related ‘mental health condition’ and ‘physical problems’ among women.

Study limitations

The present study has several limitations. Participants were students drawn from a single university. Future studies should include more socio-demographic and ethnically diverse samples.

Second, the study design was cross-sectional which allows us to examine the association between drinking and related harm, but not the causal links. Third, because answers to the questionnaires were self-reported, this might entail some level of self-report bias and/or incorrect recall of information.

Conflict of interest and funding

The authors have not received any funding or benefits from industry or elsewhere to conduct this study.

References

  1. WHO (2011). Alcohol. Fact sheet. Available from: http://www.who.int/mediacentre/factsheets/fs349/en/index.html [cited 13 February 2012].
  2. WHO (2007). WHO expert committee on problems related to alcohol consumption. Geneva: World Health Organization.
  3. Wechsler H, Davenport A, Dowdall G, Moeykens B, Castillo S. Health and behavioral consequences of binge drinking in college. JAMA 1994; 272: 1672–7. PubMed Abstract | Publisher Full Text
  4. Kypri K, Paschall MJ, Langley J, Baxter J, Cashell-Smith M, Bourdeau B. Drinking and alcohol-related harm among New Zealand university students: findings from a national web-based survey. Alcohol Clin Exp Res 2009; 33: 307–14. PubMed Abstract | Publisher Full Text
  5. Pham DB, Clough AR, Nguyen HV, Kim GB, Buettner PG. Alcohol consumption and alcohol related problems among Vietnamese medical students. Drug Alcohol Rev 2010; 29: 219–26. PubMed Abstract | Publisher Full Text
  6. Park CL. Positive and negative consequences of alcohol consumption in college students. Addict Behav 2004; 29: 311–21. PubMed Abstract | Publisher Full Text
  7. Best D, Manning V, Gossop M, Gross S, Strang J. Excessive drinking and other problem behaviours among 14–16 year old school childern. Addict Behav 2006; 31: 1424–35. PubMed Abstract | Publisher Full Text
  8. Perkins HW. Surveying the damage: a review of research on consequences of alcohol misuse in college population. J Stud Alcohol 2002; Suppl 14: 91–100.
  9. Park CL, Grant C. Determinants of positive and negative consequences of alcohol consumption in college students: alcohol use, gender, and psychological characteristics. Addict Behav 2005; 30: 755–65. PubMed Abstract | Publisher Full Text
  10. Bullock S. Alcohol, drugs and student lifestyle. A study of the attitudes, beliefs and use of alcohol and drugs among Swedish university students. Stockholm: Centre for Social Research on Alcohol and Drugs; 2004.
  11. Cronin CJ. Alcohol use and negative consequences among American college students in West Germany. Int J Addict 1991; 26: 1123–36. PubMed Abstract
  12. Livingston M. Recent trends in risky alcohol consumption and related harm among yougn people in Victoria, Australia. Aust N Z J Public Health 2008; 32: 266–71. PubMed Abstract | Publisher Full Text
  13. Assanangkornchai S, Mukthong A, Intanont T. Prevalence and patterns of alcohol consumption and health-risk behaviors among high school students in Thailand. Alcohol Clin Exp Res 2009; 3: 2037–46. Publisher Full Text
  14. Tho LH, Singhasivanon P, Kaewkungwal J, Kaljee LM, Charoenkul C. Sexual behaviors of alcohol drinkers and non-drinkers among adolescents and young adults in NhaTrang, Vietnam. Southeast Asian J Trop Med Public Health 2007; 38: 152–60.
  15. Kaljee LM, Genberg BL, Minh TT, Tho LH, Thoa LTK, Stanton B. Alcohol use and HIV risk behaviors among rural adolescents in Khanh Hoa Province Vietnam. Health Educ Res 2005; 20: 71–89. PubMed Abstract | Publisher Full Text
  16. Rehm J, Room R, Graham K, Monteiro M, Gmel G, Sempos CT. The relationship of average volume of alcohol consumption and patterns of drinking to burden of disease: an overview. Addiction 2003; 98: 1209–28. PubMed Abstract | Publisher Full Text
  17. Rehm J, Baliunas D, Borges GLG, Graham K, Irving H, Kehoe T, et al. The relation between different dimensions of alcohol consumption and burden of disease: an overview. Addiction 2010; 105: 817–43. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  18. Presley CA, Pimentel ER. The introduction of the heavy and frequent drinker: a proposed classification to increase accuracy of alcohol assessments in postsecondary educational settings. J Stud Alcohol 2006; 67: 324–31. PubMed Abstract
  19. Bellis MA, Phillips-Howard PA, Hughes K, Hughes S, Cook PA, Morleo M, et al. Teenage drinking, alcohol availability and pricing: a cross-sectional study of risk and protective factors for alcohol-related harms in school children. BMC Public health 2009; 9: 380. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  20. Caitlin A, Beau A, Rob T. The impact of parental modeling and permissibility on alcohol use and experienced negative drinking consequences in college. Addict Behav 2009; 34: 542–7. Publisher Full Text
  21. McBride N, Farringdon F, Midford R. What harms do young Australians experience in alcohol use situations? Aust N Z J Public Health 2000; 24: 54–9. PubMed Abstract | Publisher Full Text
  22. Bendtsen P, Johansson K, Akerlind I. Feasibility of an email-based electronic screening and brief intervention (e-SBI) to college students in Sweden. Addict Behav 2006; 31: 777–87. PubMed Abstract | Publisher Full Text
  23. Clement R. Prevelance of alcohol-use disorder and alcohol-related problems in a college student sample. J Am Coll Health 1999; 48: 111–8. Publisher Full Text
  24. Keller S, Maddock JE, Laforge RG, Velice WF, Basler H-D. Binge drinking and health behavior in medical students. Addict Behav 2007; 32: 505–15. PubMed Abstract | Publisher Full Text
  25. Kypri K, Paschall MJ, Langleyb JD, Baxterd J, Bourdeau B. The role of drinking locations in university student drinking: findings from a national web-based survey. Drug Alcohol Depend 2010; 111: 38–43. PubMed Abstract | Publisher Full Text
  26. White AM, Jamieson-Drake DW, Swartzwelder HS. Prevalence and correlates of alcohol-induced blackouts among college students: results of an e-mail survey. J Am Coll Health 2002; 51: 117–31. PubMed Abstract | Publisher Full Text
  27. Xing Y, Ji C, Zhang L. Relationship of binge drinking and other health-compromising behaviors among urban adolescents in China. J Adolesc Health 2006; 39: 495–500. PubMed Abstract | Publisher Full Text
  28. Kuntsche E, Knibbe R, Engels R, Gmel G. Drinking motives as mediator of the link between alcohol experiences and alcohol use among adolescents. J Stud Alcohol 2007; 68: 76–85.
  29. Schelleman-Offerman K, Kuntsche E, Knibbe RA. Association between drinking motives and changes in adolescents's alcohol consumption: a full cross lagged panel study. Addiction 2011; 106: 1270–8. Publisher Full Text
  30. Nolen-Hoeksema S. Gender differences in risk factors and consequences for alcohol use and problems. Clin Psychol Rev 2004; 24: 981–1010. PubMed Abstract | Publisher Full Text
  31. Zakletskaiaa LI, Mundt MP, Balouseka SL, Wilsona EL, Fleminga MF. Alcohol-impaired driving behavior and sensation-seeking disposition in a college population receiving routine care at campus health services centers. Accid Anal Prev 2009; 41: 380–6. Publisher Full Text
  32. Johnson TJ. College students’ self-reported reasons for why drinking games end. Addict Behav 2002; 27: 145–53. PubMed Abstract | Publisher Full Text

*Pham Bich Diep
No. 1 Ton That Tung Street
Dong Da District
Hanoi, Vietnam
Tel: +84 912365666
Email: phambichdiep@hmu.edu.vn; phambichdiep@gmail.com

Appendix

Alcohol-related harm questions.
During previous 12 months, how often have you experienced the 23 following harms after drinking?
(Relative frequency for each of the 23 harms was rated on a 0 to 3 scale: 0=never, 1=one time, 2=two times, 3=at least three times).

Negative influence on daily activities
1 Negative influence on academic study/miss study/miss night duty in hospital 0 1 2 3
2 Negative influence on housework/activities 0 1 2 3
3 Lost interest in daily activities 0 1 2 3
4 Cannot concentrate on doing anything as normal 0 1 2 3
5 Getting sick then having to stop daily activities 0 1 2 3
6 Financial problems 0 1 2 3
Social conflict
7 Negative influence on family member/flat mate 0 1 2 3
8 Negative influence on friend/colleagues 0 1 2 3
9 Lost close friends/lovers 0 1 2 3
10 Accidents/traffic accident/problems 0 1 2 3
11 Fighting with somebody 0 1 2 3
Loss of control, acute consequences, withdrawal
12 Cannot stop once you have started drinking 0 1 2 3
13 Be drunk and passed out 0 1 2 3
14 Have headache/dizziness/nausea 0 1 2 3
15 Unclear speaking/unsteady step 0 1 2 3
16 Cannot remember what happened when you were drinking 0 1 2 3
17 Feeling guilty or remorseful because of drinking 0 1 2 3
18 Try to cut down on alcohol drinking 0 1 2 3
Mental health condition and physical illness
19 Feeling that mental and physical health is reduced 0 1 2 3
20 Having a memory problem even when one does not drink 0 1 2 3
Medical health problems
21 Problems with physical numbness 0 1 2 3
22 Liver problems 0 1 2 3
23 Stomach problems 0 1 2 3

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Road traffic injury among young people in Vietnam: evidence from two rounds of national adolescent health surveys, 2004–2009

Linh Cu Le1* and Robert W. Blum2

1Department of Demography, Hanoi School of Public Health, Hanoi, Vietnam; 2Department of Population and Family Health Science, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA

Abstract

Background: Based on previous data, road traffic injury (RTI) was a leading cause of non-fatal injury in all-age groups in Vietnam, and among the top causes of injury in children and adolescents. Specific analysis on RTIs in young people, however, has yet to be fully investigated. Using the results of two surveys in 2004 and 2009, the present study aims to describe the current situation of non-fatal, unintentional RTIs among Vietnamese youths. In addition, it explores RTI-related risk and protective factors.

Methods: This study utilized the nationally representative Survey Assessment of Vietnamese Youth 2009 (SAVY2) of 10,044 youths aged 14 to 25 from all 63 provinces in Vietnam. The indicators were compared with data from SAVY1 in 2004 of 7,584 youths. Bivariate and multivariable statistical techniques were applied.

Results: Overall, 75% of youths used a motorcycle in SAVY2 compared with 54.2% in SAVY1. Of the SAVY2 sample, the proportion that had experienced an RTI was 10.6% vs. 14.1% in SAVY1. While the proportion of RTIs for both sexes decreased, the decline was greater for males (11.9% vs. 17.8% in SAVY1) than in females (9.2% vs. 10.4%). The proportion of rural youths aged 22–25 who experienced an RTI increased slightly in the 5 years between the two study intervals. The percentage of youths reporting frequent helmet use increased significantly from 26.2% in SAVY1 to 73.6% in SAVY2. Factors related to the likelihood of ever having experienced an RTI included: older age, male, ever being drunk, and ever riding motorcycles after drinking.

Conclusion: While improvements in RTIs appear to have occurred between 2004 and 2009, more attention should be paid, particularly, in maintenance and supervision of law enforcement to helmet use and drunk driving.

Keywords: road traffic injury; adolescent; risk and protective factors; Vietnam

Received: 13 May 2012; Revised: 21 December 2012; Accepted: 21 December 2012; Published: 17 January 2013

Glob Health Action 2013. © 2013 Linh Cu Le and Robert W. Blum. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 18757 - http://dx.doi.org/10.3402/gha.v6i0.18757

 

It is estimated that road traffic crashes kill about 260,000 children and adolescents worldwide under the age of 18 every year. Recognizing the huge burden of such loss, the United Nations launched an initiative in 2011 called a ‘Decade of Action for Road Safety’ with the aim of stabilizing and then reducing global road deaths by 2020 (1). With dramatic economic growth and the improvement of living standards, Vietnam has experienced a rapidly growing number of motor vehicles on its roads in recent years; however, motorcycles still represent the main mode of transport with 26,869,025 motorcycles registered in 2010 (2). The pattern of unintentional injury of adolescents and young adults in Vietnam had previously been periodically studied in several surveys. The first was the Vietnam Multi-center Injury Survey (VMIS) conducted in 2001. That study showed that among young people under 20 years of age in Vietnam, injury accounted for 70% of the burden of disease using the years of potential life lost (YPLL) measurement, compared to 17% due to chronic diseases and only 13% due to communicable diseases (3). Among all injuries, traffic-related fatal injury is the leading cause of death in Vietnam (4). Road traffic injuries (RTI) is also a third leading cause of non-fatal injury in the under 20 age group. Children and adolescents in Vietnam experienced an alarmingly high rate of non-fatal injury with 4,818 episodes per 100,000 population, with a mortality rate of 26.7 per 100,000 inhabitants per year.

Policy implications and recommendations

  • Injury control and prevention efforts will improve significantly if more attention is directed at the maintenance and supervision of law enforcement on helmet use and drunk driving control measures in the promotion of traffic safety behaviors for young people.
  • As injury patterns now show that females are becoming more vulnerable to road-traffic-related injury, no adolescent health policy should overlook this issue in its injury control design.

It is also worth mentioning that from 2001 to 2003 the Ministry of Transport of Vietnam had issued several guidelines stipulating types of roads on which helmet use is mandatory when motorcycling. The circular was then complemented by Resolution No. 13/2002/NQ-CP on traffic safety (5, 6). The government's regulations have forced, to some extent, motorcycle riders to use helmets in the inter-provincial roads and national highways. These regulations, however, did not order helmet use in urban areas. As a consequence, rural residents have tended to produce better adherence than their urban counterparts.

In 2004, the Ministry of Health undertook the first ever comprehensive national survey of youth in Vietnam, covering a sample ranging from 14 to 25 year olds across the country which became known as the Survey Assessment of Vietnamese Youth (SAVY1). SAVY1 data showed that injury most likely takes place on the highway and street, accounting for 59.8% of all injuries (higher in urban: 68.2% vs. 55.8% in rural settings) followed by injury in the home and at the work place (16.7% each). The key findings from SAVY1 suggested that RTI accounted for a vast majority of all injury types and motorcycle injuries predominated. Overall, 54.2% of youths had used a motorcycle as a driver or passenger (7). In 2004, when SAVY1 data were collected, there was no compulsory helmet use requirement in metropolitan areas. SAVY1 data made it clear that legislation was central to helmet use: 51.9% of youths indicated that the law influences their decision compared with 37% who said injury avoidance was a major motivation for helmet use, and only 4% who said road traffic safety education was key. School-based education programs and free helmets were found to have very little influence overall (8).

Recognizing the burden of injury, in June 2007, the Vietnamese Government decreed that by the end of that calendar year all motorcycle drivers and passengers had to wear a helmet on all roads (9). Several studies have been conducted to investigate the impact of this resolution. Some were conducted before and some after the decree in order to assess the changes. Pervin et al. (10) found that the frequency of helmet use in the four study locations ranged from 90 to 99% among adults, but only 15–53% among children 7 years of age or younger, and from 38 to 53% among children aged 7–14. It has been said that legislation to penalize adults whose children do not wear motorcycle helmets has been proposed in Vietnam. However, ongoing advocacy and social marketing efforts should be improved to disseminate information about the safety benefits of helmets to combat erroneous public perceptions (10).

To follow-up results from SAVY1, SAVY2 was conducted 5 years later in 2009. The major findings on SAVY2 have since been reported (11). However, no in-depth analysis has been performed specifically for motorcycle riding, helmet use and prevalence of RTI by demographic characteristics, particularly the risk factors associated with RTI among Vietnamese youths. This paper (1) compares the prevalence of motorcycling, helmet use and prevalence of RTI among youths in Vietnam in 2004 and again in 2009 and (2) determines the risk and protective factors associated with RTI among Vietnamese youths.

Methods

Study sites and sampling strategy

In 2004, the Survey Assessment of Vietnamese Youth (SAVY1) was conducted by the Ministry of Health of Vietnam. The study involved a cluster-sample of 7,584 youths aged 14 to 25 years from 42 provinces (out of 61 provinces) across the country. SAVY2 was then conducted in 2009 as the follow-up. A household sample was used based on the Vietnam Living Standard Survey (VNLSS 2008) sampling frame. SAVY2 was also a multi-stage cluster sample covering all 63 provinces throughout Vietnam. Data collection was from mid-May to the end of June 2009. Similar to SAVY1, young people were invited to come to a central location to complete both an interview and a self-administered survey. Of those invited to participate in SAVY2, 86% agreed, resulting in a final total of 10,044 young people.

Instruments

The SAVY2 instrument was designed to assure comparability with SAVY1 questions and covered a wide range of topics: demographics; education; work; puberty; dating and relationships; reproductive health; HIV/AIDS; injury, illness, and physical health; knowledge/attitudes/beliefs regarding a range of issues; violence; mental health; and mass media and aspirations. The questionnaire was anonymous and contained both interviewer-led and self-completed questions (the latter for more sensitive questions).

Ten specific questions related to unintentional injury, particularly RTI, included 1) Have you had any accident or injury during the last 12 months that required medical treatment? 2) Where did you have that accident/injury? 3) Have you ever ridden a motorcycle? 4) Do you wear a helmet when driving or as a passenger on a motorcycle? 5) In the last 6 months, did you ever travel on a motorcycle without a helmet? 6) Give the main reason that may make you wear a helmet regularly? 7) Have you ever had a road traffic accident because of which you had to take at least 1 day off for treatment? 8) In the last 12 months, did you have a road traffic accident because of which you had to take at least 1 day off for treatment? 9) Have you ever ridden a motorcycle/car after drinking alcohol? 10) Have you ever traveled with a driver who had drunk alcohol? Additionally, respondents were also asked to report their alcohol intake.

Data analysis

The present analysis included socio-demographic characteristics of the respondents (sex, age, ethnic group, place of residence, household economic status), their alcohol and motorbike riding behaviors, and the RTI-related variables of the SAVY2 dataset. The results of SAVY2 were then compared to the findings in SAVY1.

SPSS package version 16.0 was used for data management and manipulation. The individual record in the dataset was weighted to adjust for complex sampling design, making the results nationally representative. Data were first analyzed using univariate and bivariate statistical techniques (with Chi-square statistics). Then, to understand associated factors, multivariable analyses (using binary logistic regression) were performed to identify the strongest predictors of RTI in these young people. In two regression models, a selected set of independent variables were explored: age group (three categories: 14–17, 18–21, 22–25 years old); sex (males vs. females); area of residence (urban vs. rural); ever been drunk (yes/no); ever ridden a motorbike after drinking alcohol; and geographical regions (Red River Delta, North East, North West, North Central, Central Coast, Central Highland, South East, and Mekong River Delta). Of the geographical regions, the most disadvantaged regions are the two mountainous areas in the north (North East and North West) and Central Highland in the south. In contrast, the most urbanized and economically developed regions are Red River Delta (where the capital city of Hanoi is located) and the South East (where the largest and most prosperous city, Ho Chi Minh City, is located).

These multiple logistic regression models were performed in enter mode. The first model predicted the likelihood that the respondent has ever suffered an RTI as the dependent (outcome) variable. The second model analyzed the likelihood that the respondent suffered an RTI in the last 12 months prior to the survey. A chi-square goodness-of-fit Hosmer and Lemeshow test was performed, and an adjusted odds ratio (OR) and 95% confidence interval of OR were reported and compared with the crude OR.

Results

Sample characteristics and overall injury context

Approximately 75% of youths in the SAVY2 sample lived with their biological parents. Among 10,044 respondents, males accounted for 51% (vs. 49% females), and age groups 14–17, 18–21 and 22–25 accounted for 48%, 29% and 23%, respectively. The majority of the sample was unmarried (83%) and, consistent with the population as a whole, 75% resided in rural areas. Similar to SAVY1, SAVY2 asked young people whether they had an injury requiring medical treatment in the previous 12 months; 6.6% of respondents answered yes (significantly lower than the 7.4% in SAVY1, p<0.05). Males had a higher prevalence of injury than females (8.0% vs. 5.2%, p<0.05), but this difference was less than that found in SAVY1 (11.0% male vs. 3.7% female). As in the previous survey, SAVY2 data show that injury most likely takes place on the highway and street. SAVY2 findings show that the gap between road and home or work injury is much greater than previously reported. Specifically, SAVY2 showed that road injury accounted for 73.3% of all injuries (compared to only 59.8% in SAVY1) and was more common among females than males both in urban (75.3% male vs. 85.9% urban female) and rural settings (67.6% male vs. 76.4% rural female).

Motorcycling and helmet use

Overall, 75% of youths used a motorcycle as a driver or passenger increasing from 54.2% in SAVY1 (p<0.01). Not surprisingly, motorcycle use increased with age. Likewise, males were more likely to ride or drive motorcycles than females, and the Kinh – the majority population – was more likely to ride motorcycles than were ethnic minorities. Additionally, SAVY2 showed an increase of motorcycle use in urban areas among both females and males to where now it is nearly universal (see Table 1). Though not quite as universal, in rural areas dramatic increases were seen as well in the years between 2004 and 2009. For example, among rural females the increase went from 45.5% motorcycle use in 2004 to 80.5% 5 years later). Likewise, over the 5-year interval the gap between rural and urban motorcycle use narrowed.


Table 1.  Percent of ever used motorcycle by gender and locationa
  SAVY1 SAVY2
Age groups 14–17 18–21 22–25 14–25 14–17 18–21 22–25 14–25
Urban male 53.9 89.2 95.8 77.4 66.9 95.8 99.0 83.7
Urban female 43.5 72.8 86.8 64.2 50.0 94.4 94.2 74.2
Rural male 40.6 74.3 80.2 59.5 68.9 93.4 94.7 82.0
Rural female 25.6 50.6 45.8 38.4 50.3 79.1 80.5 65.0
All groups 36.6 67.1 70.5 54.2 59.3 88.7 90.4 75.0
aAll percents in SAVY2 were significantly higher than those in similar age group, sex, and location found in SAVY1 (p<0.05).

Turning to helmet use we see a dramatic increase when comparing SAVY1 and 2; and that increase is seen among all age groups. Overall the increase was from 26.2% reported helmet use in SAVY1 to 73.6% helmet use 5 years later. For the oldest age group (22–25 year olds), the increase was from 35.9 to 82.7%.

Helmet use is analyzed by age group and geography, and it is clear that the pattern of use has changed. Specifically, in SAVY1, 22–25-year-old rural males accounted for the highest proportion of helmet use (44.5%), followed by same age urban males (36.4%). This pattern has now changed. In 2009, the helmet use proportion for rural males was more than twice that seen in SAVY1 (62.9% vs. 30.9%), and for rural females the increase was three-fold (78.9% vs. 23.1% in SAVY1). Among urban youth, the changes were even more dramatic. In 2009, 79.8% of urban males reported helmet use compared with 25.7% 5 years earlier. And for females the rate climbed to 89.3% from 21.4% in SAVY1. Urban females are now the group with highest proportion of helmet use – a dramatic shift in 5 years (Fig. 1).

Fig 1
Fig. 1.  Percent of always wearing helmet by gender and location.

Beyond urban and rural changes, there is also another shift in helmet use over the years before and after compulsory helmet use was instituted. While in SAVY1 helmet use was disproportionately higher in the northern provinces, by 2009 that had shifted to the southern regions of the country (e.g. 94.4% in the Mekong River Delta region). This is of significance since the population and motorcycle density is greater in the south (see Table 2).


Table 2.  Percent of always wearing helmet while using motorcycle by age group and regiona
  SAVY1 SAVY2
Age groups 14–17 18–21 22–25 14–25 14–17 18–21 22–25 14–25
Red River Delta (highly urbanized) 19.9 29.8 41.6 28.0 37.6 55.6 70.2 53.2
North East (mountainous region) 34.9 40.4 42.3 38.5 49.5 67.8 78 64.3
North West (mountainous region) 20.3 26.1 26.1 23.2 38.4 57.5 56.9 50.4
North Central 27.4 34.0 41.1 31.8 63.5 76.3 78.2 71.2
Central Coast 16.4 23.5 34.3 22.9 67.1 80.3 87.4 76.6
Central Highland (mountainous region) 11.5 28.1 31.9 20.9 67.7 74.5 77.5 71.8
Southeast (highly urbanized) 11.2 17.6 26.9 17.2 78.5 87.1 93.5 86
Mekong River Delta 15.9 25.6 32.9 23.3 92.5 94.5 96.7 94.4
All groups 19.9 28.2 35.9 26.2 63.9 76.8 82.7 73.6
aAll percents in SAVY2 were significantly higher than those in similar age group, sex, and location found in SAVY1 (p<0.05).

There is another interesting finding when frequency of helmet use is compared with reports of ever having been fined for non-helmet use. While the majority of adolescents and youths indicate that they do not use helmets all the time, those who live in the south – where there is a greater population density – report more consistent use. Also those living in the south report a greater likelihood than their northern peers of being fined for non-helmet use in the previous 12 months (see Table 3).


