1 Faculty of Public Health, Hanoi Medical University, Vietnam, 2 Department of Social Medicine, Gothenburg University, Sweden, 3 Department of Public Health Science, Karolinska Institutet, Sweden and 4 Department of Science and Training, Ministry of Health, Vietnam
* Author to whom correspondence should be addressed at: Department of Public Health Sciences, Norrbacka, Karolinska Instituet, SE-171 76 Stockholm, Sweden. E-mail: Peter.Allebeck{at}phs.ki.se
(Received 22 June 2005; first review notified 06 July 2005; in revised form 21 July 2005; accepted 26 July 2005)
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In Vietnam, hospital records have indicated an increase of severe mental disorders due to alcohol misuse during the past decades (Thiem, 2004). The Vietnamese government also has started to pay attention to policy development to reduce the health and social burden due to alcohol misuse (National Commission of Social Affairs of Vietnam, 2003
). However, data on alcohol consumption and alcohol problems in the general population are limited, as is the case in most developing countries. Knowledge of the pattern of drinking and problem levels in the community is important for planning intervention and treatment programs. Thus, there is a need to develop and adapt instruments to measure alcohol problems in the general population in Vietnam. Instruments are also needed to help general practitioners and other health care workers to identify excessive drinkers and reduce problem drinking.
Many instruments have been developed over the past decades to assess alcohol consumption and alcohol problems (National Institute on Alcohol Abuse and Alcoholism, 1995). However, the use of these is particularly limited in developing countries (Babor et al., 2001
). There is a need to adapt these and develop capacity for using these at the outpatient and community level in developing countries, and the next step is to implement the use of these as a part of programmes for secondary prevention in the health services. AUDIT has been found in many studies to be a valid and convenient instrument to screen for alcohol problems in primary care settings in many countries (Saunders et al., 1993b
; Allen et al., 1997
; McPherson and Hersch, 2000
; Babor et al., 2001
; Reinert and Allen, 2002
).
The present study is part of an epidemiological community survey in rural Vietnam. A number of health problems have been addressed through regular screening surveys of the general population (Chuc and Diwan, 2003). Since mental health problems and alcohol abuse by local people were reported to be important health problems in the area, although not systematically analysed, development work is ongoing to assess these problems in the community.
This study aims (i) to adapt the Vietnamese version of AUDIT for identifying alcohol problems in Vietnam; (ii) to assess the accuracy and performance of AUDIT in detecting alcohol problems, as defined by ICD-10 and DSM-IV diagnostic criteria, in a rural setting in Vietnam.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In this district, almost 80% of the population are farmers, some are handicraft-makers (6%), small traders (3%), and government staff (2%). The illiteracy rate among people >15 years is 0.4%. As in other rural areas of Vietnam, the most common use of alcoholic beverage is rice wine, which generally is self-produced, sold cheaply, and is easily accessible.
Translation of