Table 3.  Multiple logistic regression models to predict the likelihood of having road traffic injury, SAVY2 (2009)
  Model 1: Life time experience of RTI (N=10,034)a Model 2: RTI in the last 12 months (N=10,021)b
Independent variables Crude OR 95% CI of crude OR Adjusted OR 95% CI of adjusted OR Crude OR 95% CI of crude OR Adjusted OR 95% CI of adjusted OR
Gender
  Male 1.335 1.175–1.518 1.297 1.101–1.529 1.489 1.218–1.821 1.442 1.112–1.871
  Female* 1   1   1   1  
                 
Age groups
  22–25 2.869 2.444–3.368 1.905 1.594–2.277 2.354 1.827–3.034 1.508 1.165–1.953
  18–21 2.215 1.892–2.594 1.609 1.359–1.905 2.305 1.813–2.933 1.377 1.038–1.827
  14–17* 1   1   1   1  
                 
Area
  Urban 1.536 1.339–1.763 1.241 1.063–1.448 1.203 0.965–1.498 1.032 0.809–1.317
  Rural* 1   1   1   1  
                 
Had ever been drunk
  Yes 2.486 2.186–2.427 1.723 1.456–2.040 3.056 2.500–3.737 2.137 1.643–2.779
  No* 1   1   1   1  
                 
Had ever ridden motorbike after drinking
  Yes 3.022 2.641–3.457 1.932 1.614–2.312 3.464 2.834–4.235 2.271 1.729–2.982
  No* 1   1   1   1  
                 
Region
  Red River Delta 0.901 0.731–1.111 0.966 0.771–1.209 0.829 0.604–1.139 0.881 0.627–1.239
  North East 0.909 0.715–1.154 0.849 0.662–1.088 0.829 0.575–1.195 0.723 0.496–1.056
  North West 0.690 0.452–1.053 0.692 0.448–1.069 0.876 0.492–1.559 0.784 0.432–1.424
  North Central 0.733 0.569–0.943 0.862 0.664–1.118 0.537 0.353–0.817 0.613 0.399–1.118
  Central Coast 1.324 1.026–1.708 1.266 0.972–1.650 1.210 0.825–1.774 1.103 0.742–1.638
  Central Highland 1.067 0.803–1.419 1.060 0.790–1.422 1.191 0.793–1.787 1.099 0.724–1.670
  South East 1.735 1.423–2.116 1.514 1.229–1.866 1.300 0.957–1.766 1.158 1.842–1.594
  Mekong River Delta* 1   1   1   1  
*Reference category.
aGoodness-of-fit Hosmer & Lemeshow test χ2=4.533; df = 8; p=0.806.
bGoodness-of-fit Hosmer & Lemeshow test χ2=8.204; df = 8; p=0.414.

Road traffic injury and drinking behaviors

The SAVY1 questionnaire did not provide specific questions for identifying RTI prevalence and incidence rates. Instead, questions only asked lifetime traffic injury. These questions were revised in SAVY2. Regarding RTI lifetime injury there is a significant decline over the 5-year interval between SAVY 1 and 2 (10.6% vs. 14.1%, p<0.01). Not surprisingly, the figures in both sexes decreased but the gap between males and females in SAVY2 narrowed as well. This decline is reflected in all regions of the country, for all ages and for both males and females (see Fig. 2).

Fig 2
Fig. 2.  Percent ever had traffic accidents by gender, age group, and location.

In general, 38.8% of those who ever had an RTI reported it occurred in the previous 12 months. The overall rate was 4.1%; recent injury was highest for males, youths from middle-income families, and those between 18 and 21 years of age. However, there was little ethnic difference seen among those who experienced an RTI in the previous year (Fig. 3).

Fig 3
Fig. 3.  Percent ever had traffic injury in the last 12 months by gender, age group, and location, SAVY2 (2009).

Drinking and driving is the great concern in any RTI program. SAVY2 data showed that alcohol use is quite common in young people; about 80% of males and 36.5% of females have drunk a glass of beer/liquor. Among those who have tried alcohol, a third report that they have ever ridden or driven a motorbike after drinking. The age-specific proportion of youths riding a motorcycle after drinking alcohol rose from 19% among the youngest group of males to 68.1% among the oldest. For females it was much lower (4.1% among those 14–17 years and 12.7%, among 22–25 year olds). This experience was more common as the economic status of the young people increased. Drinking and driving appears to be more common in urban settings (35.8% vs. 31.6% in rural areas, p<0.01). Equally worrisome is the practice of riding with a drink driver, which nearly two thirds of males (6.12%) and nearly half of females (42.9%) admit having done.

Factors associated with RTI

Multivariable analyses were performed to identify the factors associated with ever having experienced an RTI both over one's lifetime and in the previous 12 months. The two models summarized in Table 3, consistently showed the following risk factors: male, urban, older, ever been drunk, ever ridden motorcycle after drinking, and those who were from the South East region of Vietnam (Model 1) and in the North Central region (Model 2). Specifically, regarding lifetime experience of RTI, youths in the two older age groups (18–21 and 22–25) were found, respectively, at 1.6 and 1.9 times higher risk for RTIs than those in the 14–17 age group. Men were a 1.3 times higher risk than women. Those who lived in urban settings were at 20% more likely to experience an RTI. Youths who reported ‘ever been drunk’ had an RTI risk almost 1.7 times to that of peers who reported to never having been drunk. Importantly, those who acknowledged having ridden a motorbike after drinking were almost two times more likely to have had an RTI. Regarding regional variations, young people in the South East region (which is the most urbanized and economically developed region) were at 1.5 times higher risk of RTI than their counterparts in the Mekong River delta. A similar pattern was found in Model 2 for the likelihood of an RTI in the last 12 months, except that urban/rural settings were not significantly different and the North Central region reported a higher risk.

Discussion

Motorcycling and helmet use behaviors

With the rapid pace of economic development and urbanization over the 5-year period between 2004 and 2009, young people use motorcycles more frequently today, and the gaps between rural and urban and between males and females have significantly decreased. In 2009, there was no significant difference in the proportion of motorcycle use between urban and rural males and only a 9% difference between urban and rural females compared with more than a 20% difference only 5 years before. Clearly, this relates to the rapid urbanization as well as the improvement of living standards and income in rural areas. On the other hand, this also helps explain the higher risk of traffic-related injury in youth, especially among young urban males. Given that this is cross-sectional data, it is not possible to determine causality; however, there is a strong possibility that the imposition of a helmet law was what accounted for most of the uptake in use. This conclusion is reinforced by the finding from SAVY2 that, more than anything else, the existence of a law was seen by adolescents and youths to be the most influential factor on their behavior. Likewise, the data suggest that the enforcement of the law (as measured by the likelihood of being fined for infractions) may influence use, thereby explaining the regional differences and the highest use rate in the south where fines were also most common. Only future research will be able to determine causal pathways.

The pattern of RTI, alcohol drinking, and other related factors

The percentage of those who have ever experienced an RTI when comparing two rounds of SAVY showed improvement. Overall, the figures in urban settings declined in SAVY2. In contrast, the situation in rural areas showed little change. In fact, the proportion of RTIs in both rural males and females aged 22–25 was slightly higher compared with SAVY1. This pattern probably corresponds to greater motorcycle use in rural regions in 2009 when compared with 5 years earlier.

Ever having been drunk is a significantly associated risk for traffic-related injury (12). Given the fact that adolescent alcohol use in Vietnam is high compared to many countries in developing world (13), it is important to analyze the drink driving issue. In our present analyses, we see a strong association between ever having been drunk and RTI. Again, the data do not allow us to conclude a causal pathway. It may be that the association is because of drinking and driving or it may be that the factors that are associated with alcohol use in adolescence are parallel factors to those that predispose to RTI. Only future analyses will allow for clarification. What we do know from SAVY2 is that there is an increasing trend for youths to have had ridden a motorcycle after drinking. Similarly, 62.1% of males and 42.9% of females admitted that they had ridden with a driver who had consumed alcohol. This alarming finding is comparable to other studies in Vietnam that reported a significant proportion of respondents aged 17 and above (44.9%) were drink and drive (14). Multivariable models in the present analyses confirms the known risk factors (15, 16): males, older age groups, experiencing being drunk and having ridden a motorcycle after drinking, which again highlight the serious issue of drunk driving. These data strongly suggested that interventions aimed at curbing drinking and driving could have a dramatic impact on injury outcomes.

Some limitations of this study should be noted. The cross-sectional nature of both SAVY1 and SAVY2 does not allow us to analyze further the causal relationships among those variables. In addition, drinking alcohol was asked as a ‘life-time experience’ disallowing a time-sequence construction of potentially related behaviors. Additionally, no data were collected to allow for calculation of the amount of alcohol consumption. Thus, it is not possible to know for certain the relationships between the timing of drinking, the amount of alcohol consumed, and driving behavior. Recall may be an additional limitation since SAVY asked about both drinking and RTIs in the past year or ever before. The long recall period may lead to bias. So too, the reporting of both alcohol drinking and helmet use may be subject to over- or under-reporting bias depending on perceived social desirability of respondents. The sampling design of these two rounds of SAVY did not allow us to interview the same individuals or panel of youths over time. Therefore, it is not possible to measure change of behavior for the same individuals over time.

Conclusion

SAVY2 reported a decreased rate of road-traffic-related injury from 5 years previously. As was true in SAVY1, higher injury rates persist in urban areas and in the more populous south of the country. That having been said, the rates of injury appear to be declining especially among urban males. In contrast, there appears to be little change in rural areas. Importantly, SAVY2 data confirmed that alcohol use is strongly associated with risk of experiencing an RTI.

Policy implications

Within the framework of the national policy for accident and injury prevention (which is now in the process of revision and updating) injury control and prevention will improve if attention is paid to enforcement of helmet laws as well as those laws that prohibit the use of alcohol prior to driving. Enforcement should be a high priority for reduction of RTI. Additionally, special attention should be paid to adolescent and young adult males where both alcohol use and driving are the highest. However, broader-based campaigns need also to be planned since our data indicate that there is an increase in RTI among adolescent females.

Acknowledgements

This study was supported by the Ministry of Health of Vietnam. The authors would like to thank the Ministry of Health for their permission to use the SAVY data.

Conflict of interest and funding

The authors have not received any funding or benefits from industry or elsewhere to conduct this study.

References

  1. Tiwari G. Road safety: decade of action with research. Int J Inj Contr Saf Promot 2011; 18: 1–2. PubMed Abstract | Publisher Full Text
  2. Boufous S, Ali M, Nguyen HT, Stevenson M, Vu TC, Nguyen DTY, et al. Child injury prevention in Vietnam: achievements and challenges. Int J Inj Contr Saf Promot 2012; 19: 123–9. PubMed Abstract | Publisher Full Text
  3. Linnan MJP, Cuong V, Le LC, Le PN, Le AV. Report to UNICEF on the Vietnam multi-center injury survey. Hanoi: Hanoi School of Public Health; 2003.
  4. Le AV, Le LC, Pham CV. [Injury: primary findings from the first national injury survey in Vietnam]. Vietnam J Public Health 2004; 1: 18–25.
  5. Ministry of Transport, Guiding the compulsory wearing of helmets [circular no. 08/2001/TT-BGTVT]. Hanoi: Ministry of Transport; 2001.
  6. Ministry of Transport, Guiding the compulsory wearing of helmets [circular number 01/2003/TTBGTVT]. Hanoi: Ministry of Transport; 2003.
  7. Linh LC. Illness, injury, and violence among Vietnamese youth. Specific topic report, SAVY. Hanoi: Ministry of Health; 2006.
  8. Ministry of Health, GSO, UNICEF, WHO. Survey assessment of Vietnamese youth. Hanoi: Ministry of Health; 2005.
  9. Government of Socialist Republic of Viet Nam. Resolution on a number of urgent countermeasures to curb traffic safety and alleviate traffic congestion [Resolution number 32/2007/NQ-CP]. Hanoi: Government of Socialist Republic of Viet Nam; 2007.
  10. Pervin A, Passmore J, Sidik M, McKinley T, Nguyen THT, Nguyen PN. Viet Nam's mandatory motorcycle helmet law and its impact on children. Bull World Health Organ 2009; 87: 369–73. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  11. Ministry of Health, GSO, ADB. Survey assessment of Vietnamese youth 2 (SAVY2). Hanoi: Ministry of Health; 2010.
  12. The World Health Organization. Drinking and driving: a road safety manual for decision-makers and practitioners. Geneva: The World Health Organization; 2007.
  13. Bich TH, Quyng Nga PT, Quang LN, Minh HV, Ng N, Juvekar S, et al. Patterns of alcohol consumption in diverse rural adults populations in the Asian region. Glob Health Action 2009; 2. DOI: 10.3402/gha.v2i0.2017. Publisher Full Text
  14. Tran NT, Bachani AM, Pham VC, Lunnen JC, Jo Y, Passmore J, et al. Drinking and driving in Vietnam: public knowledge, attitudes and practices. Traffic Inj Prev 2012; 13: 37–43. PubMed Abstract | Publisher Full Text
  15. The World Health Organization. Youth and road safety. Geneva: The World Health Organization; 2007.
  16. Novoa AM, Pérez K, Santamariña-Rubio E, Borrell C. Effect on road traffic injuries of criminalizing road traffic offences: a time-series study. Bull World Health Organ 2011; 89: 422–31. PubMed Abstract | PubMed Central Full Text | Publisher Full Text

*Linh Cu Le
Department of Demography
Hanoi School of Public Health
138 Giang Vo Street
Hanoi, Vietnam
Tel: 844 6266 2320
Fax: 844 6266 2385
Email: lcl@hsph.edu.vn; leculinh@gmail.com

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Factors associated with health risk behavior among school children in urban Vietnam

Tran Bich Phuong1*, Nguyen Thanh Huong2, Truong Quang Tien2, Hoang Khanh Chi2 and Michael P. Dunne3,4

1Center for Health and Community Development, Vietnam Union of Science and Technology Associations, Hanoi, Vietnam; 2Faculty of Social Science and Health Behavior, Hanoi School of Public Health, Hanoi, Vietnam; 3Center for Community Health Research, Hue University of Medicine and Pharmacy, Hue, Vietnam; 4School of Public Health, Queensland University of Technology, Brisbane, Australia

Abstract

Background: Health risk behavior among young people is a public health problem in Vietnam. In addition, road traffic injuries are the leading cause of death for those aged 15–29 years. The consequences can be devastating for adolescents and their families, and can create a significant economic burden on society.

Objective: The aim of this study was to identify protective and risk factors that may influence three health risk behaviors among school children: suicidal thinking (ST), drinking alcohol (DA), and underage motorbike driving (MD).

Methods: A cross-sectional survey of 972 adolescents (aged 12–15 years) was conducted in two secondary schools in Hanoi, Vietnam. The schools were purposely selected, one each from the inner city and a suburban area, from which classes (grade 6 to 8) were randomly selected. All students attending classes on survey days took part in the survey. The anonymous, self-completed questionnaire included measures of risk behavior, school connectedness, parental bonding, and other factors. Multivariable regression models were used to examine associations between the independent variables and the three health risk behaviors controlling for confounding factors.

Results: Young people in the inner city school reported a higher prevalence of all three risk behaviors than those in the suburban area (ST: 16.1% [95% confidence interval, or CI, 12.9–19.3] versus 4.6% [95% CI 2.7–6.5], p<0.001; DA: 20.3% [95% CI 16.8–23.8] versus 8.3% [95% CI 5.8–10.8], p<0.001, and MD: 10.1% [95% CI 7.4–12.8] versus 5.7% [95% CI 3.6–7.8], p<0.01). School connectedness and mother and father care appeared to be significant protective factors. For males, bullying in school was associated with suicidal thoughts, whereas for both males and females, school connectedness may be protective against suicidal ideation.

Conclusion: This study supports findings from other nations regarding suicidal thoughts and alcohol use, and appears to be one of the first to examine risk and protective factors forMD. Health promotion within schools should be introduced to improve students’ feelings of connectedness in combination with communication and education campaigns focusing on parental care and engaging teachers for the promotion of safer, supportive school environments.

Keywords: risk behavior; risk factor; protective factor; school children; Vietnam

Received: 30 May 2012; Revised: 12 December 2012; Accepted: 21 December 2012; Published: 18 January 2013

Glob Health Action 2013. © 2013 Tran Bich Phuong et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 18876 - http://dx.doi.org/10.3402/gha.v6i0.18876

 

It is clear from research worldwide that many children and adolescents experience behavior disorders. There are differences in prevalence estimates across studies, but overall 10–20% of children and adolescents in developed and developing countries have mental and behavioral disorders (1). Substance use, alcohol consumption, smoking, fighting, and suicide attempt were all found to be prominent risk behaviors among the young (24). In Asia, the prevalence of major risk behaviors has been found to be high in Bangkok, Thailand (4). Chinese adolescents appear to be at high risk of suicide ideation and suicide attempt (5, 6). Many factors influence risk behaviors in children and adolescents, especially characteristics of families and schools. Support from family equips young people to deal with stressful situations and communication with parents is key in establishing the family as a protective environment (7). Conversely, having poor relationships with parents or experiencing parental divorce or separation is associated with behavioral problems (4). Findings from the Health Behavior in School Aged Children Study (HBSC) show that experiences in school can be crucial to the development of health behavior; in particular, school connectedness is linked with positive health practices (7).

Policy recommendations

  • More attention from government and different sectors in society is required to tackle risk behaviors among Vietnamese adolescents. Financial resources are needed to gather evidence on mental health and risk behaviors among Vietnamese adolescents and the potentially modifiable causal factors. This requires public health agencies to mobilize budgets from national and international sources to facilitate the research.
  • Capacity improvement for health and counseling services, especially at the school level is needed, to help identify health risk behaviors, including ST and behavior among adolescents.
  • Collaboration among education, transportation, police, and local government is needed to prevent underage driving of motorbikes.
  • Develop specific programs to assess, manage, and prevent young people engaging in health risk behaviors. These intervention programs need to incorporate more than one strategy, such as school-based health education, human relationships education, and bullying prevention, alongside parental involvement, and with very high risk youth, social welfare support is also needed.
  • Community awareness-raising strategies with mass organizations (such as the youth union) and youth-focused media should be implemented with rigorous evaluation.

In Vietnam, there is increasing awareness of youth risk behaviors in the media and in public debate. Recently, results from limited research have shown that a relatively large proportion of Vietnamese children exhibit behavioral problems (811). Of the 2,591 adolescents aged between 12 and 18 years who participated in Huong's survey in 2004–2005, 9.2% had thought seriously about suicide in the previous 12 months and 8.7% had consumed alcohol in the past 30 days (10). Nearly half (48.7%) of 515 attempted-suicide patients admitted to the biggest hospital in Hanoi were aged between 15 and 25 (11). Findings from two national population-based surveys, the Survey Assessment of Vietnamese Youth (SAVY) I (2003–2004) and II (2009–2010), show that there was a significant increase in the prevalence of suicidal behavior among Vietnamese adolescents over this time period (from 5.28 to 12.21%). Another serious problem for Vietnamese youth is traffic-related injury, which is a leading cause of death in children and adolescents (12, 13). Seventy percent of the traffic accidents in Vietnam are related to motorcycle crashes, and 88.14% of motorcycle crash-related deaths are due to head trauma (14).

Behaviors established during adolescence often continue into adulthood, eventually resulting in substantial morbidity and mortality (7). Given that youth under 18 years of age account for 30% of the Vietnamese population (15), the mental health and risk behaviors of school children should receive attention and resources from the government and many sectors in society. In this study, we examined the factors that influence three health risk behaviors of school children in Vietnam: (a) thinking about suicide in the past 12 months, suicidal thinking (ST); (b) drinking alcohol (DA) in the past month; and (c) underage motorbike driving (MD) in the past month. Underage driving is a particularly important health risk among Vietnamese adolescents. This study examined these behaviors in two schools located in inner city and suburban Hanoi. The overall purpose was to produce insights that can be used by government and non-government agencies in the design and delivery of programs to prevent harmful behaviors among children and adolescents.

Methodology

Study setting

A cross-sectional survey was conducted in two secondary schools in Hanoi city. One school was in one of the four ancient districts of Hanoi, which is also the national political administration center of Vietnam, and the other was in a suburban district that has an average level of socioeconomic development for Hanoi. The purpose of choosing two schools in different locations was to gain some diversity in school characteristics. The bus is the main means of public transport in Hanoi. It was introduced about a decade ago and has not met the transportation needs of the city. In Hanoi, majority of families have at least one motorbike and use it as their main means of transportation. According to Vietnamese Law, children aged below 18 years are not allowed to drive motorbikes; however, given the availability of motorbikes and the weak enforcement of this law, under-aged driving is common in big cities in Vietnam.

Wine and beer are widely available and often served at family parties or weddings. School children in the study locations could buy wine or beer from shops and small kiosks in the areas very easily and at relatively cheap prices.

Sampling

The study applied one-stage cluster sampling (school class). A total of 972 children aged 12–15 years from the two schools participated in this survey. The schools were purposely selected from the inner city and a suburban area. In the inner-city school, one class consisted of 45–55 children, whereas in the sub-urban school a class consisted of 30–35 children. Thus, 10 out of 30 classes and 15 out of 28 classes were selected from each school for the study, respectively. The total number of classes selected in each school was equally divided across the three grades. Classes were randomly selected within the grades. All students in the randomly selected classes at school on the survey dates took part in the survey.

This study assessed the association between five main independent variables (bullying, parental bonding, school connectedness, inter-parent and sibling conflicts, and self-reported academic achievement) and three health risk behaviors (dependent variables). The questionnaire consisted of 61 items divided into four domains: demographic characteristics, family environment, school environment, and behavior. The following scales and questions were included.

Health risk behaviors

Three questions asked whether the youth had experienced suicidal thoughts, alcohol consumption, or MD. The questions included dichotomous responses (Yes, No) and were adapted from the Youth Risk Behavior Survey (YRBS) developed by the Centers for Disease Control and Prevention in the United States. The YRBS has been widely used in research with adolescents in Asian countries, such as China, Thailand, and also in Vietnam, and has demonstrated itself as a suitable tool (10). The questions ‘Did you drink wine/beer/other alcohol in the past 30 days?’, ‘Did you drive a motorbike (by yourself) in the past 30 days?’, and ‘Did you ever think seriously about suicide in the past year?’

Parental bonding

We used the Parental Bonding Instrument developed by Parker et al. in 1979 (16). This scale has been widely used in many projects and is found to have good reliability and validity (in this study, Cronbach's alpha coefficients for this scale equaled 0.83 and 0.84 for mother and father care, respectively). It consists of 25 items to be completed by the respondent about their relationships with their father and mother separately, assessing two dimensions of perceived parental bonding: ‘care’ and ‘over-protection’.

School connectedness

This scale consists of seven items adapted from the California Healthy Kids Survey 2004 (17). In this study, the Cronbach's alpha was 0.84.

Bullying exposure scale

This scale was developed by the research team based on an extensive literature review of bullying scales that have been used in Vietnam and other countries. The scale consisted of five items about past-month experiences of bullying and used a three-point response format that included ‘No’, ‘Sometimes’, or ‘Often’. Cronbach's alpha for this scale was 0.79, suggesting good internal consistency.

Inter-parent and sibling conflicts

Three questions were asked about parental quarrelling, parental fighting, and sibling conflict using a 4-point scale (Never; Rarely; Sometimes; Often) (10). Academic achievement included a single question, which asked respondents to rate their academic achievement in the previous semester on a 4-point scale (Excellent; Good; Fair; Poor).

Data collection

Anonymous questionnaires were self-completed in classrooms. Before answering the questionnaires, the researchers explained the purpose of the survey and the procedures, and students were asked to focus on their own response without discussion. To protect confidentiality and to ensure standard administration procedures, questionnaires were distributed by researchers in the absence of class teachers. Study participants put the completed questionnaires into sealed envelopes that had been distributed beforehand together with the questionnaires. The questionnaire took 25–30 min to complete and the in-class survey procedure was completed within one normal school session.

Data management

To ensure data quality, the research team screened all returned questionnaires before starting data entry. Data were entered by experienced research assistants and 10% of the returned questionnaires were randomly selected to verify the data entry.

Data analysis

SPSS for Windows version 12 was used to analyze the data. Multivariable logistic regression was conducted to examine factors that may influence the three health risk behaviors separately by gender. The hierarchical forward stepwise method was used to fit four blocks of independent variables to each outcome. The four blocks represented: (1) demographic characteristics that include age, religion, grade, school location, and economic status; (2) family characteristics that include parent marital status, mother education, mother occupation, father education, father occupation, parent alcohol problems, parent drug problems, and emotional support; (3) family variables that include mother care, mother overprotection, father care, father over protection, parental quarrelling, parental fighting, and sibling conflict; and (4) school variables that include school connectedness, academic achievement, and bullying. The forward stepwise selection procedure determines if a variable statistic significantly contributes to the odds of taking risky behaviors. At the end of the analysis procedure, only significant variables remained in the models.

Ethical considerations

This research received approval from the Human Research Ethics Committee of Queensland University of Technology (QUT) in Brisbane, Australia, and the Hanoi School of Public Health, Vietnam. The research was also approved by the Health Departments of the two districts and leaders of the two secondary schools. The study applied a ‘passive parental consent’ method. First, the research team worked with the schools and their parent associations to gain agreement for conducting the survey. With help from the parent associations, study information sheets and consent forms were sent to all parents whose child/children had been selected for the study. Parents who did not respond that they would not allow their children to participate in this study were considered to have agreed to the study. In fact, no parents of the sampled children actively refused.

Results

Description of study participants

Demographic characteristics of the participants are presented in Table 1. Of the 966 students who rated their academic achievement in the last semester, less than 2% reported unsatisfactory performance. With regard to family economic status, a proxy measure related to ownership of means of transportation showed that 12.3% had ‘high economic status’ (family owns a car), 74.1% were at medium level (owned one or more motorbikes), and 13.6% were rated as poor (the family only owned a bicycle). With regard to the family environment, more than 95% of participants reported living with both natural parents. Almost all participants reported that they were Kinh, the most prominent ethnicity in Vietnam. When participants were asked about who they would seek out when they need emotional support, more than a third (36.6%) said they would seek it from friends, 20% from their mother, 14.6% from siblings, and only 4.9% from their father. More than 15% of the participants reported that they would not seek emotional support from anyone (Table 2).


Table 1.  Demographic characteristics of the participants
Characteristics %
Region (n=972)
  Urban 51.7
  Rural 48.3
Sex (n=952)
  Male 51.3
  Female 48.7
Grades (n=972)
  Grade 6 33.3
  Grade 7 36.0
  Grade 8 30.7
Self-reported academic achievement in the school (n=966)
  Distinction 32.8
  Credit 44.9
  Satisfactory 20.5
  Unsatisfactory 1.8
Ethnic group (n=960)
  Kinh 99.5
  Others 0.5
Religion (n=932)
  No 73.9
  Yes 26.1
Family economic status (n=961)
  High 12.3
  Medium 74.1
  Low 13.6


Table 2.  Family characteristics and environment
Characteristics % Characteristics %
Parent marital status (n=965)   Family arrangement (with whom the child currently lives) (n=964)  
  Living together 91.3   Living with both natural parents 95.1
  Divorced 2.7   Living with one natural parent 3.5
  Separated 3.5   Not living with natural parent 1.3
  Death (one or both) 2.5   Number of siblings (n=957)  
Parental education     Alone 8.4
  Mother (n=946)      
  University and college degree 22.2   One 48.6
   Technical/vocational education/high school 22.9   More than one 43.1
   Completed secondary school 22.7 Parent quarrelling (n=965)  
   Completed primary school 32.1   Never 42.9
  Father (n=942)     Rarely 42.6
   University and college degree 25.2   Sometimes 12.8
   Technical/vocational education/high school 20.4   Often 1.7
   Completed secondary school 22.0 Parent fighting (n=967)  
   Completed primary school 32.5   Never 79.9
Parent occupation     Rarely 15.5
  Mother (n=954)      
   Government staff 18.3   Sometimes 3.8
   Self-employed 27.5   Often 0.8
   Farmer 33.0 Sibling conflict (n=959)  
   Housekeeper/unemployment/other 21.2   Never 40.1
  Father (n=956)     Rarely 29.9
   Government staff 24.7   Sometimes 23.1
   Self-employed 34.4   Often 6.9
   Farmer 30.9 Emotional support (Who do you talk to when you need help?) (n=954)  
   Housekeeper/unemployment/other 9.7   Father 4.9
Parent alcohol problems (n=964)     Mother 20.0
  No 89.9   Brother/sister 14.6
  Yes 10.1   Friends 36.6
Parent drug problems (n=963)     Relatives/others 8.5
  No 91.1   No one 15.4
  Yes 0.9    

Prevalence of health risk behaviors by gender, school location, and grade

Between genders, there was one statistically significant difference for ‘consumed alcohol in the past 30 days’, with a larger proportion of males than females consuming alcohol. The differences between the city and suburban schools were also examined. Table 3 shows higher proportions of all risk behaviors in the inner city school compared to the suburban school. Prevalence of driving a motorbike in the past 30 days increased significantly in the higher grades and was highest in grade 8 (about five times more likely than grade 6 and more than two times higher than grade 7) (Table 3).


Table 3.  Prevalence of health risk behaviors among school children by gender, school location, and grade
  Gender School location Grade
Health risk behaviors Male (n=488),% Female (n=465),% Overall sample (%) (Average) City school (n=496),% Suburb school (n=457),% Overall sample (Average),% Grade 6 (n=321),% Grade 7 (n=341),% Grade 8 (n=291),% Overall sample (Average),%
Thought about suicide in the past 12 months 9.9 11.3 10.6 16.1; 95% CI 12.9–19.3 4.6***; 95% CI 2.7–6.5 10.6 9.9 10.6 11.4 10.6
Consumed alcohol in the past 30 days 19.7; 95% CI 16.2–23.2 9.7***; 95% CI 7.0–12.4 14.8 20.3; 95% CI 16.8–23.8 8.3***; 95% CI 5.8–10.8 14.5 13.6 16.3 13.4 14.5
Drove motorbike in the past 30 days 8.5 7.4 8.0 10.1; 95% CI 7.4–12.8 5.7**; 95% CI 3.6–7.8 8.0 2.8; 95% CI 0.9–4.6 6.5; 95% CI 3.9–9.1 15.5***; 95% CI 11.3–19.7 8.0
χ2 test comparing prevalence of health risk behaviors among male and female, among students in City School and Suburban School and among students by grade (**p<0.01; ***p<0.001).

Factors associated with the health risk behaviors

Significant factors (with odds ratio, or OR, and confidence interval, or CI, presented) associated with each health risk behavior are presented separately by gender in Tables 4, 5, and 6.


Table 4.  Multivariable logistic analysis of factors associated with thinking about suicide (male and female)
Variables OR (CI)
Male
  School location 3.9 (1.7–8.7)**
  Father care 0.9 (0.8–1.0)**
  School connectedness 0.9 (0.8–1.0)*
  Bullying 1.2 (1.1–1.4)***
Female
  Economic status **
   High economic status 1.00
   Medium economic status 0.4 (0.2–0.8)
  School location 3.0 (1.2–7.5)**
  Emotional support **
   Mother/father 1.00
   Brother/sister (1) 0.9 (0.8–4.8)
   Friends (2) 1.7(0.6–4.9)
   Relative/others (3) 3.4 (0.8–13.7)
   None (4) 2.5 (0.7–9.5)
Mother care 0.9 (0.8–1.0)**
Father over protection 1.1 (1.0–1.2)***
School connectedness 0.9 (0.8–1.0)***
Note: *p<0.05; **p<0.01; ***p<0.001.


Table 5.  Multivariable logistic analysis of factors associated with drinking alcohol (male and female)
Variables OR (CI)
Male
  School location 2.2 (1.1–4.6)***
  Father education *
  Uni/college degree 1.0
   Technical/vocational/high school (1) 0.5 (0.2–1.0)
   Complete secondary school (2) 0.3 (0.2–0.7)
   Complete primary school (3) 0.6 (0.2–1.5)
Parental alcohol*
  No 1.0
  Yes (1) 1.9 (0.9–4.0)
Father care 0.9 (0.8–1.0)**
Siblings conflict *
  Never 1.00
  Rarely (1) 2.2 (1.1–4.1)
  Sometimes (2) 2.5 (1.2–4.9)
  Usually/often (3) 2.8 (1.1–7.2)
Female
School location 6.3 (2.4–16.6)***
Mother care 0.9 (0.8–1.0)***
Note: *p<0.05; **p<0.01; *** p<0.001.

Thinking about suicide

For males, Table 4 shows that father care, school connectedness, and bullying were associated with ST. Of those, bullying was associated with an increased risk of ST (OR=1.2), whereas father care and school connectedness were associated with a reduced risk. Inner city school location was also found to be associated with high risk of ST. For females, mother care and school connectedness were associated with a reduced risk of suicidal thoughts, whereas father over-protection had the opposite association. High economic status and seeking emotional support outside of the immediate family (parents or siblings) were also associated with ST among females. Notably, female participants who did not receive emotional support at home had higher risk of ST than those who sought emotional support from mother/father or brother/sister. As in the case of males, inner city school location was also associated with a higher likelihood of ST among female participants (OR=3.0).

Alcohol consumption by adolescents

Father care and mother care were both associated with a reduced risk of alcohol consumption among males and females. Sibling conflicts were associated with an increased risk of DA among males. Inner city school location was associated with DA in both genders. Males whose fathers had relatively high education (university or college degree) were at the most risk of DA. Males who reported their parents had an alcohol problem were also at risk of DA themselves.

Underage driving of motorbikes

All adolescents in the survey were under the legal age for driving in Vietnam. For males, mother care was associated with a reduced risk of driving a motorbike. Father occupation was also associated with risk: those males whose fathers were government staff were most likely to report driving motorbikes. Males whose mothers had been educated at technical/vocational/high school, had completed primary school only, or had no education had a higher risk of MD than males whose mothers had a university/college degree. It is clear that the age of male students strongly predicted the risk of underage driving. Among females, mother care was associated with low risk while age was associated with higher risk of driving motorbikes. Female participants who sought emotional support from people other than mother/father or brother/sister were more at risk. Parental alcohol problems also significantly predicted the risk. Inner city school location, again, was associated with the risk of driving motorbikes in females (Table 6).


Table 6.  Multivariable logistic analysis of factors associated with underage driving motorbike (male and female)
Variables OR (CI)
Male
Age 2.8 (1.7–4.6)***
Mother education *
  Uni/college degree 1.0
  Technical/vocational/high school (1) 3.0 (1.0–9.0)
  Complete secondary school (2) 0.4 (0.9–1.4)
  Complete primary school (3) 1.4 (0.4–4.9)
Father occupation *
  Government staff 1.00
  Self-employed (1) 0.6 (0.2–1.5)
  Farmer (2) 0.2 (0.1–0.8)
  Housekeeper/unemployed/others (3) 0.5 (0.1–2.0)
Mother care 0.9 (0.8–1.0)***
Parental fighting *
  Never 1.00
  Rarely (1) 2.5 (1.0–6.4)
  Often/sometimes (2) 0.000
Female
Age 1.9 (1.1–3.3)***
School location 7.0 (2.2–21.0)***
Emotional support from *
  Mother/father 1.00
  Brother/sister (1) 0.5 (0.1–5.6)
  Friends (2) 3.1 (0.9–11.5)
  Relative/others (3) 2.6 (0.4–15.5)
  None (4) 1.7 (0.3–9.1)
Mother care 0.9 (0.8–1.0)***
Note: *p<0.05; ***p<0.001.

Discussion

Recent economic development has enabled Vietnamese children to enjoy better living conditions, but it may also subject them to various risks that negatively influence their health. This study indicates that the prevalence of ST in young people (in the preceding 12 months) is about 11%, and that students in inner city schools are at a significantly higher risk than those in suburban areas (suburban 4.6%; city 16.1%, p<0.001). This difference appears to be remarkably large, but it is quite consistent with several other studies in Vietnam, including the national youth surveys (SAVY I and II) (10, 18, 19), and it reflects a pattern in some other Asian countries (5, 20).

Alcohol consumption (past month) was reported by about 15% of the adolescents in this sample, with significant differences by gender and school location, with highest prevalence among inner city males. The estimate for alcohol consumption is higher than that reported in the 9% found in the Vietnamese study conducted by Huong (10) and a prevalence of 8% found by Choo in Malaysia (20). However, all three studies were consistent in terms of the association among drinking, gender, and location.

This study is the first in Vietnam to estimate the prevalence of MD (past month). We have been unable to find other community-based surveys in Asia that examine family and social characteristics of underage drivers. This risk behavior was more prevalent among older, inner city students, although there was no significant difference between males and females. Our data show that maternal care is protective. However, the behavior was not strongly associated with range of family and school variables included in this study. This strongly suggests the need for further research into potentially modifiable factors that influence underage driving in this environment, especially where there is high risk of injury or death from traffic accidents (13).

Children's mental health and risk behavior are influenced by a wide range of factors. This study indicates that for both males and females, school connectedness and parental care were protective factors. Importantly, father care appeared to influence males and mother care influenced females. The findings are generally consistent with the two national surveys (SAVY I and II). The SAVY analysis revealed that positive family relationships and school connectedness correlate with good mental health among Vietnamese youth (21). A study of 1,432 secondary school children aged 12–16 years, examining the relative contribution of parental bonding and peer victimization at school in Adelaide, Australia, by Rigby, Slee and Martin (2007) found that poor mental health in both male and female students was linked with low mother and father care (22). However, that study did not examine the association between parental care and the risk behaviors such as suicidal behaviors and DA. Data from a cross-sectional household survey carried out in six European countries with 7,740 respondents were similar, showing that father and mother care reduced the risk of suicidal ideation (23).

One previously unobserved finding was that mother care had a positive influence on MD in both sexes. From another angle, the SAVY II survey examined the use of motorized vehicles by youth, specifically after drinking, and found that staying in school may be protective against driving under the influence of alcohol (24). Together, these findings suggest that young people with positive mother care and engagement in school may have reduced risk of traffic-related injury and the social consequences of legal violations.

With regard to the negative correlates of behavioral risk, this study found that, while bullying is linked to ST in males, father over-protection had a negative influence on suicidality in females. These findings are similar to research with school children in Adelaide, Australia (22). They found that for both sexes, mother control and father control correlated with anxiety, social dysfunction, and depression, all common precursors of suicidal ideation, In Hong Kong, Lai and McBride-Chang (25) found that parental over-control was associated with high risk of suicidal ideation in both male and female adolescents.

There are some notable points of difference with prior research. The Vietnamese adolescents’ self-reports of parental drug problems, alcohol problems, and divorce had very weak associations with risk behaviors. Also in Vietnam, the SAVY I & II surveys found that there was no association between having a family member with a history of drug or alcohol problem and youth drinking behaviors. This pattern is very different from trends in Western research, where parental drug abuse is often found to have a strong negative influence on their children's behaviors. The difference might be attributed to cultural differences that lead to a lower prevalence of these risk factors, especially very low prevalence of alcohol abuse by mothers, and the relatively low rate of divorce compared to economically developed countries (15).

Inner city school location had a strong influence on behaviors of the school children, especially for girls. There may be a number of contributing factors. Although no associations were found between self-reported (single item) ‘academic achievement’ and any of the three risk behaviors, media reports and other research in Vietnam (26, 27) have highlighted educational stress to be a significant problem in Vietnam. Pressure to succeed academically has negative effects on the well-being and behavior of young people, especially children living in urban areas (18, 19, 21, 26). In inner urban areas, the pressure from parents, teachers, and society may be higher than that in less developed suburban and rural areas. In many families, educational success, especially admission to university, is perceived as essential for their children's future and family pride. Further investigation of the prevalence, indicators, and effects of this problem is needed.

Health risk behaviors need to be recognized and identified early. Therefore, emphasis should be put on increasing capacity for mental health services and on human resource development in this field, including individualized and group support for students, parents, and teachers. For underage driving of motorbikes, there should be involvement not only from the school and family but also enforcement from the government and different sectors such as the police and transportation authorities. The enforcement should be continuous and strong.

Strengths and limitations of the study

This is one of the few school-based studies on health risk behavior among secondary school children in Vietnam to examine a range of family and school determinants. The interviews were based on standardized measures of behavior and social correlates that have been validated and widely used in similar studies in other countries. The main limitation is that just two schools in Hanoi city participated; they were not randomly selected and therefore the results may not be representative of other cities and rural areas. As with many other studies of this type, all data were gathered by retrospective self-report, so there may have been recall bias that reduced the accuracy of the findings.

Conclusion

This study revealed a number of factors that influence risk behaviors among Vietnamese school students. Findings of importance to potential intervention programs are that school connectedness and high quality parental care have positive influences, whereas bullying, family conflict, and parental over-protection appear to increase risks to health. The problem behaviors were especially common among young people in inner urban areas.

To prevent young people from MD, it is suggested that general population awareness campaigns should emphasize good parental (particularly maternal) communication about traffic safety, focusing on adolescents aged 15 or 16 years. Involvement from the school and family is important, but enforcement by the government and different sectors, such as police and transportation authorities, is needed to prevent school children from driving motorbikes. With regard to prevention of suicide, it is recommended to improve school connectedness among children and implement programs that have been proven in other countries to reduce bullying. These initiatives could be implemented effectively through health promotion approaches in schools. More in-depth study in this field is necessary to expand the evidence base for action that is urgently needed to promote mental health and well-being among young people in Vietnam.

Acknowledgements

This study was supported by a grant from the Ford Foundation. The authors would like to acknowledge Hanoi School of Public Health for supporting this work.

Conflict of interest and funding

The authors report no conflicts of interest. The authors have not received any funding or benefits from industry or elsewhere to conduct this study.

References

  1. WHO. World health report 2001 – mental health: new understanding, New Hope. Geneva: World Health Organization; 2001.
  2. Chen C-Y, Storr CL. Alcohol use and health-related quality of life among youth in Taiwan. J Adolesc Health 2006; 39: 752, e9–e16.PubMed Abstract
  3. Faeh D, Viswanathan B, Chiolero A, Warren W, Bovet P. Clustering of smoking, alcohol drinking and cannabis use in adolescents in a rapidly developing country. BMC Public Health 2006; 6: 169. DOI: 10.1186/1471-2458-6-169. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  4. Puangkanchanasetr S, Plitponkarnpim A, Hetrakul P, Kongsakon R. Youth risk behaviour survey: Bangkok, Thailand. J Adolesc Health 2005; 36: 227–35. Publisher Full Text
  5. Hesketh T, Ding QJ, Jenkins R. Suicide ideation in Chinese adolescents. Soc Psychiatry Psychiatr Epidemiol 2002; 37: 230–5. PubMed Abstract | Publisher Full Text
  6. Dunne MP, Chen JQ, Choo WY. The evolving evidence base for child protection in Chinese societies. Asia Pac J Public Health 2008; 20: 267–76. PubMed Abstract | Publisher Full Text
  7. Currie C, Zanotti C, Morgan A, Currie D, Looze MD, Roberts C, et al. Social determinants of health and well-being among young people. Health behaviour in school-aged children (HBSC) study: international report from the 2009/2010 survey. Copenhagen: WHO Regional Office for Europe; 2012.
  8. Cuong TV. Clinical epidemiology study on mental illnesses in different geopraphical regions in Vietnam. Hanoi: Ministry of Health; 2002.
  9. Hanh VT, Guillemin F, Cong DD, Parkerson GRJ, Thu PB, Quynh PT, et al. Health related quality of life of adolescents in Vietnam: cross-cultural adaptation and validation of the adolescent duke health profile. J Adolesc 2005; 28: 127–46. DOI: 10.1016/j.adolescence.2003.11.016. PubMed PMID: 15683639. PubMed Abstract | Publisher Full Text
  10. Huong NT. Child maltreatment in Vietnam: prevelence and associated mental and physical health problems. Brisbane: Queensland University of Technology; 2006.
  11. Huong TTT, Jiang G-X, Tuong NV, Duc PTM, Rosling H, Wasserman D. Attempted suicide in Hanoi, Vietnam. Soc Psychiatry Psychiatr Epidemiol 2005; 40: 64–71. Publisher Full Text
  12. Torre E. How to reduce traffic fatalities among Vietnamese children. Hanoi: Asia Injury Prevention Foundation (AIPF); 2009.
  13. World Health Organization. Representative office in Vietnam. A leading cause of death and disability in Vietnam; 2009. Available from: http://www2.wpro.who.int/vietnam/sites/dhp/injury/ [cited 19 May 2012].
  14. Huong NL. Data leads to successful introduction of Vietnam Helmet Law 4th IRTAD Conference, 16–17 September 2009, Seoul, Korea.
  15. UNICEF. At a glance: Vietnam – demographics 2012. Available from: http://www.unicef.org/infobycountry/vietnam_statistics.html [cited 7 March 2012].
  16. Parker G, Tupling H, Brown LB. Parental bonding instrument (PBI): Black Dog Institute; 1979. Available from: http://www.blackdoginstitute.org.au/public/research/researchtools.cfm [cited 11 March 2012].
  17. California Department of Education. California Healthy Kids Survey. California: California Safe and Healthy Kids Program office; 2004.
  18. Vietnam Ministry of Health, General Statistics Office of Vietnam, WHO, UNICEF. Survey assessment of Vietnamese youth (SAVY). Hanoi: UNICEF Vietnam; 2005.
  19. Minh THL, Huong TN, Thach TD, Fisher JRW. Experience of low mood and suicidal behaviors among adolescents in Vietnam: findings from two national population-based surveys. J Adolesc Health 2012; 51: 339–48. Publisher Full Text
  20. Choo WY. Child maltreatment in Malaysia: prevalence, correlates, and probable health and behavioural consequences. Brisbane: Queensland University of Technology (QUT); 2007.
  21. Huong NT. Report on mental health of Vietnamese endolescent and youth – survey assessment of Vietnam youth (SAVY 2). Hanoi: Hanoi School of Public Health; 2010.
  22. Rigby K, Slee PT, Martin G. Implications of inadequate parental bonding and peer victimization for adolescent mental health. J Adolesc 2007; 30: 801–12. PubMed Abstract | Publisher Full Text
  23. Heider D, Bernert S, Matschinger H, Haro JM, Alonso J, Angermeyer CM. Parental bonding and suicidality in adulthood. Aust N Z J Psychiatry 2007; 41: 66–73. PubMed Abstract | Publisher Full Text
  24. Liem NT, Nguyen VC, Nguyen HN. Thematic report – using alcohol, beer and cigarette among Vietnamese youth – the analysis result of survey assessment of Vietnamese youth 2003 and 2009 (SAVY). Hanoi: Ha Noi Asia Development Bank, General Office of Population and Family Planning, United Nation Population Fund. 2010.
  25. Lai KW, McBride-Chang C. Suicidal ideation, parenting style, and family climate among Hong Kong adolescents. Int J Psychol 2001; 36: 81–7. Publisher Full Text
  26. Michaelson R. Child Abuse in Vietnam: summary report of the concept nature and extent of child abuse in Vietnam. Hanoi: UNICEF Vietnam; 2004.
  27. Truc TT, Loan KX, Nguyen DN, Dixon J, Sun JD, Dunne MP. Validation of the educational stress scale for adolescents (ESSA) in Vietnam. Asia Pac J Public Health 2012; DOI:10.1177/1010539512440818. PMCID: PMC22500036(2012). Publisher Full Text

*Tran Bich Phuong
Center for Health and Community Development
Vietnam Union of Science and Technology Associations
No 59, Lane 406 Au Co Street
Tay Ho, Hanoi, Vietnam
Email: phuongtb@healthCD.org

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Exploring quality of life among the elderly in Hai Duong province, Vietnam: a rural–urban dialogue

Nguyen Thanh Huong1*, Le Thi Hai Ha2, Nguyen Thai Quynh Chi2, Peter S. Hill3 and Tara Walton4

1Department of Health Policy, Faculty of Social Sciences, Behaviour and Health Education, Hanoi School of Public Health, Hanoi, Vietnam; 2Department of Health Sociology, Faculty of Social Sciences, Behaviour and Health Education, Hanoi School of Public Health, Hanoi, Vietnam; 3School of Population Health, The University of Queensland, Brisbane, Australia; 4Masters of Public Health Candidate (Global Health Stream), Faculty of Health Sciences, Simon Fraser University, Burnaby, Canada

Abstract

Background: Quality of life (QoL) is an important health index for the elderly, necessary for assessing interventions, and prioritising medical and social care needs. As the ageing population in Vietnam continues to increase, understanding important dimensions of QoL for the elderly is essential. There is a paucity of research in this area, however, and the available literature focuses on functional capacities. The purpose of this article is to explore perceptions on the dimensions of QoL among the elderly in Vietnam, to use these perceptions to broaden the concept, and to explore similarities and differences between those living in urban compared to rural areas.

Method: Qualitative methods included in-depth interviews (IDI) with experts in ageing and elderly persons, as well as focus group discussions (FGDs) in three communes in Hai Duong province. IDIs and FGDs were recorded and transcribed. NVivo software was used to analyse the data.

Results: Thematic analysis identified physical, psychological, social, environmental, religious, and economic as important dimensions of QoL. For elderly participants in both urban and rural areas, physical health, social relations, finances and economics, the physical and social environment, and psychological health were reported as important. Rural participants also identified religious practice as an important dimension of QoL. In terms of relationships, the elderly in urban areas prioritised those with their children, while the elderly in rural areas focussed their concerns on community relationships and economic conditions.

Conclusion: Isolating individual factors that contribute to QoL among the elderly is difficult given the inter-relations and rich cross-linkages between themes. Elderly participants in urban and rural areas broadly shared perspectives on the themes identified, in particular social relationships, but their experiences diverged around issues surrounding finances and economics, their respective physical and social environments, and the contribution of religious practice. The study findings may help provide guidance for the development of a socially and culturally relevant instrument for measuring QoL among the elderly in Vietnam. The results will also be useful for developing policies and interventions that are responsive to the needs of the elderly, and reflect the themes perceived to be important.

Keywords: quality of life; elderly; perception; qualitative study; Vietnam

Received: 30 May 2012; Revised: 28 October 2012; Accepted: 1 November 2012; Published: 22 December 2012

Glob Health Action 2012. © 2012 Nguyen Thanh Huong et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2012, 5: 18874 - http://dx.doi.org/10.3402/gha.v5i0.18874

 

Vietnam is in the process of social and economic transition. Social and political change has been profound, opening up the country's socialist economy through the process of đi mới (innovation), and moving away from socialist collectivism towards private enterprise and a market economy. The changes are reflected not only in their rising Gross Domestic Product, but also in almost all aspects of social and economic life (1, 2). With its recent economic growth and effective primary health coverage, life expectancy has increased from 66.5 in 1999 to 72.8 in 2009 (3). People aged over 60 years currently represent around 9% of the total population and are predicted to increase to 13.3% by 2,024 (4, 5).

Policy Recommendations

In order to improve the lives of ageing people in Vietnam there is a pressing need to develop comprehensive policies that cover the aspects identified as important to QoL. The following points summarise the key policy recommendations for the elderly drawn from this study:

  • National policy promotes the development of inte-grated services by setting out a clear vision with the goals of maximising ageing people's QoL.
  • Involvement of older people should be modelled in the way in which policy is developed, monitored, and evaluated. National policy supports integrated approaches that are centred on older people in the way in which they are planned and delivered, and in their quality. This involves supporting innovative approaches that offer choice, flexibility, and control by older people.
  • Improving service system efficiency for service users with complex needs by taking a ‘whole-system’ approach, where services recognise their interdependencies, plan together to provide a comprehensive range of services, establish clear links between these services, and provide ways of tailoring services to local older population.
  • Resource allocation supports the development of balanced service systems and is not directed to acute health care at the expense of prevention, comprehensive primary, and community services.
  • Additionally, it is necessary to have prioritised policies focussing on the different needs of the elderly, such as improving social activities and connection for urban older people while creating mechanisms to provide economic stability for the rural aged.

However, this growth and associated urbanisation has also been associated with an increase in inequity, with a widening rural–urban income gap, as well as growing disparities related to geographic, social, ethnic, and linguistic isolation (6). Urban migration has been one response to these changes, offering opportunities to redistribute income between rural and urban areas (7, 8). On the other hand, however, migration has left the elderly isolated in rural communities without their traditional family support, as the younger generation seeks to take advantage of new opportunities. Family structures have altered, with shifts in the patterns of interdependence, and the elderly face many challenges in the transition because of changes in social structure and value systems, family structures, and living arrangements (8).

The negative synergy between an ageing population and an increasing burden of non-communicable disease has major implications for the health system, for the economy and for the QoL of the aged themselves. Previous studies have examined health needs, utilisation of health services and counselling for the aged, and the influence of lifestyle factors on health in ageing (912). While in urban centres, aged care facilities are now being promoted as an innovative alternative to direct support from the family, options for the elderly living in rural areas are limited, with most currently unable to access health care insurance, and dependent on basic state primary health care facilities (12).

QoL is an important health index for the elderly in every country, playing a key role in assessing interventions, and establishing essential medical and social care needs for the ageing population. Building on the recognition by the World Health Organization (13) that health is not merely the absence of disease but the ‘state of complete physical, mental, and social well-being’, international efforts over the past three decades have led to the development of multiple scales combining objective and subjective elements to measure functional capacity, broader health status, psychological well-being, social support, and the broader concept of QoL (14, 15). With increasing aged populations in most developed countries, understanding QoL of the aged has become a particular area of interest. However, research in the Vietnamese context is currently lacking, though the traditional sayings included in the headings of this article suggest that there is a long history of observation of issues of ageing, relationships, and perspectives on living that inform local debate.

Although the sense of psychological well-being is increasingly seen as influencing self-perception of health, much of the existing research on QoL among the Vietnamese elderly has focussed on the contribution of functional capacity to health. More recently however, there has been some emerging literature addressing the broader concepts of QoL among the elderly in Vietnam (1619). Exploring socio-economic determinants of health, two studies conducted in 2010 examined the association between health-related quality of life (HRQoL) and socio-economic factors among elderly rural populations in Vietnam (16, 17). While these studies used different tools to assess QoL, the results were similar, with both finding inequalities in health status and QoL, specifically noting age, being female, and poor household economic status as being determinants of poor HRQoL. To gain a cross-cultural understanding of QoL, two studies conducted in 2011 compared HRQoL/QoL among the elderly living in rural areas in Vietnam with those in rural areas of Bangladesh (16) and Indonesia (18). Both found significant differences in HRQoL/QoL of the elderly between countries and concluded that socio-cultural differences were likely to account for these. While the results of these recent studies underscore the role of culture and socio-ecological context in influencing health and QoL among the elderly, a deeper understanding of the dimensions of QoL as a function of the local social and cultural realities specific to Vietnam is still needed.

This article describes research undertaken to explore the important dimensions of QoL that reflect the culture, life-experience, and views of the rural and urban elderly in Hai Duong province, Vietnam.

Methods

The study was conducted in Hai Duong province about 60 km Northeast of Hanoi. The province has both urban and rural districts, with rural mountainous areas comprising 11% of the province. Socio-economic development and urbanisation processes are similar to that of other Red River Delta provinces in the North of Vietnam (20).

The study employed qualitative methods to explore and compare the dimensions of QoL for urban and rural dwelling elderly Vietnamese. In the first stage, key informant IDIs were undertaken with five Hanoi-based experts in ageing to seek their opinions on the QoL for the aged. Experts included sociologists, geriatricians, demographers, and sexual and psychological counselors for the elderly who had been identified during the literature review process, and were then contacted by email or phone to invite them to participate. During the interview, participants were asked to discuss their knowledge and experience of QoL research in Vietnam, and other countries, as well as the dimensions they perceived to be important and the factors that may serve to influence them. The aim of the key informant interviews was to obtain the experts’ experience about QoL so that these could later be used for triangulating the findings with the experiences of the elderly collected during the second stage.

In the second stage, IDIs were conducted with male and female elderly participants, who were purposively selected using a maximum variation sample (retirees/farmer/informal sector, urban–rural). Mixed gender FGDs were conducted concurrently to broadly explore the general themes identified from the literature and expert IDIs, while the aim of the elderly IDIs was to explore in greater depth and specificity more sensitive issues such as sexuality, finance, and social relationships. Eligibility for both the FGDs and IDIs required participants to be over age 60 years, physically and mentally capable of participating in an approximately one and a half hour interview.

Heads of the Elderly People's Union were contacted by phone, provided with the eligibility criteria, and then asked to identify potential participants. Eligible individuals were then asked by Heads of the Elderly People's Union to attend a meeting, where they were provided with information about the study, and invited to participate in either an IDI or a FGD.

Three communes were selected representing urban, rural, and rural–mountainous areas of Hai Duong, and two IDIs, and three FGDs (with 8–10 male and female aged persons per group) were conducted in each, a total of six IDIs and nine FGDs overall. As some themes included areas that were at times controversial – particularly in relation to sexuality – use of both IDIs and FGDs provided a mechanism for triangulating the findings. FGDs and IDIs with elderly people were conducted in a conversational manner by experienced interviewers and were tape-recorded and transcribed in the Vietnamese language. The IDIs and the FGDs included similar questions, which asked the participants to discuss their perception and understanding of the term QoL, as well as its potential dimensions, they were also asked to discuss what would make life better or worse for elderly persons, and to compare how gender or location would affect how one values QoL.

Thematic analysis of the literature search and key informant IDIs provided key themes used in the development of the question guides for elderly IDIs and FGDs. Transcripts of the elderly IDIs and elderly FGDs were coded using NVivo software by a group of five experienced researchers, with emergent themes not previously flagged for discussion. Selected transcripts were independently re-coded by a second researcher, to ensure consistency between coders. Reconciliation of codes, including emergent themes, was agreed by consensus of the whole group before all transcripts were finally coded.

Results

Thematic analysis revealed a set of six themes of QoL perceived to be most important for the elderly in Vietnam: physical health, social relationships, finance and economics, physical and social environment, psychological health, and religious practice. These themes along with the sub-themes identified as corresponding with them can be found in Table 1. Descriptions of each theme are provided below, and each of these begins with a Vietnamese idiom, expressions which we found to be consistent with the contents of the particular theme.


Table 1.  QoL themes and sub-themes identified from research findings
No. Themes Sub-themes
I Physical health 1. Body pains
    2. Ability to move around
    3. Tiredness
    4. Dependence on treatment
    5. Ability to work
    6. Ability to serve themselves
    7. Sleep/easy to sleep
    8. Ability to hear and see
    9. Ability to remember
    10. Ability to do housework
II Social relationships 11. Support non-economics for others
    12. Support non-economics from children
    13. Spouse intimacy
    14. Family sentimenta
    15. Role in the family
    16. Community relationshipsa
    17. Relative relationshipsa
    18. Role in the community
    19. Participation in community's activitiesa
    20. Sexual activitya
III Finance and economics 21. Stable income
    22. Economic support from childrena
    23. Economic life is assured
    24. Economic support for others
    25. Economic dependence on childrena
    26. Having favourite food to eata
    27. Daily expensesa
    28. Expenses for community activitiesa
    29. Expenses for health care
    30. Adequate facilitiesa
IV Physical and social environment 31. Physical environment
    32. Housing
    33. Social security
    34. Social service accessibility
    35. Information accessibility
    36. Health care service accessibility
V Psychological health 37. Satisfaction of social relationships
    38. Feeling bored
    39. Satisfaction of family/children
    40. Being respected
    41. Not having anxiety about the after life
VI Religious practice 42. Spiritual belief and religious practice
aEmphasised in urban; bemphasised in rural; the remaining are mentioned in both rural and urban.

Physical health

Khôn đâu tới trẻ, khỏe đâu tới già: ‘Young people are never wise, older people never strong’

Informants considered physical health a higher priority than material assets, linking it to harmony within the family and a sense of purposive community engagement. Access to health services was perceived to be important for maintaining health.

Within urban FGDs, concerns for material needs were minimised, with health issues and access to services seen as more important. Hypertension, heart disease, diabetes, rheumatism, and hearing problems were cited as common, and hearing problems or joint pain presented obstacles to participation in the ‘morning exercise clubs’ – groups meeting in local parks for low impact calisthenics.

The significance of health was not limited to the physical, but to be healthy is the most important, but not only physically, but also spiritually (FGD 2_Urban). In one FGD an older man offered a poem, reminiscing on a life well spent, and the importance of a peaceful soul and a settled mind.

Rural FGDs supported the view of health as having enough to eat, sleeping well, and to have a mind free from worry (FGD 6_Rural). But health within the family also means harmony and the maintenance of tradition, important in a society still influenced by Confucian values1: I care firstly about my family, secondly about keeping family tradition. The family tradition will affect my health: if there is no argument or conflict in the family, then I feel happy, and free from headaches (IDI 4_Male_Rural).

Good health for the elderly was understood by informants to be purposive, the key to their continued usefulness and engagement in rural society. It allows them to remain part of the community, keeping their garden, picking up the grandchildren, and doing the housework. But these roles also come with prescriptive expectations, that their health will enable them to practice behaviours that are ‘a mirror for the younger generations’.

Key informant interviews with experts tended to reflect their focus on the clinical, with arthritis and muscle pain, prostate disease, nocturia, and bladder control identified as key issues affecting QoL. Problems with seeing and hearing were recognised as impacting on social interactions and compromising the confidence of the elderly in engaging in community activities. Experts acknowledged that disease and disability were often accepted by the elderly as a natural part of aging – and not seen to be negatively affecting life.

Social relationships

Con chăm cha không bằng bà chăm ông: ‘The children can't care for their father like his wife can’

The analysis of the social dimension showed the largest disparity in opinion between the elderly informants in both urban and rural areas and experts. Within FGDs, in contrast to expert IDIs, discussion around the theme of social relationships revealed quite prescriptive marital and filial relationships, with clear expectations around roles, and pleasure found for the elderly in the successful lives of their children and in their sense of connectedness to their communities and ancestors. Elderly IDIs gave insight into the tensions between social expectations and the personal experience of individuals coping with the compromises demanded by social change, and in particular the impact of rural–urban migration of the more economically productive family members.

The issue of sexuality divided opinions. The experts – largely clinically oriented geriatricians – insisted that continuing sexual activity is a key element of QoL for the elderly. Urban and rural informants raised the issue in FGDs and in individual interviews, but not approvingly, arguing that the elderly should be ‘models’ for their families, putting those desires behind them and concentrating more on their relationships with their children. All groups saw these inter-generational relationships as crucial to QoL.

When urban participants in the research raised issues related to sexual activity in the elderly, it was to indicate the inappropriateness of the idea. In one case, the opposition mixed implicit concern around the appropriateness of a widowed parent's intention to re-marry a younger woman – with clear sexual implications – and the explicit financial risks to his children's inheritance, with them insisting that he reimburse them for the value of their shared home if he was to re-marry. In another, there are clear prescriptions for what is expected of the elderly – even by their siblings:

I have a brother in Hanoi who calls me continuously asking me to find him a woman who is hard-working and honest for him to marry. I advise him ‘You shouldn't get married again. Your children are all grown up, and you should be a mirror for them to look in and a strong shoulder for them to lean on. [Getting married] would only be to have someone to live with – you wouldn't be up to anything else’. (FGD 3_Urban)

In urban areas, the elderly described tensions between independent living and access and influence in their families’ lives. Older people indicated that they want to live close to their children but not necessarily together. Increasingly, they may live together but not eat together – keeping their own kitchen to accommodate different diets, and more independent living – though sharing meals reasonably regularly. Some multi-generational households persist, though less than in rural areas. The extent to which modern lifestyles challenge ‘traditional’ Vietnamese family values was more evident in urban areas or in rural households where adult children had emigrated to find employment. Respondents reported situations where elderly had to ‘fend for themselves’, where ill parents were not given the immediate attention they needed, where their opinions expressed in community forums were discounted as irrelevant. Most confronting was the refusal to accept the obligation to care for elderly parents:

In our day, people always cared for each other regardless of whether they were poor or rich, but now, they do not. They only feel responsible for their own immediate [nuclear] family, not thinking of their parents. I've heard someone tell their father: ‘I earn money to feed my wife and children. Did you have to feed your father when you were my age?’ (FGD 2_Urban)

At times, the urban elderly were seen to impose their more conservative values and attitudes on their families, explicitly expressing their concerns over the ‘Westernisation’ of children's and grandchildren's life styles. With increasing age, strong traditional desires for a son and grandson to continue the family line are often expressed. In one IDI, a grandparent expressed a desire for their children to break the national ‘two child policy’ in order to secure a male grandson: I don't feel comfortable about my grandchildren. I have two grandchildren but both are girls (IDI 2_Female_Urban).

The issue of sexual activity in the elderly was downplayed in rural interviews and discussions; if anything, this was considered more of an issue in urban areas, linked to a loss of culture:

In the city, their way of living seems to be westernised. In the rural, we think sexual life is not necessary at this age. What important now is how to live healthy and merry. Old people should live pure, should go to pagoda2 to pray for their children and grandchildren. (IDI 1_Male_Rural)

The relationships between elderly parents and their children and grandchildren were seen to be of prime importance for both urban and rural informants. Given the more limited cash economy in rural areas, there was less discussion of how this was expressed financially – more in terms of respect, physical assistance, the links to grandchildren and the need for contact between generations. With closer living more common in rural contexts, tensions in the relationships with sons or daughters-in-law were raised here. Financial independence was exceptional, as the rural elderly were less likely to have pensions, but if their children are rich or they can earn more money to support their parents, the elderly can live comfortably (sướng) (FGD 9_Rural).

But this comfort in living is dependent on harmony within the household, with harmonious relationships between elderly parents creating a positive environment for their children. In contrast, the changing socio-economic environment brings uncertainty – particularly where children migrate to urban centres and are open to the corrupting influences of modern life.

Now we are worried about of our children when they go out of the community to work or study. Social evils are present everywhere. Before, if my children travelled, I knew they would come back, but now, I cannot be sure. The society now is more complex than in our time. (IDI 6_Male_Rural)

Clearly psychological health is interconnected with social relationships, with the financial context for rural elderly dependent on their children an important influence:

My neighbour has four children: two of them are alcoholic, two have gone away to look for work, so my neighbour worries all the time about his children. He can't sleep at night and now he is very sick. (FGD 6_Rural)

Both rural and urban informants spoke of the importance of status in the community and family to their QoL – respect for the old is traditionally important – but increasingly eroded. The elderly valued being able to contribute to debates at home and in the community. Moreover, they expressed concern when young people no longer acknowledged them on the streets or valued their contributions to community meetings.

Despite the constraints that they experienced compared to their urban counterparts, the rural elderly saw their QoL enhanced by their ongoing social networks, their access to community activities such as morning exercise, singing or poetry clubs.

People in rural areas have a sense of neighbourhood. They can just drop-in and chat with their neighbour when they want to. Not like in the city: people close their doors to neighbours. They have no close relationships with their neighbours even though they live next to each other. That's why some people don't feel happy. (FGD 6_Rural)

The sense of community extends beyond their neighbours into a sense of community that extends back through generations. While pagodas tend to be more directly associated with formal religion and religious practice, community temples are erected around the memory of significant community ancestors, providing links through them for the communities that they have shaped, and integral to community identity: Old men go to the community temple; old women go to the pagoda. There they talk – about their community culture and customs (IDI 4_Male_Rural).

Expert interviews, however, based on their research and experience, tended to emphasise marital relationships over other communal relationships. For them, the importance of marital relationships lay in continued sexual engagement, a ‘natural’ part of life, though they conceded that this may not be popularly accepted:

If old people talk about love, or marriage or sex, then people will think they are depraved. I think we need to be more humane – our thinking should be more realistic. If old people have good health then they will have sexual desire. Sexual life is the most important aspect of quality of life. (IDI 2_Expert)

While experts saw the primary social relationship being between elderly marital partners, in common with other informants, they also saw their relationship with their children as significant, with the evidence for the quality of these relationships very much in the public eye:

Respect from children and society is very important for old people. We can actually measure this respect by counting how often the children call their parents or visit them each week, and their willingness to help when older people have problems. (IDI 2_Expert)

The expert panel considered that the increasing distance within families could be offset by relationships in the broader community and that this occurred with more ease in rural areas than in urban areas. In urban areas, even living closely may not ensure satisfying contact: their children may go to work all day and their grandchildren are at school. In the evenings, their children go to English class and their grandchildren go out with friends, and they are still alone (IDI 2_Expert). What they did acknowledge was that the elderly highly valued their right to express opinion or judgement and have it considered by family or community. They keenly sensed the perceived loss of respect for the opinions of the elderly.

Finances and economics

Trẻ cậy cha già cậy con: ‘Children rely on their parents, older people rely on their children’

All groups consulted identified financial independence for the elderly as integral to their QoL, linking this theme to their happiness, their sense of security, their sense of independence, their ability to contribute to public debate and to engage in community life. Both urban and rural elderly expressed their fear of being sick and the need to go to hospital. For many, they do not have financial reserves, and costs are unpredictable – and involve more than just the direct costs. The differences between urban and rural are most acute here – with urban dwellers more frequently able to access government pensions, and rural elderly marginalised in the cash economy. While they contribute in non-financial ways to the economic status of households through child-care and household chores, these contributions are often not formally acknowledged.

Pensions were more likely to be received by the elderly who have worked in government positions in urban areas, and as a result, concerns around financial independence were less frequently raised in urban than in rural interviews:

Almost all the elderly in city have pension so they have relatively sufficient money for living. But almost elderly in the rural are farmers, they do not have pension and they have to live in poorer conditions than in the city. (FGD 2_Urban)

The economic opportunities for the elderly are clearly greater in the city, though in both urban and rural areas situations they make significant household contributions.

For rural respondents, the more constrained financial circumstances are expressed in quite concrete expressions of QoL. Financial dependence, and the lower prevalence of pension support, means the elderly continue to work long past prescribed retirement age. Old people in the urban have pensions. They are happier (sướng) than us. We are farmers. If you ask each one in this room, we have the same life condition (FGD 8_Rural).

Having not worked in paid employment, the rural elderly are ineligible for government health insurance. While the rural elderly do not benefit from pensions, the elderly who continue to farm are still liable for taxes and often need to pay additionally for work they are no longer capable of performing. Their QoL is largely determined by the financial standing of the family.

For both urban and rural, disposable finance is reflected in their diets: food is important culturally and the focus of social interactions. As would be expected in rural communities, the quality of food is clearly important, particularly for those rural informants with a long history in produce. Rural informants were acutely aware of the difference that marginal financial changes make: more money means better quality in food – meat at every meal, rather than occasionally. The lack of financial liquidity impacts in simple ways – such as the simple luxury of a snack to follow morning calisthenics:

In the city, after doing morning exercise, old people can have breakfast with noodles or something else. But old people in the rural areas don't have [can't afford] anything to eat for breakfast. (FGD 6_Rural)

But for some rural elderly, the solutions are simple: I am happy when I have a glass of [herbal tonic] wine to drink every day. If you drink it regularly it keeps you healthy (FGD 9_Rural).

Physical and social environment

Ði hỏi già v nhà hỏi trẻ: ‘Ask your elders before going out; ask children when you come back home’

The physical and social environments reflected concerns around safety – both in terms of the physical environment itself – with its noise, pollution and lack of comfort – and social changes towards more nuclear family living spaces, with less communal interaction. For respondents in rural IDIs and FGDs, there was a greater emphasis on the physical environment, both in terms of increasing risks, and in terms of comfort for living. Interestingly, the social environment issues were mentioned more by urban than by rural, with urban living increasingly globalised, with both Western, but also Japanese and Korean cultures influencing changes in fashion, lifestyle, media and diet. For them, the rapidly changing social environment presented a challenge in terms of locating themselves meaningfully.

In urban centres, the environment in which the elderly now find themselves was the focus of a number of anxieties, with industrialised centres a particular concern. The cost and quality of fruit/vegetable reflected uncertainty of the origin of foods, and concerns about the increased use of pesticides in foods purchased in urban markets were reported by urban respondents and listed as a risk of urban living by their rural counterparts. But uncertainty regarding the influences on the city was not limited to the quality of air, food and water – the social and physical environments interact in ways that can be intimidating for the elderly:

The social and cultural environment is polluted now. I went out and saw not only my grandchildren at home but also many younger people in the street have blue and red hair. I asked them why they do that and they told me I am a ‘backward man’. I feel broken hearted for that. (FGD 1_Urban)

Elderly respondents vocalised their ambivalence around current social trends expressing strong concerns around the influence of the media, which promotes the desire for increasingly consumerist lifestyles. Their persisting socialist commitment contrasts with the more individualist orientation of their grandchildren. For a generation which has lived through post-war austerity, there is concern that the media promotes these ‘modern’ lifestyles without building skills to achieve and sustain them:

The TV doesn't teach young people how to make themselves prosperous—it only shows fighting, dancing and hip-hop and travelling. It shows them how to make themselves seem beautiful, but not how to earn the money so that they can afford beautiful clothes and travel. They learn how to spend a fortune, but not how to make their fortune. In our time, our parents taught us practical things: how to work properly in the fields, how to reduce the weeds for better crops. (FGD 2_Urban)

Some of the elderly, however, recognised that this transformation is grounded in socio-economic changes that the young are better prepared for, and that personal hope for the future has to be linked to confidence in young people: I believe in young people. They have different awareness of politics and culture and they have the knowledge to solve the everyday problems we are facing (IDI 1_Male_Urban).

Rural respondents were acutely aware that their environment had advantages, in particular air quality and the relative quietness compared to urban centres. There is recognition by rural elderly of the material benefits of city life, but the growing economy and remissions from migrant family workers means that much of the advantages of the city are now accessible in the country:

If you compare our place with the city, of course the city is better than here. However, I come to the city many times and I see that some places in the city are very poor. Honestly I think those places are not as good as in the rural area. Here we also have television, refrigerator, and etc. We have fish, pork, beef, and fresh fruit. I think it's even better than the city! (FGD 4_Rural)

As in the cities, however, increasing industrialisation has brought increasing risk – often associated with insecticide contamination of food or water, though sometimes the threats of rural life, real or perceived, persist.

Before, we weren't sick so often. Today, because we are sick so frequently, we have to avoid many things. We used to prepare tea with rain water, it was delicious, but now it tastes tart. The living conditions in rural areas are not comfortable for old people. The bathroom and toilet is not in the house—if they wake during the night, they risk getting sick because of the strong winds outside or even snake bite. (IDI 5_Male_Rural)

The expert group also expressed a similar view on the advantages and disadvantages of physical and social environment on both rural and urban elderly QoL.

Psychological health

Gùng càng già càng cay: ‘The older the ginger, the spicier it is’

Constructs around psychological health exposed the impact of rapid social and economic change on the worldview of the elderly interviewed for this study, and their need for a sense of continued relevance and usefulness in society. The erosion of traditional values concerned some, though the centrality of ‘sướng’ – happiness or contentedness – to their lives, reflected the persistence of deeper values.

In urban centres, the elderly were distressed that their participation in community activities was compromised because of perceptions that their opinions were not seen as relevant. Within local Communist Party Cells, they felt – many of whom had long associations with the political struggle for Vietnam – were perceived to lack currency or value, and were seen as ‘out of touch’. In part this reflected a shift in ideological values, accentuated since the introduction of ‘đi mới’ (innovation):

Lifestyles now are affected by the influence of the West. Young people don't maintain traditions, and the elderly worry about their families and society and worry that they will be abandoned. (FGD 3_Urban)

Rural informants similarly acknowledged that psychological health depended on feeling ‘useful’ and that the current economic situation offered greater opportunities for an active role than ever before. In part, this window of opportunity is attributed to political change with individual efforts contributing directly to family wealth, rather than being absorbed through communal obligations to a state based collective.

However, family was seen as the key to psychological contentment in terms of living positively and securing the future economically. ‘Sướng’ was used often by the elderly to describe this happiness, though the word is more complex than happy, and implies a sense of completeness, of shared joy, of all aspects of life being ‘right’.

Older people should think about their children when they do anything. When the family and their children's family are happy, this is what the older people want to see the most. This will make them feel at ease (sướng). (IDI 5_Male_Rural)

This suong is reflected in the consensus from all groups that the most important is a mind free from worry (FGD 6_Rural). The ‘worry’ reiterated in the statements by experts and both elderly groups certainly included the need for basic needs to be met but emphasised health and evidence of success within the family. Psychological health is not considered an individual attribute, but a shared one, with obligations for the elderly to be models for others in terms of prosperity, but more importantly, harmony:

In general I am satisfied with my life. My family is not rich, but my children are good: they do nothing to make us sad. Not like some families that are rich, but their children cause them headaches! (IDI 5_Male_Rural)

Surprisingly, expert opinion was divided on psychological health. While the need for freedom from worry and care was broadly accepted, some experts challenged the idealised view of the ‘simple life’ of the elderly, making the point that minimal needs must be met before contentment can be considered:

A mind free from worry is very important to old people. If they have a good material life they can then think about entertainment, but if they are living in a poor rural area, the material and spiritual are linked. (IDI 3_Expert)

Religious practice

Trẻ vui nhà, già vui chùa: ‘Young people are happy at home; the elderly are happy at the pagoda’

The theme of religious practice placed emphasis on spiritual practice and personal contentment, shared with the previous themes of psychological health and social relationships. While religious practice was not raised as an issue in expert panel discussions, and only twice in references from urban focus informants, it was spontaneously offered in rural elderly focus groups.

In urban FGDs, going to the pagoda was associated with feelings of contentment, of relief. In traditional thinking for both urban and rural informants, age frees people from their responsibilities and pre-occupations in the home to spend more time in the pagoda, reflecting more on the spiritual aspects of life.

In the rural groups interviewed, spiritual practice was emphasised as important to the essential identity of the elderly. Religious practice offers psychological as well as spiritual benefits. The temple gives a sense of community to older men, a connectedness to their past. While it does have a religious dimension, the communal temple is a social space, a meeting place particularly for older men that also connects them with the history of their community or profession, through reverence offered to their ‘ancestors’. Regular pagoda attendance or church attendance was seen to provide this social contact, but also to offer other transformative benefits. Religious practice was described as having an impact on the elderly themselves, but also transforming lives of others.

For rural respondents – both Buddhist and Roman Catholic – righteous living brings health for themselves and benefits for their families: People need to love each other and those more vulnerable than they are …. If we live well now, the future of our children will be better. If the father eats salt, the children will be thirsty (FGD 4_Rural).

For those informants committed to religion, purity of thought and of intent was integral to their discussion of its transforming impact, often defined in terms of the surrender of sexual desire consistent with popular expectation: You need to be pure when going to the pagoda. Pure means that you need to give up your sexual desire: it is not important to old people (IDI 3_Female_Rural).

Discussion

Using a qualitative approach in the current study, we analysed and triangulated narratives from experts and elderly people living in both urban and rural areas to get an in-depth understanding of the meaning of QoL and to identify the dimensions of QoL that are of importance and relevant for older people in Vietnam. The results from this study reveal six key themes that are common among elderly people in both rural and urban Vietnam when talking about their QoL: physical health, social relationships, finance and economics, physical and social environment, psychological health and religious practice. The findings of our study identified similar themes to a study conducted in rural Bangladesh by Nilsson and colleagues (14), though the detailed expression of these was clearly shaped by local social, cultural and economic factors. Thematic analysis of the data demonstrated the difficulty in isolating factors contributing to QoL, and the complex interdependence of the themes. Rich cross-linkages were apparent among psychological health, social relationships, and finance and economics. Despite this, two issues relating to QoL appear to dominate the responses of both rural and urban respondents: finances and economics, and social relationships.

Financial security appears to be a key concern in a state moving from a socialist economy to a mixed economy under đi mới, with the loss of the previous protection afforded by socialist structures. The economic impact of đi mới was also apparent in the lived experience of the elderly, with urban retirees from (usually State) salaried positions entitled to a State pension, but rural elderly not eligible for these benefits.

Urban elderly report greater access to funding and a higher potential for continued economic activity than their rural counterparts, who suffer the limitations of a subsistence economy, with fewer options for activities, and no financial resources to take advantage of them (8). The inverse relationship between socio-economic status and QoL is well known and has been confirmed in recent comparative studies among the elderly populations of Vietnam and Indonesia as well as Bangladesh (16, 18). Urban migration provides some limited compensation – with the younger generation returning remittances from their urban employment – but at the cost of isolation and familial fragmentation for the rural elderly.

In the accounts of our respondents, rural–urban migration represents a cost that is acutely felt in terms of the second most significant issue for QoL: social relationships. For all informants, it was clear that healthy social relationships – marital, filial, and communal – were crucial to creating the sense of suong – completeness, contentedness – on which QoL for the elderly depended. However, the area of social relationships also produced the most contested issues, particularly in dealing with sexual activity in the aged. While experts valued sexual relationships as a major contributor to QoL, this issue did not emerge in discussion with both rural and urban informants. The tension between prescriptive traditional roles and more liberal trends is evident in the examples offered in the FGDs and supported in IDI. While overt sexual interest was seen as inappropriate to the elderly, and in some way clouded the ‘mirror’ the elderly were meant to hold up to the younger generation, all acknowledged the primacy of marital relationships to QoL. The proverb Con chăm cha không bằng bà chăm ông: ‘The children can't care for their father like his wife can’, suggests that a discreet sexual relation may be implicit in the superior care a wife offers her husband, though this may not be overtly acknowledged. Although there are similar between the perception of elderly and expert group, the discord between these two participants groups clearly exists on social relationship theme. This difference can be explained by strongly influence of Confucian philosophy, social structure, the ways Vietnamese elderly express emotion and somatisation tendency.

The importance of familial relationships, however, remained critical to QoL among both rural and urban participants, with the elderly needing to be both valued by their children, and dependent on their children's happiness and prosperity for their own. Urban and rural informants saw differences between their physical and social environments, their perceptions shaped by their own experience; in finances and economics, inequities between advantaged urban elderly and rural elderly with less access to a cash economy were marked. However, the commonalities in the area of psychological health and social relationships suggest a shared deep valuing of links with family and community. Having a link and role in the family and the community is also emphasised as the most important aspect of QoL of elderly people in rural Bangladesh (21). The elderly's concerns around having a ‘voice’ within the decision making frameworks of their own extended families was also reflected in their concerns around their place within the broader social framework, with rapid social and economic change seeming to easily marginalise them in terms of their social and economic contributions. Despite these concerns, there is sense in which the elderly continue to play a stabilising and supportive role within families, and that they are negotiating their way into new relationships in a rapidly changing world, and that their capacity to maintain their QoL is intimately linked to this. Traditional Vietnamese culture exerts a positive influence on elderly people by providing guidance in dealing with the process of ageing.

Although considerable attention has recently been paid to the elderly in Vietnam, the vast majority of this attention and related research has focussed primarily on the physical health of this population. This study has examined older people's perception of QoL and contributes to increasing and ongoing efforts to better understand the dynamics of ageing. This study also indicates that interdisciplinary studies should be conducted to assess the needs of the diverse elderly population with respect to specific needs related not only health services but also social services, in order to develop policies and identify specific programs at different settings.

Given the importance of QoL among the elderly, it is not surprising that there has been increased attention in assessing this variable, with general instruments validated in ageing populations, as well as the development of specific instruments for the aged (21). However, the search for culturally compatible instruments that maintain links with international understandings of QoL while acknowledging local social and cultural realities, places researchers in a difficult position of deciding whether to adapt existing scales or to develop independent instruments, with experience from Bangladesh, India, Lebanon, Taiwan, and Thailand presenting a range of approaches and solutions (15, 2227). Consequently, the findings from this study may prove useful in guiding the adaption of an existing measure, or, alternatively, in the development of a new culturally relevant tool, specific to the Vietnamese context.

Additionally, this study was undertaken to raise awareness to the issues surrounding QoL of this population and create appropriate changes within elderly care. The results of this study have a number of important policy implications. First, the responses from the elderly participants provide a first-hand account of the multiple dimensions of QoL that are perceived to be most important among this population, perspectives which are invaluable for developing appropriate interventions that are responsive to the needs of this population. The identified issues and implications allows institutional policy makers, administrators, and program planners to better understand the needs of this population, undertake informed policy and planning decisions, and pursue funding allocation in an effort to improve the QoL of the elderly. It is especially important to recognise and address the needs of the ageing population that is projected to increase, making it relevant and pertinent for ongoing gerontological research. Moreover, this information enables participation of the elderly in the policy process, providing them with the chance to actively contribute to society and that will have a benefit on the QoL of the rest of their life. Secondly, the interconnectedness of multiple dimensions that were identified in this study highlight the importance of considering QoL from a broader perspective, and the need to develop comprehensive public health policies that not only encompass the functional capacity, and physical aspects of health but also incorporate psychological, social, environmental, and economical aspects. More importantly, these findings emphasise the need for an intersectoral approach to developing policies for improving the lives of the ageing population in Vietnam. For these public policies, the issues at stake are not only to undertake appropriate changes in the parameters of the welfare system but also to reconsider the balance between public and private initiatives, and notably to thoroughly analyse to what extent future necessary adjustments can be achieved through the play of market forces and to what extent policy intervention is required.

Strengths and limitations

One of the strengths of this study was that the collection of qualitative data provided an opportunity for a more thorough understanding of the similarities and differences between urban and rural populations, contrasting them with the perspectives of experts in the care of the elderly. This triangulation of the views of experts on ageing, in addition to the views of the elderly themselves, helped to provide a more comprehensive understanding of the dimensions of QoL perceived to be important among the elderly.

However, the study findings are limited to the perspectives of elderly individuals living in rural and urban regions of Hai Duong province; and given the social and cultural diversity of this country, national generalization would require careful consideration. Moreover, those living in remote areas may have been missed. Additionally, the results are limited to those who were physically and mentally capable of participating in an IDI or FGD, and as such individuals with disabilities or those who were too sick may not have participated. It is important for future research to explore these missing perspectives, particularly since these individuals represent potentially disadvantaged groups, and needs may differ for these vulnerable populations.

Conclusions

In this study, six themes including physical, social, financial and economic, environmental psychological, and religious, emerged as the important aspects of QoL among both rural and urban elderly in Vietnam. Person–environment interaction is a major variable in the evaluation of QoL in elderly people. Economic aspects and family ties are also important components of QoL. While older populations share many of the elements of QoL across cultures, this research demonstrates clearly how not only socio-cultural issues, but also geographic issues impinge on understandings of QoL. This research has attempted to bridge this gap in Vietnam by providing insights into both rural and urban perceptions of the dimensions of QoL that are most important to them.

The findings are significant for developing a QoL instrument for the elderly that could be used for comprehensive and long-term outcome evaluation of health promotion intervention programs to this population in Vietnam. The suggested instrument would include subjective assessments of the six themes that were identified as relevant for QoL among elderly Vietnamese, taking into account the differences in level of emphasis among these themes. Visual scales could be used to accommodate for variations in literacy, and recall period would be limited to the present or a short period of time to make the instrument easy for elderly individuals to respond to.

Acknowledgements

We would like to express our sincere thanks to Ford Foundation for their financial support, without which this report could not have been done. We are also grateful to the Hanoi School of Public Health for its management support. We would like to acknowledge the anonymous elderly in the three communes and experts who voluntarily participated into this study, without them this study would have been impossible.

Conflict of interest and funding

The authors have not received any funding or benefits from industry or elsewhere to conduct this study.

References

  1. Witter S. ‘Doi moi’ and health: the effect of economic reforms on the health system in Vietnam. Int J Health Plan Manage 1996; 11: 159–72.
  2. Haughton J. Living standards during an economic boom (Vietnam 1993–1998). Geneva: UNDP; 2001, pp. 9–10.
  3. General Statistical Office. The 2009 Vietnam population and housing census: completed results, 2009. Available from: http://www.gso.gov.vn/default.aspx?tabid=512&idmid=5&ItemID=10798
  4. General Statistical Office. VIE/97/P14: population estimation for country, regions and 61 provinces/cities in Viet Nam in the period of 1999–2024. Hanoi: General Statistical Office; 2001.
  5. United Nations Development Programme. Vietnam at a glance. Available from: http://www.undp.org.vn/undpLive/Content/UNDP/About-Viet-Nam/Viet-Nam-at-a-Glance [cited 1 December 2008].
  6. Dang NA. Gender dimension of labour migration in the period of modernization and industrialization of Vietnam. Vietnam Sociol Rev 2005; 2: 23–32
  7. Nguyen PB, Hill PS. The role of temporary migration in rural household economic strategy in a transitional period for the economy of Vietnam. Asian Popul Stud 2008; 4: 57–75. [Crossref]
  8. Bui TC. In the space of social security: a research of the elderly in Vietnam. Hanoi: National University Press; 2005.
  9. Trieu C. Prevalence of using health care and counselling service of elderly in rural area in Hanoi. Hanoi School of Public Health, Hanoi; 2003 (unpublished observation).
  10. Ha HT. Prevalence of elderly with hypertension in Tien Du District, Bac Ninh province. Hanoi School of Public Health, Hanoi; 2004 (unpublished observation).
  11. Thuy NT. Prevalence of elderly with hypertension in Long Bien District, Hanoi. Hanoi School of Public Health, Hanoi; 2007 (unpublished observation).
  12. HSPI. Assessment of health care for the elderly in Vietnam. Hanoi: Health Strategy and Policy Institute, Medical Publishing House; 2007.
  13. WHO. World Health Organization constitution. Geneva: World Health Organization; 1948.
  14. Nilsson J, Parker MG, Kabir ZN. Assessing health-related quality of life among older people in rural Bangladesh. J Transcult Nurs 2004; 15: 298–307. [Crossref]
  15. Bowling A, Banister D, Sutton S, Evans O, Windsor J. A multidimensional model of the quality of life in older age. Aging Ment Health 2002; 6: 355–71. [Crossref]
  16. Nilsson J, Rana AM, Luong DH, Winblad B, Kabir ZN. Health-related quality of life in old age: a comparison between rural areas in Bangladesh and Vietnam. Asia Pac J Public Health 2012; 24: 610–9. [Crossref]
  17. Minh HV, Byass P, Chuc NTK, Wall S. Patterns of health status and quality of life among older people in rural Viet Nam. Glob Health Action 2010; 3. [Crossref] .
  18. Minh HV, Ng N, Byass P, Wall S. Patterns of subjective quality of life among older adults in rural Vietnam and Indonesia. Geriatr Gerontol Int 2012; 12: 397–404. [Crossref].
  19. Hoi LV. Health for community dwelling older people: trends, inequalities, needs and care in rural Vietnam. Umeå University Medical Dissertations. New Series No 1437; ISSN 0346-6612; ISBN 978-91-7459-257-3. Epidemiology and Global Health, Department of Public Health and Clinical Medicine. Umeå University, Umeå, Sweden; 2011.
  20. Hai Duong E-information Portal. Available from: http://www.haiduong.gov.vn/vn/thongtin/Pages/Thongtintongquan.aspx
  21. Nilsson J, Grafstrom M, Zaman S, Kabir NZ. Role and function: aspects of quality of life of older people in rural Bangladesh. J Aging Stud 2004; 19: 363–74. [Crossref]
  22. Saxena S, Chandiramani K, Bhargava R. WHOQOL Hindi: a questionnaire for assessing quality of life in health care settings in India. Natl Med J India 1998; 11: 160–5.
  23. Sabbah I, Drouby N, Sabbah S, Retell-Rude N, Mercier M. Quality of life in rural and urban populations in Lebanon using SF-36 health survey. Health Qual Life Outcomes 2003; 1: 30. [Crossref]
  24. Hwang H-F, Liang W-M, Chiu Y-N, Lin M-R. Suitability of the WHOQOL-BREF for community-dwelling older people in Taiwan. Age Ageing 2003; 32: 593–600. [Crossref]
  25. Ingersoll-Dayton B, Saengtienchai C, Kespichayawattana J, Aungsuroch Y. Measuring psychological well-being: insights from Thai elders. Gerontologist 2004; 44: 596–604. [Crossref]
  26. Ministry of Health. Health Statistics Yearbook 2006. Vietnam: Ministry of Health; 2007, p. 13.
  27. McDowell I, Newell C. Measuring health: a guide to rating scales and questionnaires. Oxford: Oxford University Press; 1996.

*Nguyen Thanh Huong
Faculty of Social Sciences
Behaviour and Health Education
Hanoi School of Public Health
138 Giang Vo, Ba Dinh
Hanoi, Vietnam
Tel: +84 4 62662321
Fax: +84 4 62662385
Email: nth@hsph.edu.vn

Footnotes

1The teachings of the Chinese sage Confucius continue to influence Vietnamese society, with their strong emphasis on moral virtues and normative familial and social relationships that ensure a harmonious society.

2Vietnamese culture broadly differentiates between the pagoda, the communal site for formal religious practice (usually Buddhist, or Taoist) and the community temple, more frequently the focus of respect for communal ancestors.

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

Social capital and mental health among mothers in Vietnam who have children with disabilities

Nguyen Thi Minh Thuy1* and Helen L. Berry2

1Department of Community Based Rehabilitation, Hanoi School of Public Health, Hanoi, Vietnam; 2Faculty of Health, The University of Canberra, Australia

Abstract

Background: Having a child with a disability is a heavy burden for mothers, especially in developing countries, where there is little available financial or other government support. Having a child with a disability is also linked to mental health problems and poor quality of life. Communities rich in social capital and individuals who have high levels of personal social capital generally enjoy day-to-day and long-term health and social benefits but this has not been investigated in Vietnam among mothers of children with disabilities. This study aims to investigate these mothers’ distress in terms of their social capital.

Methods: A cross-sectional study based on an interviewer-assisted survey included 172 mothers of children with moderate/severe disabilities in two provinces of Vietnam (one in the North and one in central Vietnam), using a newly translated and modified version of the Australian community participation questionnaire, several measures of personal social cohesion, and Kessler's 10-item measure of general psychological distress. Hierarchical linear regression modelling was used to explore the relationships among socio-demographic factors, multiple components of structural and cognitive social capita, and mothers’ distress controlling for a wide range of socio-demographic characteristics, the nature of the child's disability, and mothers’ personality (extroversion).

Results: Mothers in this study were highly and multiply disadvantaged, and they had very high levels of distress and low levels of community participation. Furthermore, most forms of participation were associated with greater, not less, distress. Socio-demographic characteristics, child's disability, and mothers’ personality did little to explain variance in mothers’ distress, but types and amounts of participation were important predictors. The final regression model explained 29% of variance in distress, with major contributions made by living in a mountainous area, having a ‘reserved’ personality, and frequency and types of participation.

Conclusion: Vietnamese mothers whose children have disabilities are extremely marginalised and distressed. They have only modest social capital, but the little they have tends to be related to better mental health. Being from the mountains; being ‘reserved’; spending time with friends, neighbours, and in educational activities; and trusting others are related to better mental health among these women. However, several types of participation are associated with worse mental health. Such activities should be avoided in any interventions designed to increase social capital as a mental health promotion strategy.

Keywords: distress; social capital; mothers; children with disabilities; community participation; personal social cohesion; women's health; mental health

Received: 30 May 2012; Revised: 4 December 2012; Accepted: 20 December 2012; Published: 11 February 2013

Glob Health Action 2013. © 2013 Nguyen Thi Minh Thuy and Helen L. Berry. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2013, 6: 18886 - http://dx.doi.org/10.3402/gha.v6i0.18886

 

Family members’ expectations, hopes, and plans, including work and financial arrangements, rest on the expectation of having a ‘normal’ child, not a child with disabilities. The birth of a child with a disability has adverse consequences for the lives, emotions, and behaviours of all family members. These consequences arise because looking after a child with a disability is very difficult and families are not prepared for it (1). Families with children with disabilities experience very high levels of stress and depression (24). Throughout their lives, parents of these children report a wide array of difficult emotions: protectiveness towards someone helpless; revulsion at the abnormal; reproductive inadequacy; child-rearing inadequacy; anger; grief; shock; guilt; and embarrassment (5). The high prevalence of psychological distress these parents endure stems from the interplay of multiple factors. These include child factors, such as age, sex, type of disability, caregiving ‘burden’, the presence of externalising behaviour problems, and emotional disorders. Parental factors, such as personality and coping styles, and social factors, such as marital harmony, social support, and socio-economic circumstances, also play an important part (4).

Policy Recommendations

Results of the study suggested the following recommendations:

  • Vietnamese women with children with disabilities have very poor mental health. Where possible, mental health intervention services should be made freely available to these women.
  • These women also have very limited social contact. Providing a setting in which they could interact socially could help meet their need for social contact and friendship in a non-threatening environment through being among women in similar circumstances. The social contact, support, and opportunity to develop closeness and trust could help buffer their mental health.
  • Educational participation was associated with better mental health among Vietnamese women with children with disabilities. Educational opportunities specially designed for these women could provide educational opportunities as well as offer social contact.
  • Poverty is a real problem among Vietnamese women with children with disabilities. These women, their children and their whole families need financial support to relieve some of the pressure they experience daily. It could also mean that they need to work a little less hard, freeing up a small proportion of their time for mental health-promoting social and educational contact.

The impact of children with disabilities on parental wellbeing

The birth of a normal or abnormal child is viewed as the mother's personal success or failure, respectively. When a child is born with a disability, the mother is often blamed and belittled by society, and false explanations may be attributed to her child's disability, leading to her feeling hopeless about the future (1). Mothers, more so than fathers, are at higher risk of depression because they tend to take active roles in caring for their children with disabilities, even relinquishing their jobs or leisure activities (6). They also report significantly greater anxiety than other mothers (2) and worry about their children's future and ability to find a place in society. Depression, these difficult emotions, and the many hardships they face in addressing their children's problems are associated with low levels of energy and physical activity, affecting their general quality of life (3). Compared with parents of healthy children, parents (especially mothers) of children with disabilities have reported impairment in physical activity, health and social relationships, as well as worse overall perceptions of their quality of life (7).

Family influences on the development of children with disabilities

In addition to child impacts on parents’ wellbeing, family factors, especially mothers’ circumstances, affect the children's development (8). Whitely (9) explored the relationship between family stressors and dysfunctionality and psychosocial adjustment in children with disabilities. Family factors were more predictive of children's psychosocial adjustment than their disabilities. The strain on families, poor maternal physical or mental health, and poverty were the strongest correlates of maladjustment in these children. Mothers’ mental health is particularly important. Compared with other children with disabilities, those with depressed mothers show more behavioural disturbance, poorer cognitive functioning, more insecure attachment, a more difficult temperament, and greater risk for developing depression when older (10).

Despite the evident need for assistance for families with children with disabilities, providing supplementary financial support for these families may not, ultimately, much improve their health and wellbeing. In the United Kingdom, ‘Family Fund’ grants were given to families of children with disabilities, providing them with washing machines, tumble dryers, refrigerators, and telephones. In addition, financial grants were made to purchase items, such as bedding and clothing, or holidays and day outings. The evaluation of these grants showed that the support did not improve the extent to which the disability adversely affected families’ lives (11). Nevertheless, mothers reported a greater sense of wellbeing and fewer symptoms indicative of mental and physical ill-health. The evaluation concluded that it was essential to provide an integrated package of services to help parents cope effectively with the many difficulties associated with caring for their children.

Social capital in the lives of mothers with children with disabilities

To understand how living with children with disabilities affects family wellbeing, it is important to consider the family's broad social context (12). Much of the time that might be available to caregivers for social participation is consumed by caring responsibilities. With the arrival of a child with disabilities, family networks and support often diminish quickly (13). More significantly, people with disabilities and their families endure substantial stigma and social exclusion, leaving them unable to participate in society or to access services equitably (13). Some families choose not to attend community events to avoid confronting the stigma or their own difficult feelings towards those who exclude them (13). These dynamics can lead to a dramatic deficit in social capital for these families.

Social capital comprises two components, community participation and social cohesion (23). A deficit in social capital matters because social capital is widely considered a critical element of public health promotion and a reliable predictor of health, happiness, and life satisfaction (14). It is made up of two separate but connected components: community participation and personal social cohesion (15), also respectively known as the structural and cognitive components of social capital (1618) or what people ‘do’ and what they ‘feel’ (19). The structural component has to do with multiple ways of participating in the community, the networks of association that participation generates, and the quality of relationships (18) within and between those networks. Participation is thought to be linked to cohesion in that greater participation leads to greater cohesion (20), creating a virtuous circle of social capital creation and maintenance. Social capital can be measured at various group levels (such as at neighbourhood or provincial level) or at the individual level. In the latter case, this may be termed ‘personal social capital’ (21). It comprises individuals’ patterns of community participation and their ‘personal social cohesion’ (20, 22). Because personal social capital is strongly protectively related to health, especially to mental health (23), loss of social capital is a substantial concern for these at-risk families.

The stigma and exclusion that follows having a child with disabilities harm these families because they inhibit participation (24), breaking the positive cycle of social capital creation and maintenance. A study of people with disabilities in New York found that deficits in social capital, especially fewer social ties, were more influential than disability or economic circumstances in predicting homelessness (25). A recent review of mothers with children with disabilities reported similar conclusions, finding that participation and social support mediated and predicted health and quality of life (26). The review found that (i) greater social network size, activity, and satisfaction with network were related to greater wellbeing, particularly to life satisfaction and health; (ii) caregiver depression was significantly related to perceived inadequacy of social support; and (iii) social support was protective against declines in physical health, those caregivers who reported higher initial levels of social support showed improved health over time. Indeed, mothers of children with disabilities who participated in support groups often generated new social capital by establishing informal support systems for themselves and their families. They were assertive and expressive, and fairly confident that their actions could make a difference in their children's lives. Non-participating mothers were more depressed, confused, isolated, and overwhelmed. They thought of their children's disabilities as random events or genetic accidents, and could not find meaning in, or make sense of their circumstances (27).

Children with disabilities in Vietnam

Over the past decade, numerous estimates have been made of the prevalence of disability among children in Vietnam. Estimates indicate that approximately 2–6% of the child population have disabilities (28). According to the 1999 national survey, just over 3% of Vietnamese children aged 0–17 had disabilities, nearly 1 million such children, and their situation was poor (29). Almost 50% of school-aged children (6–17 years) with disabilities were illiterate, over one-third had never attended school, and another one-sixth had dropped out of school. About one-third of these children's families had never sought treatment for their disability. Only 5% in urban areas and 10% in rural areas received any form of financial support from the government or the community, such as monthly allowances, free or subsidised education, or free health care cards. This additional disadvantage adds to pressures on these families, especially on mothers, who work very hard in Vietnam (30). Little is known about the social and health circumstances of these mothers, despite their obvious hardship and consequent need for support.

Aims of the study

Participating in the community and a sense of cohesion, of feeling connected to the community (core components of social capital), so influence the lives of mothers of children with disabilities that interventions to improve their quality of life must include elements that enhance their social capital. However, there has not yet been any research into the relationship between social capital, health, and wellbeing in Vietnam among mothers with children with disabilities. Our study aims to (i) explore the relationship between mental health and social capital in this vulnerable group of mothers in Vietnam and (ii) consider whether and, if so, how, social capital might be promoted to enhance their wellbeing.

Method

Respondents and procedure

Respondents included 172 mothers aged 25–65 years (M=40.22 years, SD=8.13, Md=38 years) of children with moderate to severe disabilities living in two provinces in Vietnam, Ninh Binh (in the north) and Quang Nam (central Vietnam). The families came from four districts within these provinces, one remote mountainous, one coastal, and two rural delta districts. The women were identified from a list of children with moderate and severe disabilities provided by Catholic Relief Services, an American NGO, which supports the inclusion of children with disabilities in mainstream community schools. The children's disabilities were screened, initially, through a community-based survey and then confirmed by specialist doctors. There were approximately 220 children with moderate and severe disabilities in the project provinces, and all of their mothers were invited to participate in the study via interviewer-assisted questionnaire. Of the 175 women that attended, three declined to take part, leaving 172, a response rate of approximately 80%.

Procedure

As this is the first time, the measures described below have been used in survey-based epidemiological research in Vietnam, and as the questionnaire was developed in English, we undertook extensive piloting. Two assistant lecturers from the Rehabilitation Department of the Hanoi School of Population Health assisted in translating and piloting the questionnaire, and collected the data. We translated the questionnaire into Vietnamese in July 2007, piloting it the following month with 11 mothers whose children with disabilities were receiving rehabilitation services at the National Institute of Paediatrics in Hanoi. We included the full 63-item version of the Australian community participation questionnaire (‘ACPQ’, see below) on the same persons on different days to identify problems with the questionnaires and its test–retest reliability. The mothers found this tiring because the questionnaire was too long and there were problems with terminology: some of the words, as we had translated them, and contexts were not appropriate in Vietnamese (for example, some items referred to attending ‘church’, though almost all Vietnamese people are Buddhist). Some concepts in the questionnaire were new, too, such as ‘self-efficacy’. Also, mothers of children with disabilities are typically poorly educated and not familiar with completing surveys, making the questionnaire much longer to complete.

We piloted a second, revised and shortened, questionnaire, which included a 30-item short-form of the ACPQ, with 88 in-service students at the Hanoi school of Public Health. The students were asked to complete the questions and to provide written comments on the questionnaire itself. The second questionnaire was more easily answered, but one-quarter of the students commented that some items still did not make sense. For example, items about ‘signing petitions’ or volunteering for ‘non-profit organisations’ were not meaningful in Vietnam where such activities are uncommon. A third, further revised questionnaire was piloted among five Catholic Relief Services staff with proficient English who were experienced in working with families with children with disabilities in Vietnam. Their detailed comments helped us revise the questionnaire to produce a fourth and final version.

In April and May of 2008, the assistant lecturers helped respondents complete the questionnaire at the children's community primary schools. To ensure that respondents could understand the items, the lecturers gave each a copy of the questionnaire to read for 15 min, and then talked about the purpose of the research. The lecturers then read each question aloud, checking whether respondents understood (Rural women in Vietnam are very shy and do not speak out when they do not understand or are confused.). Respondents completed the questionnaire themselves and returned it to the lecturers, waiting for a few minutes so that the lecturers could ensure all questions had been answered. The lecturers collected feedback from the respondents so that we could improve the questionnaire for future studies.

Measures

Mental health

We measured general psychological distress (hereafter, ‘distress’) as a general indicator of mental health. Distress was measured using the Kessler 10-item (K10) (31) measure which taps symptoms of non-specific psychological distress, each scored on a 5-point scale from 1=‘none of the time’ to 5=‘all of the time’. Final summed scores have a possible range of 10–50 with higher scores indicating higher levels of distress. Scores ranged from 14 to 47 in the present sample, with the scale exhibiting a satisfactory degree of internal consistency (Cronbach alpha, hereafter ‘α’,=0.76). The K10 is a widely used and validated measure, including in Australia, where Vietnamese ancestry is around 0.7% of population (32) and where Vietnam is among the top five countries of birth (other than Australia).

Social capital

Community participation

We measured frequency and breadth of community participation, and perceptions about participation. We used the 30-item short-form of the ACPQ (20), which taps 14 types of participation: contact with household members, extended family, friends, neighbours, social contact with workmates, organised community activities, religious observance, adult learning, volunteering, giving money to charity, interest in current affairs, expressing opinions, community activism, and political protest. Each item is answered on a 7-point scale from 1=‘never, or almost never’, to 7=‘always, or almost always’. We computed mean scores for each type of participation, with higher scores indicating more frequent participation.

We also computed a breadth of participation index as described elsewhere (20). Briefly, using a multiple regression analysis in which mean scores for all 14 types of participation were entered simultaneously, controlling for key socio-demographic characteristics (see below) and with distress as the dependent variable, types of participation were removed from the equation one-by-one until only significant predictors of distress remained. These items were then dichotomised by median split (to take account of skew in some of the distributions of scores). A score of 1=‘participator’ was assigned to those scoring at or above the median and 0=‘non-participator’ to those scoring below it. These scores were summed to create an index (M=6.1, Md=6.0, SD=3.2). The index displayed an acceptable level on internal consistency (α=0.74), with higher scores indicating greater breadth of participation.

Participation perceptions

We measured perceptions (thoughts and feelings) about seven types of community participation that were independently associated with distress in a previous study (20). These were: contact with household members, extended family, friends and neighbours; organised community activities; religious observance; and interest in local affairs. Respondents indicated whether they considered that they participated too little or too much in these types of participation (‘thoughts’), and whether they enjoyed each of them (‘feelings’). Each response was measured on a 7-point scale from 1=‘definitely disagree’ to 7=‘definitely agree’. Mean scores were calculated (M=4.9, Md=5.0, SD=0.92 for ‘too little’ and M=4.8, Md=5.0, SD=1.1 for ‘enjoying’), each demonstrating acceptable internal consistency (0.63 for ‘too little’ and 0.76 for ‘enjoying’ participation).

Elements of personal social cohesion

Social trust, which is trust in people in general rather than in known others, was measured using the 12-item short-form of the Organisational Trust Inventory as adapted for use in the general population (33, 34). Three 4-item sub-scales tap separate dimensions of trust: believing that most people (i) avoid taking excessive advantage of others; (ii) try to negotiate honestly; and (iii) are reliable. Each item is scored on a 7-point scale from ‘definitely agree’ to ‘definitely disagree’. Final mean scores for each sub-scale, and for the full scale, range between 1 and 7, with higher scores indicating higher levels of trust. Mean scores were calculated with the values of 4.7, 5.1, and 4.4 (Md=5.0, SD=0.7; Md=5.25, SD=1; and Md=4.5, SD=0.7 for ‘take advantage’, ‘negotiate honestly’, and ‘reliable’, respectively). The sub-scales and full scale exhibited a lesser degree of internal consistency than found previously in the Australian studies, with α-values ranging from 0.38 to 0.53.

Generalised reciprocity, the belief that people tend to help each other out without expecting immediate repayment, was measured using the item from the World Values Survey (35). As used, it was ‘generally speaking, would you say that, most of the time, people try to be helpful, or are they mostly looking out for themselves?’ The item is scored 1=‘people try to be helpful’, 0=‘people are mostly looking out for themselves’, with higher scores meaning greater sense of generalised reciprocity (M=4.9, Md=4.5, SD=1.1). As this is a single-item measure, α cannot be calculated.

Optimism was measured using the Life Orientation Test (36). It includes six items tapping trait (underlying, rather than situation-specific) optimism and includes three positively and three negatively worded items. Consistent with the trust measure, each optimism item was scored on a 7-point scale from 1=‘definitely disagree’ to 7=‘definitely agree’. Final average scores for the scale range between 1 and 7, with higher scores indicating greater optimism (M=4.9, Md=4.83, SD=0.8). The scale exhibited a lower degree of internal consistency than has been found in Australian samples (37), with α= 0.35.

Sense of belonging and tangible support were measured using two sub-scales of the interpersonal support evaluation list (38). The sub-scales assess the degree to which respondents perceive themselves to be included in a social group (sense of belonging) and have people available to offer practical and material support when needed (tangible support). Each sub-scale contains 10 items scored 1=‘yes’ or 0=‘no’. Total summed scores for each sub-scale range between 0 and 10 with higher scores representing higher levels of belonging (M=6.1, Md=6, SD=1.16) and tangible support (M=6.5, Md=6, SD=1.04). With α= 0.20 for belonging, this sub-scale demonstrated poor internal coherence, while, with α=0.53, tangible support displayed only modest internal coherence.

Social cohesion

We defined personal social cohesion, using the measures above, as a combination of respondents’ sense of belonging, generalised reciprocity, social trust, confidence (self-efficacy), and hope for the future (optimism). We also included tangible social support because it is related to personal social capital and mental health among women (34), including mothers with children with disabilities (see introduction).

Socio-demographic characteristics

Respondents reported their sex, age, level of education, ethnic origin, responsibility for dependent persons (under 18 and over 60 years), being in paid work at least 6 h per week, living alone, and having a government poverty benefit card. They also provided information about their role in the family (primary income-earner and/or caregiver), the condition of their housing (from good to very poor: brick wall and tiled roof; flat house, storeyed house, or other), housing tenure, location of residence (mountains, coast or rural), and financial problems in the last 12 months (such as being unable to pay the rent, selling something to get money, going without a meal, or being unable to afford warm clothes in winter) as well as ownership of each of the seven basic household appliances (television, bicycle, motorbike, car, home telephone, mobile telephone, and washing machine). We asked respondents about family income and the size of family land. However, they could not answer these questions because most were subsistence farmers who rarely measure their land, or they measure it in an unconventional manner (for example, a metric metre equals about 14 metres among these farmers).

Extroversion

We controlled for extroversion in this study because extroverted people tend to be more sociable than introverts and, therefore, may participate more in their communities. Because participation is a core part of social capital, this may mean that extroverts have more social capital than introverts. We measured extroversion using the two extroversion items from a 10-item very brief measure of the ‘Big Five’ personality domains (39).

Analytic strategy

We start by presenting descriptive statistics for this sample. Relationships between the independent (‘predictor’) and dependent variables were assessed, in the first instance, by examining zero-order (unadjusted) Pearson product-moment correlations. We then examined partial correlations adjusted for all other variables (including socio-demographic characteristics) derived from multiple regression analyses with distress as the dependent variable. The data met the assumptions for the various statistical tests.

Hierarchical linear regression analyses were employed to evaluate the independent contribution made by each predictor variable to explaining variance in distress. Variables were added in blocks. Blocks 1–3 included the control variables and blocks 4–6 the social capital variables, entered in the order suggested by social capital theory (in which participation is hypothesised to generate greater cohesion). Socio-demographic control variables were entered in block 1, type of children's disability (mental, physical, or multi-disability) in block 2, and extroversion in block 3. Block 4 introduced the first of the predictor variables, frequency of participation, followed by perceptions of participation in block 5, and personal social cohesion in the final block, block 6. Changes in standardised beta values from one block to the next were examined to assess the degree of potential mediation, if any, from block to block in the analysis.

Results

Socio-economic and demographic characteristics

Respondents’ ages ranged from 25 to 65 years, with nearly three-quarters (72.1%) aged 25–44 years. One-half (50.6%) lived in the mountains, just over one-third (36.6%) on the delta, and 12.8% on the coast. Most mothers (N=119, 69.2%) had children with mental disabilities, one-quarter (N=44, 25.6%) with physical disabilities, and a few (N=9, 5.2%) with multiple disabilities. More than 90% of respondents were living with spouses. The mothers reported poor educational attainment: only 7% of them had completed high school or higher. Most (N=90, 52.3%) had completed lower secondary school, and one-quarter (N=42, 24.4%) primary school, while a sizeable minority (N=28, 16.3%) were illiterate. Most of the respondents were famers (N=153, 89%). A few (N=6, 3.5%) were government employees or self-employed (N=5, 2.9%), with the remainder (N=8, 4.7%) being home-makers. Nearly one-half of respondents were both primary income earners (who make money for family living) and caregivers for other family members, including their children with disabilities. Most of the mothers owned their own home which was in good condition (brick wall, tiled roof) and had at least one of the seven essential household appliances (N=141, 82% and N=158, 91.9%, respectively). However, nearly one-half (N=83, 48.3%) of respondents reported having a government poverty benefit card, much more than the average of 16.1% for the rural population in 2008 (40), and nearly three-quarters reported that they had experienced financial problems during the past 12 months.

Mental health

Distress

Scores for distress in the present sample (Table 1) ranged from 14–47 (M=25.07, Md=25, SD=6.02) with very high morbidity rates: only 3.5% of respondents scored <15 (indicating little or no distress), while 81.4% scored 16–30 (moderate distress), and 15.1% scored >30 (severe distress).


Table 1.  Mean scores and standard deviations for frequency and perceptions of community participation, social cohesion, and general psychological distress
          95% CI
Items Min Max Mean SD Lower Upper
Community participation (frequency)
 Contact with household members 2.5 7.0 5.94 1.01 5.79 6.1
 Contact with extended family 1.0 7.0 3.93 0.98 3.93 3.78
 Contact with friends 1.0 7.0 3.54 1.05 3.38 3.67
 Contact with neighbour 1.0 6.0 3.38 1.15 3.21 3.55
 Social contact with workmates 1.0 7.0 3.41 1.32 3.21 3.61
 Adult learning 1.0 5.0 1.60 1.06 1.44 1.76
 Religious observance 1.0 6.5 2.50 1.54 2.27 2.73
 Organised community activities 1.0 7.0 2.98 1.39 2.77 3.19
 Volunteer sector activity 1.0 7.0 1.61 1.23 1.42 1.8
 Giving money to charity 1.0 7.0 3.81 1.38 3.60 4.02
 Active interest in current affairs 1.0 7.0 2.60 1.57 2.37 2.83
 Expressing opinions publicly 1.0 6.0 1.89 1.26 1.7 2.08
 Community activism 1.0 7.0 2.38 1.76 2.12 2.65
 Political protest 1.0 7.0 1.43 1.14 1.26 1.6
 Breadth of participation 1.57 5.47 2.96 .67 2.86 3.06

           
Community participation indices (perceptions)
 Too little time for participation 1.0 6.43 4.88 .92 4.74 5.02
 Enjoy participation 1.0 7.0 4.84 1.1 4.68 5.01

           
Social cohesion
 Social trust, total score 1.0 6.43 4.75 .676 4.65 4.85
 Negotiate honestly 1.0 7.0 5.08 1.38 4.87 5.29
 Don't take advantage 2.83 6.42 5.10 1.03 4.95 5.23
 Keep commitments 1.0 7.0 4.45 1.12 4.28 4.62
 World Values Survey trust 1.50 7.0 4.49 1.17 4.31 4.66
 World Values Survey reciprocity 2.0 6.75 4.92 1.14 4.75 5.09
 Optimism 1.5 7.0 4.93 .84 4.81 5.06
 Sense of belonging 2.0 10.0 6.42 1.57 6.18 6.66
 Tangible support 2.0 10.0 6.81 1.92 6.52 7.10
 General psychological distress 14 47 25.07 6.02 24.16 25.98

There are no Vietnamese data with which to compare these findings, but Australian norms are 68%, 29%, and 3% (Kessler et al., 2002). Younger mothers reported the greatest distress (Fig. 1).

Fig 1

Fig. 1.   Proportions of mothers reporting no/little, moderate, and severe distress by age group.

Social capital

Community participation

Most mothers participated infrequently in community activities (Table 1). The most common forms of participation were contact with household members, with extended family, with friends, with workmates, and giving money for charity; they undertook less community activism, adult learning, volunteer activities, expressing opinion, and political protest. Most respondents reported that they spent too little time in community participation but that they enjoyed these activities when they could. In a multiple linear regression analysis, 9 of the 14 types of participation were independently associated with distress, 3 of them with less distress (in the order: neighbours, adult learning, and friends) and 6 with greater distress (activism, extended family, immediate household, political protest, workmates, and religious observance). To assess any impact of breadth of participation, two indices were constructed by summing the items endorsed. One measured breadth of participation across types of activities that were associated with less distress and one with more.

Personal social cohesion

Mothers reported moderate to high levels of all components of personal social cohesion (trust, reciprocity, optimism, sense of belonging, and tangible support). However, the components of social capital were only modestly associated with mothers’ distress (Table 2). Frequency of participation, trust, and sense of belonging were significantly negatively related to mothers’ distress such that greater participation, trust, and sense of belonging were associated with less distress. The components of social capital were positively correlated, as expected, such that those scoring higher on one element were likely also to score higher on another.


Table 2.  Pearson Product Moment correlations among social capital variables and mothers’ general psychological distress (K10 scores)
  2 3 4 5 6 7 8 9 10 11 12 13 Distress
1. Breadth (less distress) 0.62*** −0.07 0.03 −0.04 −0.12 −0.05 0.08 0.09 0.08 −0.36*** 0.00 −0.08 −.14
2. Breadth (more distress)   −0.06 −0.07 −0.03 −0.21** 0.06 0.09 −0.10 −0.01 −0.29*** 0.05 −0.13 0.21**
3. Too little participation     0.41*** 0.32*** 0.40*** −0.01 0.21** 0.22*** 0.42*** 0.06 −0.18* 0.00 0.04
4. Enjoy participation       0.10 0.22** −0.12 0.09 0.10 0.32*** −0.07 −0.16* 0.08 −0.17*
5. Social trust (full scale)         0.65*** 0.47*** 0.76*** 0.17* 0.29*** −0.02 −0.11 0.18* 0.02
6. Negotiate honestly           −0.13 0.35*** 0.26*** 0.38*** 0.11 −0.12 0.11 −0.02
7. Do not take advantage             0.04 −0.06 −0.06 −0.03 −0.12 0.20** 0.00
8. Keep commitments               0.12 0.22** −0.12 0.01 0.04 0.06
9. WVS trust                 0.21** 0.08 −0.17* 0.15 −0.24**
10. WVS reciprocity                   −0.06 −0.15 0.19* −0.02
11. Sense of belonging                     −0.01 −0.06 0.00
12. Tangible support                       −0.34*** 0.07
13. Optimism                       −0.21**
*p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001.

Hierarchical multiple linear regression analyses were used to assess the relative contributions of the different measures in explaining the variance of mothers’ psychological distress (Table 3). In step 1, ‘living in a mountainous area’ (less distress) and having the role of combined carer and income earner (more distress), each made significant independent contributions to explaining variance in mothers’ distress. Children's type of disability made no significant contribution to the model in step 2. Extroversion (specifically, being ‘reserved’) made a small contribution to explaining mothers’ (lower) distress when added in step 3, and also resulted in the role of ‘combined carer and income earner’ becoming non-significant. This suggests that being reserved accounted for the relationship between this role and being more distress.


Table 3.  Multiple hierarchical linear regression analysis showing blocks of predictors of mothers’ general psychological distress (K10 scores)
  B SE B Beta R2 change Adj. R2
1. Socio-demographic characteristics          
  Mountainous area −2.63 0.92 −0.22** 0.08*** 0.08***
  (Combined earner and carer)## 1.92 0.92 0.16*    
2. Child's disability       ns ns
3. Extroversion          
  Mountainous area −3.18 0.87 −0.27*** 0.04** 0.09***
  Being ‘reserved’ −0.69 0.27 −0.19**    
4. Frequency of participation          
  Mountainous area −3.72 0.83 −0.31*** 0.21*** 0.27***
  Being ‘reserved’ −0.60 0.26 −0.16*    
  Immediate household 1.72 0.88 0.14*    
  Extended family 1.06 0.51 0.17*    
  Friends −1.37 0.50 −0.24**    
  Neighbours −1.28 0.41 −0.24**    
  Workmates 1.00 0.34 0.22**    
  Adult learning −2.77 1.08 −0.20**    
  Activism 3.16 0.91 0.26**    
  Political protest 2.68 1.20 0.18*    
5. Breadth of participation       ns ns
6. Perceptions of participation       ns ns
7. Social cohesion          
  Mountainous area −3.45 0.83 −0.29*** 0.02* 0.29***
  Being ‘reserved’ −0.59 0.26 −0.16*    
  Immediate household 1.80 0.87 0.15*    
  Extended family 1.07 0.50 0.18*    
  Friends −1.34 0.49 −0.23**    
  Neighbours −1.14 0.41 −0.22**    
  Workmates 0.85 0.35 0.19*    
  Adult learning −2.72 1.06 −0.20**    
  Activism 3.03 0.90 0.25***    
  Political protest 2.49 1.19 0.17*    
  World Values Survey trust −1.78 0.83 −0.15*    
## Note: Predictor shows mediation effect later in the model (displayed in brackets).
*p-value<.05, **p-value<.01, ***p-value<.001; ‘ns’ means not significant.

In step 4, frequency of participation variables were entered with 8 of the 14 types of participation making a significant and, together, large (21%) independent contribution to explaining variance in mothers’ distress. Three of them contributed to less distress (in the order, friends, neighbours, and adult learning) and five to greater distress (activism, workmates, political protest, extended family, and immediate household). Breadth of participation in step 5 and perceptions about participation in step 6 did not contribute explained variance to the model. In the final step, one component of social cohesion (believing people could ‘be trusted’) made a further, small contribution to explaining variance in mothers’ distress, bringing total adjusted variance explained to 29%. In the final model, less distress was predicted by, in the order, living in a mountainous area, seeing friends and neighbours, participating in adult learning, and trusting others. Greater distress was predicted, in the order, by engagement with activism, workmates, extended family, political protest, and immediate household.

Discussion

In Vietnam, as elsewhere, mothers with children with disabilities have very high levels of distress. Younger mothers in this sample reported the greatest distress, consistent with international findings that mental health problems are more prevalent among younger than older people. In his review (41), Bailey noted that many different tools have been used to measure mental health in different studies, complicating the comparison of findings across studies. The present study was the first to translate and use the K10 measure of general psychological distress. Compared with prevalence studies focused on depression, we found relatively high rates of distress among this sample of mothers. This is not surprising, as distress is a more-encompassing concept than depression.

Although the mothers in this study reported high levels of social cohesion, except for one measure of trust, components of cohesion were not linked to their mental health, unlike in developed economies. As in other studies (2, 3, 26, 4143), these mothers had very low levels of community participation. Participation was quite strongly linked to distress noting that, among these mothers, except for three types of activity (neighbours, friends, and adult learning), participation was linked to greater, and not to less, distress. This finding has been observed internationally in developed countries (44), and among Aboriginal Australians (20, 22, 23) who, while living in a developed economy, tend not to share in its benefits. It is possible that, being part of a severely marginalised group (a circumstance that is much more prevalent among women than among men), participation is more difficult and more costly, in many ways, to these chronically stressed and under-resourced parents (22).

We found other differences between these exceptionally pressured, low-income country mothers compared to parents of children with disabilities in other countries. For example, socio-demographic characteristics and the nature of the child's disability did not help explain variance in mothers’ distress, and neither did most of the components of social cohesion. That said, social capital (and particularly – at the individual level – social cohesion) has been found to be helpful for common mental disorders among young mothers in developing countries, including Vietnam (45). However, the mothers in this earlier study were not parents of children with disabilities. The present findings suggest, perhaps, that their circumstances are so unusual or extreme that a wholly different explanatory framework needs to be developed. For example, stigma and misunderstanding may be significant factors (46), as may the quality of relationships (47). To better understand the circumstances of these mothers, future studies would benefit from the inclusion of a comparison control group of mothers whose children do not have disabilities where the study is conducted in an otherwise identical setting.

The issue of being under-resourced and marginalised is enormously important. For example, it has been reported that families with children who need to be assisted by technology receive much less help with childcare (day care, babysitting, help from relatives, and professional nursing) than other families (48), though they need help so much more. Being isolated may also contribute to mothers’ distress: mothers who provide full-time childcare for children with disabilities report greater depression than mothers who do not (49).

There are some substantial weaknesses in the present study, key among them the understandably small sample size (which makes it difficult to be confident in the statistical estimates and impossible to conduct certain statistical tests) and the use of measures that have been translated and deployed for the first time in Vietnam (making it difficult to be sure that the correct concept is being tested, and tested correctly). Nevertheless, mothers could and did complete the survey, and it was meaningful to them. The sample was appropriately selected, with a strong response rate, and represents an important and frequently overlooked group in Vietnamese society. The study provides insights into the complicated relationship that mothers of children with disabilities have with their communities, and the many ways in which engagement can be difficult, possibly even harmful for their mental health.

Conclusion

Mothers of children with disabilities in rural Vietnam are very distressed and social capital, especially the participation component, significantly predicts this distress. Building social capital among these mothers could be helpful for their mental health but, if this strategy is used, it will be essential to ensure that women are involved in types of participation that are associated with better mental health (friends, neighbours, learning) and not encouraged to participate in those that are not. This may be especially important in a context in which financial support is not easily available or, possibly, not the most helpful kind of support (11), wherein social capital may appear a preferable alternative. Opportunities for adult education may be an alternative strategy, as this seems to be a positive experience for these highly disadvantaged mothers.

Acknowledgements

We thank Catholic Relief Services for financial support and coordination of this study. We thank BPH Dang Thi Ha Trang and Tran Thi Thu Thuy, junior staff at the Hanoi School of Public Health, who helped us collect data at field sites and undertook data input. Special thanks go to the Australian Government Endeavour Postdoctoral Research Fellowship Program for awarding a scholarship to Dr Nguyen and thus providing the opportunity to work together and development of her research capacity.

Conflict of interest and funding

The authors declare that they have no competing interests.

References

  1. Ceylan R, Aral N. An examination of the correlation between depression and hopelessness levels in nmothers of disabled children. Soc Behav Pers 2007; 35: 903–8. Publisher Full Text
  2. Altindağ FO, Işcan A, Akcan S, Köksal S, Erçin M, Ege L. Anxiety and depression in mothers of children with cerebral palsy. Turk J Phys Med Rehabil 2007; 53: 22–4.
  3. Bumin G, Günal A, Tükel Þ. Anxiety, depression and quality of life in mothers of disabled children. SDÜ Týp Fak Derg 2008; 15: 6–11.
  4. Emerson E, Llewellyn G. The mental health of Australian mothers and fathers of young children at risk of disability. Aus N Z J Public Health 2008; 32: 53–9. Publisher Full Text
  5. Veisson M. Depression symtoms and emotional states in parents of disabled and non-disabled children. Soc Behav Pers 1999; 27: 87–97. Publisher Full Text
  6. Sharghi A, Karbakhsh M, Nabaei B, Meysamie A, Farrokhi A. Depression in mothers of children with thalassemia or blood malignancies: a study from Iran. Clin Pract Epidemiol Ment Health 2006; 2(27). Available from: http://www.cpementalhealth.com/content/pdf/1745-0179-2-27.pdf
  7. Mugno D, Ruta L, D'Arrigo VG, Mazzone L. Impairment of quality of life in parents of children and adolescents with pervasive developmental disorder. Health Qual Life Outcomes 2007; 5: 22. PubMed Abstract | PubMed Central Full Text | Publisher Full Text
  8. Eriksson U, Hochwalder J, Carlsund A, Sellstrom E. Health outcomes among Swedish children: the role of social capital in the family, school and neighbourhood. Acta Paediatr 2012; 101(5): 513–7. PubMed Abstract | Publisher Full Text
  9. Whitely J. A model of general self-concept for students with learning disabilities: does class placement play a role? Dev Disabil Bull 2008; 36: 106–34.
  10. Foreman PJ, Dempsey IJ. Follow-up study of children attending a special education preschool. J Aus Res Early Child Educ 1998; 2: 15–23.
  11. Beresford BA. Easing the strain: assessing the impact of a family fund grant on mothers caring for a severely disabled child. Child Care Health Dev 1993; 19: 369–78. PubMed Abstract | Publisher Full Text
  12. Raina P, O'Donnell M, Schwellnus H, Rosenbaum P, King G, Brehaut J, et al. Caregiving process and caregiver burden: conceptual models to guide research and practice. BMC Pediatr 2004; 4(1). Available from: http://hinari-gw.who.int/whalecomwww.biomedcentral.com/whalecom0/content/pdf/1471-2431-4-1.pdf. Publisher Full Text
  13. Foley D, Chowdhury J. Poverty, social exclusion and the politics of disability: care as a social good and the expenditure of social capital in Chuadanga, Bangladesh. Soc Policy Adm 2007; 41: 372–85. Publisher Full Text
  14. Miller E, Buys L. Does social capital predict happiness, health, and life satisfaction in an urban Australian community? N Z J Soc Sci Online 2008; 3: 15–20.
  15. Ziersch AM, Baum F, Darmawan IG, Kavanagh AM, Bentley RJ. Social capital and health in rural and urban communities in South Australia. Aust N Z J Public Health 2009; 33: 7–16. PubMed Abstract | Publisher Full Text
  16. Almedom AM. Social capital and mental health: an interdisciplinary review of primary evidence. Soc Sci Med 2005; 61: 943–64. PubMed Abstract | Publisher Full Text
  17. Mitchell AD, Bossert TJ. Measuring dimensions of social capital: evidence from surveys in poor communities in Nicaragua. Soc Sci Med 2007; 64: 50–63. PubMed Abstract | Publisher Full Text
  18. Whitley R, McKenzie K. Social capital and psychiatry: review of the literature. Harv Rev Psychiatry 2005; 13: 71–84. PubMed Abstract | Publisher Full Text
  19. Harpham T, Grant E, Thomas E. Measuring social capital within health surveys: key issues. Health Policy Plan 2002; 17: 106–11. PubMed Abstract | Publisher Full Text
  20. Berry HL, Rodgers B, Dear KB. Preliminary development and validation of an Australian community participation questionnaire: types of participation and associations with distress in a coastal community. Soc Sci Med 2007; 64: 1719–37. PubMed Abstract | Publisher Full Text
  21. Berry HL, Rickwood D. Measuring social capital at the individual level: a values-based model. Int J Ment Health Promot 2000; 2: 35–44.
  22. Berry HL. Social capital elite, excluded participators, busy working parents and aging, participating less: types of community participators and their mental health. Soc Psychiatry Psychiatr Epidemiol 2008; 43: 527–37. PubMed Abstract | Publisher Full Text
  23. Berry HL, Welsh JA. Social capital and health in Australia: an overview from the household, income and labour dynamics in Australia survey. Soc Sci Med 2010; 70: 588–96. PubMed Abstract | Publisher Full Text
  24. Horton R. Launching a new movement for mental health. Lancet 2007; 370: 806. PubMed Abstract | Publisher Full Text
  25. Shinn M, Gottlieb J, Wett JL, Bahl A, Cohen A, Ellis DB. Predictors of homelessness among older adults in New York city: disability, economic human and social capital and stressful events. J Health Psychol 2007; 12: 696–708. PubMed Abstract | Publisher Full Text
  26. Zebrack B, Lim JW. Caring for family members with chronic physical illness: a critical review of caregiver literature. Health Qual Life Outcomes 2004; 2(50). Available from: http://www.hqlo.com/content/pdf/1477-7525-2-50.pdf
  27. Shapiro J. Stress, depression, and support group participation in mothers of developmentally delayed children. Fam Relat 1989; 38: 169–73. Publisher Full Text
  28. Kane T. Disability in Vietnam in 1999: A meta-analysis of the data, in USAID report 1999. Available from: http://pdf.usaid.gov/pdf_docs/PNACG781.pdf
  29. UNICEF, MOLISA. An analysis of situation of children with disabilities in Viet Nam, 2004. Hanoi: Ministry of Labor, Invalids and Social Affairs. 2006.
  30. The Centre for International Economics. Vietnam poverty analysis. Canberra, Sydney. 2002. Available from: http://www.ausaid.gov.au/publications/pdf/vietnam_poverty_analysis.pdf
  31. Kessler RC, Andrews G, Colpe LJ, Hiripi E, Mroczek DK, Normand SL, et al. Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol Med 2002; 32: 959–76. PubMed Abstract | Publisher Full Text
  32. Australian Bureau of Statistics. Cat. No. 2068.0 - 2006 Census Tables. Canberra: ABS, Government of Australia; 2006.
  33. Berry H, Rodgers B. Trust and distress in three generations of rural Australians. Australas Psychiatr 2003; 11: S131–7. Publisher Full Text
  34. Berry HL, Rickwood DJ. Measuring social capital at the individual level: personal social capital, values and psychological distress. Int J Ment Health Promot 2000; 2: 35–44.
  35. Inglehart RF, Miguel B, Jaime D-M, Loek H, Ruud L, et al. Human Beliefs and Values: a Cross-cultural Sourcebook based on the 19997–2002 Value Surveys. Mexico City: Siglo XXI. 2004.
  36. Scheier MF, Carver CS, Bridges MW. Distinguishing optimism from neuroticism (and trait anxiety, self-mastery, and self-esteem): a re-evaluation of the life orientation test. J Pers Soc Psychol 1994; 67: 1063–78. PubMed Abstract | Publisher Full Text
  37. Berry HL, Shipley M. Longing to belong: personal social capital and psychological distriess in an Australian coastal region. Social Policy Research Paper 39. Vol. Commonwealth of Australia: Canberra. 2009, Canberra: Department of Families, Housing, Communities and Indigenous Affairs.
  38. Cohen S, Mermelstein R, Kamarck T, Hoberman HM. Interpersonal support evaluation list, in social support: theory, research and application. In: Sarason GS, Sarason BR, eds. Washington, DC: Martinus Nijhoff; 1985.
  39. Gosling SD, Peter RP, Swann WB. A very brief measure of the big-five personality domains. J Res Pers 2003; 37: 504–28. Publisher Full Text
  40. General Statistics Office. Poverty prevalence by urban/rural, region and provinces. Hanoi: Ministry of Labor, Invalids and Social Affairs; 2010.
  41. Bailey DB Jr, Golden RN, Roberts J, Ford A. Maternal depression and developmental disability: research critique. Ment Retard Dev Disabil Res Rev 2007; 13: 321–9. PubMed Abstract | Publisher Full Text
  42. Buys L, Roberto KA, Miller E, Blieszner R. Prevalence and predictors of depressive symptoms among rural older Australians and Americans. Aust J Rural Health 2008; 16: 33–9. PubMed Abstract | Publisher Full Text
  43. Shin J, Nhan NV, Crittenden KS, Hong HT, Flory M, Ladinsky J. Parenting stress of mothers and fathers of young children with cognitive delays in Vietnam. J Intellect Disabil Res 2006; 50: 748–60. PubMed Abstract | Publisher Full Text
  44. Kawachi I, Berkman LF. Social ties and mental health. J Urban Health - B New York Acad Med 2001; 78: 458–67. Publisher Full Text
  45. De Silva MJ, Huttly SR, Harpham T, Kenward MG. Social capital and mental health: a comparative analysis of four low income countries. Soc Sci Med 2007; 64: 5–20. PubMed Abstract | Publisher Full Text
  46. Tufan I. Prejudices against, and social responsibilities towards, the disabled. Soc Behav Pers 2008; 36: 67–76. Publisher Full Text
  47. Kazak AE. Families with disabled children: stress and social networks in three samples. J Abnorm Child Psychol 1987; 15: 137–46. PubMed Abstract | Publisher Full Text
  48. Thyen U, Kuhlthau K, Perrin JM. Employment, child care, and mental health of mothers caring for children assisted by technology. Pediatrics 1999; 103: 1235–42. PubMed Abstract | Publisher Full Text
  49. Trute B. Gender differences in the psychological adjustment of parents of young, developmentally disabled children. J Child Psychol Psychiatry 1995; 36: 1225–42. PubMed Abstract | Publisher Full Text

*Nguyen Thi Minh Thuy
Head of Rehabilitation Department
Hanoi School of Public Health
138 Giang Vo, Hanoi, Vietnam
Email: ntmt@hsph.edu.vn

 

PUBLIC HEALTH IN VIETNAM: HERE'S THE DATA, WHERE'S THE ACTION?

The association and a potential pathway between gender-based violence and induced abortion in Thai Nguyen province, Vietnam

Phuong Hong Nguyen1,2*, Son Van Nguyen1, Manh Quang Nguyen1, Nam Truong Nguyen3, Sarah Colleen Keithly3, Lan Tran Mai2, Loan Thi Thu Luong1 and Hoa Quynh Pham1

1Thai Nguyen University of Medicine and Pharmacy, Thai Nguyen, Vietnam; 2International Food Policy Research Institute, Hanoi, Vietnam; 3Institute of Social and Medical Studies, Hanoi, Vietnam

Abstract

Background: Gender-based violence (GBV) has profound adverse consequences on women's physical, mental, and reproductive health. Although Vietnam has high rates of induced abortion and GBV, literature examining this relationship is lacking.

Objective: This study examines the association of GBV with induced abortion among married or partnered women of reproductive age in Thai Nguyen province, Vietnam. In addition, we explore contraceptive use and unintended pregnancy as mediators in the pathway between GBV and induced abortion.

Design and methods: Data were drawn from a cross-sectional survey of 1,281 women aged 18–49 years in four districts of Thai Nguyen province. Bivariate and multivariate logistic regression analyses were applied to examine the associations between lifetime history of GBV, contraceptive use, unintended pregnancy, induced abortion, and repeat abortion, controlling for other covariates.

Results: One-third of respondents had undergone induced abortion in their lifetime (33.4%), and 11.5% reported having repeat abortions. The prevalence of any type of GBV was 29.1% (17.0% physical violence, 10.4% sexual violence, and 20.1% emotional violence). History of GBV was associated with induced abortion (OR=1.61, 95% CI: 1.20–2.16) and repeat abortion (OR=2.22, 95% CI: 1.48–3.32). Physical violence was significantly associated with induced abortion, and all three types of violence were associated with repeat abortion. Abused women were more likely than non-abused women to report using contraceptives and having an unintended pregnancy, and these factors were in turn associated with increased risk of induced abortion.

Conclusions: GBV is pervasive in Thai Nguyen province and is linked to increased risks of induced abortion and repeat abortion. The findings suggest that a pathway underlying this relationship is increased risk of unintended pregnancy due in part to ineffective use of contraceptives. These findings emphasize the importance of screening and identification of GBV and incorporating women's empowerment in reproductive health and family planning programs.

Keywords: Gender-based violence; abortion; contraceptive use; unintended pregnancy; Thai Nguyen; Vietnam

Received: 19 June 2012; Revised: 26 September 2012; Accepted: 21 October 2012; Published: 29 November 2012

Glob Health Action 2012. © 2012 Phuong Hong Nguyen et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Glob Health Action 2012, 5: 19006 - http://dx.doi.org/10.3402/gha.v5i0.19006

 

Violence against women, also known as gender-based violence (GBV), is “now widely recognized as a serious human rights abuse” as well as “an important public health problem that concerns all sectors of society” (1). Internationally, attempts have been made to redress this issue, with the United Nations Declaration of Elimination of Violence against Women in 1993. In 2005, the World Health Organization (WHO) conducted a multi-country study on violence against women, which highlighted that GBV prevalence is far greater than that indicated by research prior to 1999 (1). Factors contributing to the hidden nature of GBV include victims’ fear of retaliation or other consequences, gender norms that reinforce women's subordination to their male counterparts, and social norms that justify GBV as ‘normal’ (2). The adverse health consequences of GBV are not limited to the physical and mental wellbeing of a woman but also encompass her reproductive health (1, 3). Domestic violence against women in their childbearing years can lead to a range of health problems for both mothers and their babies, including late prenatal care during pregnancy, adverse pregnancy outcomes (4), somatic disorders, neonatal problems, infant mortality (5), and even maternal death (6). Specifically, literature has shown the negative effects of domestic violence against women on sexual autonomy, unintended pregnancy (3, 7), spontaneous abortion, and induced abortion (8, 9).

Implications for policy and practice

  • Assessment and treatment of GBV should be integrated into standard reproductive health and family planning services in Vietnam. To do so successfully, health providers need guidelines and training in screening, documenting, and treating GBV.
  • Investment is needed to establish shelters and support centers for victims of GBV that provide both protection and counseling services. Linkages should be developed between these centers and reproductive health and family planning services.
  • Programs are needed to address gender inequality by advancing women's empowerment, socioeconomic status, and education as a means for curbing GBV.
  • It is necessary to increase public awareness of the physical, reproductive, mental, and societal consequences of GBV and to shift norms of acceptance of GBV.
  • The capacity of the police and local authorities to uphold GBV policies and legislation should be strengthened.

Previous studies show that GBV remains a widespread problem in Vietnam with indications that it may be on the increase (10). According to a 2010 nationally representative study on domestic violence, 58% of Vietnamese women have experienced some form of violence in their lifetime, with prevalence rates of physical, sexual, and emotional violence at 32, 10, and 54%, respectively (11).

In addition to GBV, maternal health in Vietnam suffers from conspicuously high abortion rates (an average of 2.5 induced abortions in a woman's reproductive life), which persist despite the country's high contraceptive prevalence rate (78% among married couples) (12, 13). While no official data on unsafe abortion are available for Vietnam, a 2000–2001 study found that unsafe abortion was the direct cause of 11.5% of maternal deaths in seven Vietnamese provinces (14).

Due to the scope of violence against women and its importance for reproductive health, research has increasingly focused on examining the association between GBV and abortion. Studies in developing countries, such as India (15), Tanzania (16), Bangladesh (9), China (17), and Uganda (7), have identified GBV as a significant risk factor for induced abortion. However, the literature examining the determinants of induced abortion in Vietnam is yet to explore GBV as a possible predictor (18, 19). In light of this gap, this study aims to document the prevalence of GBV among married women of reproductive age in Thai Nguyen province, Vietnam, and to examine the association of GBV with induced abortion and repeat induced abortion.

The study also aims to explore an underlying pathway linking GBV and induced abortion in which contraceptive use and unintended pregnancy act as mediating factors. Research suggests that by creating an environment of fear and intimidation, increasing psychological stress, and reducing access to services and support, GBV may limit women's control over their fertility, thereby reducing contraceptive use or interrupting the effective use of contraceptives (2022). It has also been posited that victims of GBV may feel unprepared or have reduced desire to raise a child in an abusive environment, which could increase the likelihood that pregnancies are unintended and terminated (22, 23). Further, women may be forced into having an abortion by their abusive partners (16, 22, 24). We hypothesized that GBV reduces women's contraceptive use, thereby increasing their risk of unintended pregnancy and, consequently, induced abortion (Fig. 1).

Fig 1
Fig. 1.  Conceptual model of a potential pathway between gender-based violence (GBV) and induced abortion.

Methods

Design and settings

Data were drawn from a cross-sectional household survey conducted as part of a project entitled ‘Primary Health Improvement Initiative in Thai Nguyen Province, Vietnam.’ By providing evidence for health policy planning, the project aimed to advance the national effort to attain universal access to family planning and safe abortion services in Vietnam. The survey was conducted between February and May 2011 in four of the nine districts of Thai Nguyen, a province in northern Vietnam with high proportions of indigent and ethnic minority women. These districts account for 52% of the province's total population and are representative of both urban (Thai Nguyen city) and rural areas (Dai Tu, Dinh Hoa, and Vo Nhai districts).

Sampling and data collection

A representative sample of married or partnered women aged 18–49 years was selected from the four districts by a two-stage cluster sampling technique. The first stage involved selection of 40 communes as primary sampling units, with probability of inclusion proportional to the size of the population in each district. In the second stage, 40 individual respondents were selected in each commune via systematic random sampling.

Data were collected using face-to-face interviews using a structured questionnaire. The survey collected information from respondents on sociodemographic background, sexuality and obstetric history, knowledge, attitudes and practices related to contraception, fertility preferences, history of GBV, attitudes towards and access to family planning and reproductive health services, history of abortion, and experience with abortion services. Interviewers were trained medical doctors with sufficient technical ability to distinguish among induced abortions, stillbirths, and miscarriages. A total of 1,540 women completed the survey. Among those, 1,281 women were married or living with partner and were included in this analysis.

Variables of interest

The outcomes were self-reported lifetime history of induced abortion and repeat (two or more) induced abortion. The main independent variables were self-reported lifetime experience of any GBV as well as physical violence, sexual violence, and emotional violence, based on the WHO definition (1) and adapted for the Vietnam context (11). Physical violence was assessed in terms of 11 items: slapping, throwing things, pushing or shoving, hair pulling, hitting, kicking, dragging, beating, choking, burning, and threatening with or using a weapon such as a knife or scissors. Sexual violence was assessed in terms of three items: having sexual intercourse against the respondent's will, partner using excessive physical force during sexual intercourse, and being forced to engage in sexually degrading acts. Emotional violence was assessed in terms of three items: insults or degrading activities, belittlement or humiliation, and scaring the respondent (including threats of violence). Lifetime occurrence of any kind of violence was defined as experience of any act of violence up to the date of the interview, perpetrated by a current or former husband/partner. In addition, because women reporting a particular type of violence (e.g. physical) could also be exposed to another form of violence (e.g. sexual), lifetime history of each type of violence was assessed independently.

Potential mediating explanatory factors included contraceptive use and unintended pregnancy. To measure contraceptive use, respondents were asked whether they had ever used any contraceptive method in their lifetime, and, if so, the type(s) of method used. Responses were categorized into three groups: use of any contraceptives (yes/no), use of female-only methods (yes/no), and use of couple methods (yes/no). Female-only methods included oral pills, intrauterine devices (IUDs), injectables, implants, and female sterilization. Couple methods included those that require at least the awareness and a certain degree of support and cooperation from husbands, such as male and female condoms, diaphragms, withdrawal, breastfeeding, and periodic abstinence. Unintended pregnancy was assessed by asking respondents if they would describe their current pregnancy or any pregnancy in the past as unintended (yes/no).

Additional explanatory variables included as control variables in the multivariate analyses were women's age at interview (18–24 years/25–34 years/35–49 years); ethnicity (Kinh majority/other ethnic minority groups); highest level of education completed [primary school (grades 1–5)/secondary school (grades 6–9)/high school (grades 10–12)/college or higher]; main occupation (farmers/other occupations); area of residence (rural/urban); household social economic status (SES) (lowest/lower/middle/higher/highest); number of living children (≤1 child/2 children/≥3 children); and age at first intercourse (14–19 years/20–24 years/≥25 years). The household SES index was constructed using principal components analysis using asset data such as house and land ownership, housing quality, access to services (water, electricity, gas, and sanitation services), and household assets (different types of durable goods, productive assets, animals and livestock) (25, 26). Household SES categories were formed by dividing the index into quintiles. Due to its high correlation with occupation, area of residence was not included in the multivariate analyses.

Statistical analyses

Data were analyzed using Stata 11 (27). Descriptive statistics were used to report the study population's background characteristics and to estimate prevalence of GBV and induced abortion. Proportions and chi-squared test were used to examine the binary associations between GBV and contraceptive use, unintended pregnancy, and induced abortion outcomes. Multivariate logistic regression analysis was used to explore the association of GBV variables with lifetime history of induced abortion and repeat induced abortion, controlling for other covariates. In addition, multivariate logistic regression analyses were run to study the relationship between GBV and contraceptive use and unintended pregnancy. Odds ratios (OR) with 95% confidence intervals (95% CI) are presented for the final models.

Ethical considerations

Ethical approval for this study was obtained from the Institutional Review Board of the Institute of Social and Medicine Studies in Vietnam. Written informed consent was obtained from all respondents, and measures were taken to ensure participants’ confidentiality and privacy throughout the study process.

Results

The overall response rate was high at 95%. The internal non-response rates of the study's sensitive questions – including those related to contraceptive use, unintended pregnancy and induced abortion, and exposure to GBV – were very low, ranging from 0.3–2.1%. The limited number of respondents with missing data on these variables was unlikely to have greatly affected the results.

Table 1 provides an overview of the prevalence of and associations between lifetime history of GBV with contraceptive use, unintended pregnancy, induced abortion, and repeat induced abortion. One-third of women reported undergoing at least one induced abortion, and 11.5% reported having two or more induced abortions in their lifetime. The prevalence of GBV was relatively high, with 29.1% of women reporting experiencing any types of violence by their husband/partner, 17.0% reporting physical violence, 10.4% reporting sexual violence, and 20.1% reporting emotional violence. About two-thirds of respondents had used contraception, with more women reporting use of female-only methods (48.8%) than couple methods (19.9%). Close to one-quarter of women (23.6%) reported having an unintended pregnancy currently or in the past.


Table 1.  Prevalence of lifetime history of gender-based violence (GBV) and association between GBV and lifetime contraceptive use, unintended pregnancy, and induced abortion
    Lifetime history of contraceptive use and unintended pregnancy (n=1,260) Lifetime history of induced abortion (n=1,254)
  Prevalence (n=1,281) Contraceptive use (n=865, 68.66%) Use of female contraceptive method (n=615, 48.77%) Use of couple contraceptive method (n=251, 19.89%) Unintended pregnancy (n=297, 23.57%) Abortion (n=421, 33.57%) Repeat abortion (n=144, 11.48%)
Lifetime history of GBV Percent (95% CI) Percent Percent Percent Percent Percent Percent
Any GBV              
  No 70.95 (68.44–73.46) 65.70 44.39 20.98 21.05 30.54 9.07
  Yes 29.05 (26.54–31.56) 77.57*** 50.00*** 17.14 29.56** 40.93*** 17.31***
Physical GBV              
  No 83.02 (80.94–85.09) 67.32 46.24 20.77 21.88 31.41 9.69
  Yes 16.98 (14.91–19.06) 78.02** 62.15*** 15.25+ 31.60** 44.13*** 20.19***
Sexual GBV              
  No 89.62 (87.93–91.30) 68.25 47.89 19.92 22.51 32.73 10.40
  Yes 10.38 (8.69–12.07) 76.07+ 56.41+ 19.66 32.82** 40.46+ 20.61**
Emotional GBV              
  No 79.87 (77.66–82.09) 66.99 45.72 21.02 22.20 32.22 9.56
  Yes 20.13 (17.91–22.34) 77.73** 61.97*** 15.02* 29.08* 38.74* 18.97***
+p<0.10, *p<0.05, **p<0.01, ***p<0.001.

Lifetime history of induced abortion and repeat induced abortion were significantly associated with having experienced GBV (Table 1). Women who had undergone an abortion were significantly more likely to report experiencing any types of violence, physical violence, and emotional violence and were marginally more likely to experience sexual violence. Those reporting repeat abortion had a greater likelihood of experiencing any types of violence as well as all forms of GBV. Table 2 summarizes the demographic characteristics of the study sample as well as the association between these characteristics with induced abortion and repeat abortion. As expected, women who were older and had more children were more likely to have experienced abortion. Women who worked as farmers were less likely to report abortion compared to women having other jobs (30.1% vs. 39.9% for induced abortion and 10.1% vs. 14.1% for repeat abortion). Women who had first intercourse before age 25 were more likely to experience abortion. In addition, women reporting use of contraceptives also had increased likelihood of having experienced abortion. Women with higher SES status were also more likely to report experiencing any abortion and repeat abortion.


Table 2.  Demographic characteristics, contraceptive use, unintended pregnancy, and lifetime history of induced abortion among study participants
    Lifetime history of abortion (n=1,254)
  Prevalence (n=1,281) Abortion (n=421, 33.57%) Repeat abortion (n=144, 11.48%)
Explanatory variables Percent (95% CI) Percent Percent
Age of respondent (years)      
  18–24 9.40 (7.79–11.00) 10.09 0.00
  25–34 34.38 (31.77–36.70) 23.00 5.16
  35–49 56.22 (53.50–58.95) 43.50*** 16.92***
Ethnicity      
  Kinh 67.37 (64.80–69.94) 34.75 13.00
  Other (ethnic minority) 32.63 (30.06–35.20) 31.13 8.33*
Education      
  Primary school 10.62 (8.93–12.31) 28.68 8.82
  Secondary school 50.27 (47.53–53.01) 34.34 11.39
  High school 23.50 (21.17–25.82) 30.17 11.53
  College or higher 15.61 (13.62–17.60) 39.79+ 13.61
Occupation      
  Farmer 64.77 (62.15–67.37) 30.14*** 10.09*
  Other 35.23 (32.61–37.85) 39.91 14.06
Area of residence      
  Rural 64.64 (62.02–67.26) 29.88*** 8.52***
  Urban 35.36 (32.74–37. 99) 40.32 16.89
Household SES      
  Lowest 20.02 (17.83–22.23) 25.70 7.23
  Low 20.03 (17.83–22.23) 28.97 7.94
  Middle 19.95 (17.76–22.15) 33.47 10.76
  Higher 20.03 (17.83–22.23) 37.55 13.83
  Highest 19.95 (17.76–22.15) 42.28*** 17.89***
Number of living children (range: 0–5)      
  ≤1 34.35 (31.74–36.95) 20.58 4.60
  2 52.30 (49.56–55.04) 38.36 14.48
  ≥3 13.35 (11.48–15.21) 46.20*** 16.37***
Age at first intercourse (years)      
  14–19 27.97 (25.50–30.43) 34.76 11.11
  20–24 52.63 (49.89–55.38) 35.26 12.61
  25–45 19.40 (17.23–21.58) 26.36* 8.79
Ever used contraception      
  No 31.34 (28.64–34.04) 8.20 1.64
  Yes 68.66 (65.96–71.36) 35.49*** 12.08*
Ever used female contraceptive methods      
  No 51.23 (48.32–54.14) 23.17 5.88
  Yes 48.77 (45.89–51.68) 44.85*** 17.18***
Ever used couple contraceptive methods      
  No 80.11 (77.78–82.43) 34.64 12.78
  Yes 19.89 (17.57–22.22) 31.08 6.31**
Ever had unintended pregnancy      
  No 76.43 (74.07–78.80) 32.69 11.32
  Yes 23.57 (21.20–25.93) 50.82** 14.75
+p<0.10, *p<0.05, **p<0.01, ***p<0.001.

The multivariate logistic regression models of the relationship between GBV and abortion outcomes are presented in Tables 3 and 4. After adjusting for potential confounders, women who experienced any form of GBV were 1.61 times (95% CI: 1.20–2.16) more likely to report having an induced abortion and 2.22 times (95% CI: 1.48–3.32) more likely to report having repeat induced abortion. When considering the odds of induced abortion by the three subtypes of GBV, the adjusted OR for physical violence was 1.91 (95% CI: 1.34–2.72), while the associations with the two other subtypes (sexual and emotional violence) were not statistically significant. All three subtypes of GBV were positively associated with repeat abortion, with adjusted OR ranging from 2.28 to 2.50 (Table 4).


Table 3.  Multivariate logistic regression models for the relationship between any kind of gender-based violence (GBV) and induced abortion
  Ever had abortion (n=1,118) Ever had repeat abortion (n=1,118)
Explanatory variables OR (95% CI) OR (95% CI)
Ever experienced any GBV    
  Nor 1.00 1.00
  Yes 1.61** (1.20–2.16) 2.22*** (1.48–3.32)
Ever used contraception    
  Nor 1.00 1.00
  Yes 4.10** (1.56–10.77) 4.81 (0.64–36.22)
Ever had unintended pregnancy    
  Nor 1.00 1.00
  Yes 2.69** (1.44–5.01) 1.82 (0.74–4.43)
Age of respondent (years)    
  18–24r 1.00 1.00
  25–34 3.07* (1.31–7.20)  
  35–49 7.49*** (3.11–18.06) 3.39*** (1.91–6.02)
Ethnicity    
  Kinh 0.76+ (0.56–1.04) 1.01 (0.62–1.64)
  Otherr 1.00 1.00
Education    
  Primary schoolr 1.00 1.00
  Secondary school 1.17 (0.72–1.89) 1.16 (0.53–2.53)
  High school 0.83 (0.46–1.50) 0.95 (0.38–2.38)
  College or higher 1.19 (0.60–2.34) 1.12 (0.40–3.11)
Occupation    
  Farmer 0.61* (0.39–0.93) 0.95 (0.52–1.74)
  Otherr 1.00 1.00
Household SES    
  Lowestr 1.00 1.00
  Lower 1.01 (0.64–1.61) 0.87 (0.40–1.87)
  Middle 1.42 (0.88–2.28) 1.45 (0.69–3.04)
  Higher 1.46 (0.87–2.45) 1.94+ (0.88–4.23)
  Highest 1.55 (0.86–2.81) 2.77* (1.17–6.58)
Number of living children    
  ≤1r 1.00 1.00
  2 1.39 (0.94–2.05) 1.83+ (0.95–3.53)
  ≥3 1.90* (1.11–3.25) 2.19+ (0.96–5.00)
Age at first sexual intercourse (years)    
  14–19 1.63* (1.03–2.58) 1.27 (0.65–2.50)
  20–24 1.61* (1.10–2.36) 1.37 (0.78–2.43)
  25–45r 1.00 1.00
+p<0.10, *p<0.05, **p<0.01, ***p<0.001.
rReference group.


Table 4.  Multivariate logistic regression models for the relationship between physical, sexual, and emotional gender-based violence (GBV) and induced abortion
  Model 11,2 Model 2 Model 3
  Ever had abortion (n=1,118) Ever had repeat abortion (n=1,118) Ever had abortion (n=1,118) Ever had repeat abortion (n=1,118) Ever had abortion (n=1,118) Ever had repeat abortion (n=1,118)
Explanatory variables OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Ever experienced GBV            
  Physical GBV            
  Nor 1.00 1.00        
  Yes 1.91*** (1.34–2.72) 2.50*** (1.58–3.96)        
Sexual GBV            
  Nor     1.00 1.00    
  Yes     1.35 (0.88–2.06) 2.28** (1.34–3.88)    
Emotional GBV            
  Nor         1.00 1.00
  Yes         1.31 (0.94–1.83) 2.36*** (1.53–3.63)
Ever used contraception            
  Nor 1.00 1.00 1.00 1.00 1.00 1.00
  Yes 4.31** (1.64–11.3) 5.38 (0.71–40.6) 4.28** (1.63–11.2) 5.12 (0.68–38.5) 4.28** (1.63–11.2) 5.17 (0.69–39.0)
Ever had unintended pregnancy            
  Nor 1.00 1.00 1.00 1.00 1.00 1.00
  Yes 2.80** (1.50–5.24) 1.87 (0.76–4.61) 2.72** (1.46–5.07) 1.84 (0.76–4.47) 2.74** (1.47–5.12) 1.78 (0.72–4.37)
+p<0.10, * p<0.05, ** p<0.01, *** p<0.001.
1The main independent variables for models 1, 2, and 3 were physical GBV, sexual GBV, and emotional GBV, respectively. These independent variables were highly correlated and were therefore examined in separate models.
2Models control for the following (not shown): age, ethnicity, education, occupation, household SES, number of living children, and age at first intercourse.
rReference group.

To assess contraceptive use and unintended pregnancy as potential mediators on the pathway between GBV and induced abortion, we examined the relationship between these variables and the independent variables of interest (GBV variables) and the dependent variables of interest (induced abortion outcomes). Findings show that both contraceptive use and unintended pregnancy were associated with GBV in both bivariate analyses (Table 1) and multivariate analyses (Table 5). Controlling for potential confounders, women who were exposed to any form of violence were 1.82 times (95% CI: 1.30–2.47) more likely to report contraceptive use and 1.66 times (95% CI: 1.11–2.11) more likely to report unintended pregnancy than unexposed women. Abused women were 1.76 times (95% CI: 1.33–2.33) more likely to report using female-only contraceptive methods than those unexposed to violence. In contrast, the association between GBV and use of couple contraceptive methods was statistically insignificant. In addition to GBV, the reproductive health mediators were also associated with induced abortion outcomes. Use of any form of contraception and use of female-only methods were associated with increased likelihood of having an induced abortion and repeat abortion (Tables 2 and 3). In addition, abortion-seekers were more likely to report having an unintended pregnancy in their lifetime. Multivariate logistic regression analyses show that these relationships remained even after adjusting for potential confounders (Table 3).


Table 5.  Multivariate logistic regression models for the relationship between any kind of gender-based violence (GBV) and contraceptive use and unintended pregnancy
  Contraceptive use  
  Ever used contraception (n=1,138) Ever used female only contraceptive method (n=1,124) Ever used couple contraceptive method (n=1,124) Ever had unintended pregnancy (n=1,211)
Explanatory variables OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Ever experienced any GBV        
  Nor 1.00 1.00 1.00 1.00
  Yes 1.82*** (1.33–2.49) 1.76*** (1.33–2.33) 0.89 (0.62–1.27) 1.66** (1.22–2.26)
Age of respondent (years)        
  18–24r 1.00 1.00 1.00 1.00
  25–34 1.70* (1.05–2.76) 2.35** (1.39–3.96) 0.65 (0.37–1.15) 0.71 (0.36–1.43)
  35–49 1.44 (0.84–2.48) 2.66*** (1.51–4.68) 0.43** (0.22–0.81) 0.41* (0.20–0.87)
Ethnicity        
  Kinh 1.09 (0.82–1.46) 0.92 (0.69–1.22) 1.36+ (0.95–1.96) 1.01 (0.72–1.41)
  Otherr 1.00 1.00 1.00 1.00
Education        
  Primary schoolr 1.00 1.00 1.00 1.00
  Secondary school 0.97 (0.62–1.54) 0.89 (0.58–1.37) 1.13 (0.61–2.09) 1.16 (0.73–1.86)
  High school 0.78 (0.45–1.33) 0.85 (0.51–1.43) 0.90 (0.44–1.81) 0.76 (0.42–1.38)
  College or higher 1.00 (0.52–1.93) 0.60 (0.36–1.12) 1.76 (0.81–3.84) 0.82 (0.40–1.68)
Occupation        
  Farmer 0.98 (0.64–1.50) 0.93 (0.62–1.38) 1.07 (0.66–1.73) 1.00 (0.62–1.60)
  Otherr 1.00 1.00 1.00 1.00
Household SES        
  Lowestr 1.00 1.00 1.00 1.00
  Lower 1.30 (0.86–1.96) 1.14 (0.76–1.70) 1.27 (0.72–2.23) 0.88 (0.56–1.39)
  Middle 1.80** (1.16–2.78) 1.47+ (0.97–2.24) 1.34 (0.76–2.37) 0.81 (0.50–1.32)
  Higher 1.53+ (0.94–2.48) 0.91 (0.57–1.46) 2.21** (1.22–4.02) 1.57+ (0.93–2.64)
  Highest 2.15** (1.21–3.81) 1.36 (0.79–2.34) 1.90+ (0.96–3.74) 1.33 (0.70–2.49)
Number of living children        
  ≤1r 1.00 1.00 1.00 1.00
  2 1.68** (1.18–2.39) 1.61** (1.15–2.25) 1.00 (0.67–1.51) 4.07*** (2.53–6.55)
  ≥3 1.80* (1.06–3.07) 1.82* (1.10–2.99) 0.83 (0.42–1.66) 18.05*** (9.84–33.12)
Age at first sexual intercourse (years)        
  14–19 1.37 (0.89–2.10) 2.06*** (1.36–3.11) 0.53* (0.32–0.88) 1.26 (0.76–2.09)
  20–24 1.37+ (0.96–1.96) 1.85*** (1.31–2.61) 0.63* (0.43–0.94) 1.13 (0.73–1.78)
  25–45r 1.00 1.00 1.00 1.00
+p<0.10, *p<0.05, **p<0.01, ***p<0.001.
rReference group.

Discussion

In parallel with socioeconomic development, Vietnam has seen significant gains in gender equality and women's advancement in recent years. However, Vietnamese culture remains deeply rooted in traditional Confucian principles that serve to lower women's status in relation to men, such as the declaration that wives should be subordinate to their husbands (28). Such inequities between men and women are a root cause of GBV, which is known to be perpetuated by a culture of silence in Vietnam (11). Contributing to this silence are gender norms that place women in the position of maintaining family harmony and call for women to protect the reputation of their husbands and families. GBV is considered a family and private issue, not to be discussed openly. This silence reduces public awareness of the problem and prevents abused women from getting support and protection they need (11).

This study provides evidence that GBV remains a formidable problem in Thai Nguyen province with serious reproductive health consequences for the women who live there. The prevalence rates of GBV in the present study were lower than the results from the 2010 national study on domestic violence (11) (e.g. 31% vs. 58% for any form of GBV), most likely due to variations between the two study populations. Compared to the national study, our sample had larger proportions of rural residents and ethnic minorities, and these populations may be more likely to underreport GBV.

Our findings demonstrate that women who experience GBV are at an increased risk for induced abortion and repeat induced abortion. After adjusting for potential confounders, the risk for repeat abortion remained for all three subtypes for GBV, while the risk of having an induced abortion remained significant for physical violence but not for sexual violence and emotional violence. These findings corroborate some of the findings from literature studying GBV and abortion. Findings from a study in Kenya (8) show that induced abortion was associated with all three subtypes of GBV. However, in Cameroon (29), the risk for induced abortion increased with physical and sexual violence but was not associated with emotional violence.

Evidence suggests that contraceptive use and unintended pregnancy are factors mediating the relationship between GBV and abortion. We found that GBV was positively associated with contraceptive use and unintended pregnancy, and, in turn, these factors were positively associated with induced abortion. These findings are consistent with the literature demonstrating that abused women are more likely to use contraception (5, 23, 30) and to have unintended pregnancy (9, 31) than their non-abused counterparts.

This pathway may be explained by reluctance on behalf of the abused women to raise a child in a violent setting, which would lead them to take matters into their own hands using contraception to avoid pregnancy/childbirth. This theory helps explain why use of female-only contraceptive methods was associated with abortion but not use of couple methods. Couple methods intrinsically require couple communication and cooperation, while female-only methods can be used without the partner's involvement or even knowledge. However, it has also been theorized that this finding may be due to men discovering that their partners are using contraception and then reacting violently, which would also lead to a positive association between contraceptive use and GBV (20).

While it appears paradoxical that high contraceptive use goes together with high unintended pregnancy and high abortion rates among abused women, this finding may be related to the influence of GBV on the effective use of contraception. The environment of fear and male dominance evident in abusive relationships may reduce women's ability to use contraception consistently and effectively, thereby increasing their risk of unintended pregnancy and abortion (22). In addition, evidence suggests that contraceptive failure is relatively common in Vietnam, where an estimated 36.7% of unintended pregnancies are the result of contraceptive failure, which helps explain the country's high rates of contraceptive use and abortion (32). Another recent publication from this study found that family planning services in Thai Nguyen remain limited in terms of counseling about method choice as well as contraceptive access and method availability (33). Inadequate family planning services may contribute to higher rates of contraceptive failure in this population.

The other logical hypothesis is that women exposed to GBV have reduced control over their reproductive choices due to reduced access to family planning or other fertility control resources and therefore experience higher rates of unintended pregnancy. Abortion may be one method by which these women regain control of their reproductive health. Finally, there remains the possibility that abusive partners have used violence or the threat of violence to coerce these women into having abortions.

This study contains limitations that should be taken into consideration. Due to the study's cross-sectional design, it is impossible to surmise causality from the findings. Only the associations between having experienced GBV and induced abortion can be inferred. Another limitation is the possibility of underreporting of history of GBV and induced abortion due to our reliance on retrospective reports as well as the sensitive nature of the questions. In particular, younger women are likely to have underreported their experience with abortion since having an abortion before the first birth is viewed as disgraceful in Vietnam. In addition, our data did not include single women, whose GBV experiences and induced abortion behaviors could be different from those of married women. However, the results can be generalized to married or partnered women of reproductive age in Thai Nguyen province as well as the wider region of northern Vietnam. Despite these limitations, the study findings have important implications for policy and practice in Vietnam.

The Vietnam family planning program has made impressive gains in increasing contraceptive use and reducing fertility (34). However, family planning may not be easy for women in violent relationships, and therefore policies and programs in Vietnam should consider the role of GBV in women's reproductive health. At present, there are no national treatment guidelines for health care providers to use in screening, treating, and referring GBV victims. Nationwide interventions should focus on providing guidelines and training to health workers for screening and treatment of GBV during routine family planning and antenatal care visits. Because abused women also need support and protection outside of health care settings, referrals should be provided to appropriate support centers if abuse is identified. As of now, these centers are severely lacking; it is critical to establish effective protection, support, counseling, and treatment services for abused women.

While the National Assembly issued the Law on Domestic Violence Prevention and Control in 2007, knowledge of the law is still lacking and implementation remains weak, especially in rural and poor areas (35). Furthermore, the culture of silence around GBV means that many of the serious consequences go unrecognized by policymakers and the community at large. Efforts should be made to increase public awareness of the physical, reproductive, mental, and societal consequences of GBV and to shift norms of acceptance of GBV. To transfer some of the burden away from women, it is important to hold police and local authorities responsible for the implementation of GBV policies and legislation.

Conclusions

This study is among the first to provide evidence of a positive association between GBV and induced abortion in Vietnam. In addition, the study provides insight into the role that contraceptive use and unintended pregnancy play in the underlying pathway between GBV and induced abortion. Addressing Vietnam's high prevalence of GBV has the potential to significantly improve maternal, infant, and reproductive health outcomes. In particular, reducing GBV could lower the country's stubbornly high abortion rates. Our findings emphasize the importance of reducing gender inequality, integrating routine screening and treatment of GBV into reproductive health care, providing supportive services for victims of violence, and raising awareness of the extent and consequences of GBV. Additional research is needed to further the development and improvement of policies and programs. Qualitative studies are needed to probe into the psychology and contextual factors of GBV and reproductive health in greater detail to provide a more comprehensive explanation of the mechanisms at play. Research is also needed to test the practicality and effectiveness of interventions that aim to prevent and treat GBV as a means for improving reproductive health outcomes in a Vietnamese context.

Acknowledgements

Research was conducted by Thai Nguyen University of Medicine and Pharmacy, the Institute of Social and Medical Studies, and Thai Nguyen Provincial Health Services with support from Population Council Vietnam under the Primary Health Improvement Initiative. Funding was provided by Atlantic Philanthropies through Population Council Vietnam.

Conflict of interest and funding

The authors have not received any other funding or benefits from industry or elsewhere to conduct this study.

References

  1. García-Moreno C, Jansen HAFM, Ellsberg M, Heise L, Watts C. WHO multi-country study on women's health and domestic violence against women. Geneva, Switzerland: World Health Organization; 2005.
  2. Gardsbane D, Ha VS, Taylor K, Chanthavysouk K. Gender based violence: issue paper. Ha Noi: United Nations Population Fund; 2010.
  3. Sarkar NN. The impact of intimate partner violence on women's reproductive health and pregnancy outcome. J Obstet Gynaecol 2008; 28: 266–71. [Crossref]
  4. Chambliss LR. Intimate partner violence and its implication for pregnancy. Clin Obstet Gynecol 2008; 51: 385–97. [Crossref]
  5. Alio AP, Nana PN, Salihu HM. Spousal violence and potentially preventable single and recurrent spontaneous fetal loss in an African setting: cross-sectional study. Lancet 2009; 373: 318–24. [Crossref]
  6. O'reilly R. Domestic violence against women in their childbearing years: a review of the literature. Contemp Nurse 2007; 25: 13–21. [Crossref]
  7. Kaye DK, Mirembe FM, Bantebya G, Johansson A, Ekstrom AM. Domestic violence as risk factor for unwanted pregnancy and induced abortion in Mulago Hospital, Kampala, Uganda. Trop Med Int Health 2006; 11: 90–101. [Crossref]
  8. Emenike E, Lawoko S, Dalal K. Intimate partner violence and reproductive health of women in Kenya. Int Nurs Rev 2008; 55: 97–102. [Crossref]
  9. Silverman JG, Gupta J, Decker MR, Kapur N, Raj A. Intimate partner violence and unwanted pregnancy, miscarriage, induced abortion, and stillbirth among a national sample of Bangladeshi women. BJOG 2007; 114: 1246–52. [Crossref]
  10. Budiharsana MP, Mai TQ. Project evaluation report: improving the health care response to gender-based violence – phase II. Vietnam: Population Council; 2009.
  11. GSO. Results from the national study on domestic violence against women in Vietnam 2010. Hanoi, Vietnam: General Statistics Office; 2010.
  12. Dang T, Nguyen D. Vietnam Abortion Situations: Country Report, 2001. Paper for the Conference on Expanding Access: Advancing the Roles of Midlevel Providers in Menstrual Regulation and Elective Abortion Care, South Africa, 2–6 December 2001.
  13. Johnston R. Historical abortion statistics, Vietnam. Available from: http://www.johnstonsarchive.net/policy/abortion/ab-vietnam.html [cited 15 August 2012].
  14. WHO. Maternal Mortality in Viet Nam 2000–2001: An in-depth analysis of causes and determinants. Manila, Philippines: World Health Organization, Regional Office for the Western Pacific; 2005.
  15. Yoshikawa K, Agrawal NR, Poudel KC, Jimba M. A lifetime experience of violence and adverse reproductive outcomes: findings from population surveys in India. Biosci Trends 2012; 6: 115–21.
  16. Stockl H, Filippi V, Watts C, Mbwambo JK. Induced abortion, pregnancy loss and intimate partner violence in Tanzania: a population based study. BMC Pregnancy Childbirth 2012; 12: 12. [Crossref]
  17. Wu J, Guo S, Qu C. Domestic violence against women seeking induced abortion in China. Contraception 2005; 72: 117–21. [Crossref]
  18. Gorbach PM, Hoa DT, Nhan VQ, Tsui A. Contraception and abortion in two Vietnamese communes. Am J Public Health 1998; 88: 660–3. [Crossref]
  19. Goodkind D. Abortion in Vietnam: measurements, puzzles, and concerns. Stud Fam Plann 1994; 25: 342–52. [Crossref]
  20. Hindin MJ, Kishor S, Ansara DL. Intimate partner violence among couples in 10 DHS countries: predictors and health outcomes. Calverton, Maryland, USA: Macro International; 2008.
  21. Miller E, Jordan B, Levenson R, Silverman JG. Reproductive coercion: connecting the dots between partner violence and unintended pregnancy. Contraception 2010; 81: 457–9. [Crossref]
  22. Pallitto CC, Campbell JC, O'Campo P. Is intimate partner violence associated with unintended pregnancy? A review of the literature. Trauma Violence Abuse 2005; 6: 217–35. [Crossref]
  23. Fanslow J, Silva M, Whitehead A, Robinson E. Pregnancy outcomes and intimate partner violence in New Zealand. Aust N Z J Obstet Gynaecol 2008; 48: 391–7. [Crossref]
  24. Janssen PA, Holt VL, Sugg NK, Emanuel I, Critchlow CM, Henderson AD. Intimate partner violence and adverse pregnancy outcomes: a population-based study. Am J Obstet Gynecol 2003; 188: 1341–7. [Crossref]
  25. Vyas S, Kumaranayake L. Constructing socio-economic status indices: how to use principal components analysis. Health Policy Plan 2006; 21: 459–68. [Crossref]
  26. Gwatkin D, Rutstein S, Johnson K, Suliman E, Wagstaff A, Amouzou A. Socio-economic differences in health, nutrition, and population within developing countries: an overview. Niger J Clin Pract 2007; 10: 272–82.
  27. StataCorp. Stata Statistical Software: Release 11. College Station, Texas 77845 USA. Copyright 2009 StataCorp LP. 2009.
  28. Schuler SR, Anh HT, Ha VS, Minh TH, Mai BTT, Thien PV. Constructions of gender in Vietnam: in pursuit of the “Three Criteria”. Cult, Health Sex 2006; 8: 383–94. [Crossref]
  29. Alio AP, Salihu HM, Nana PN, Clayton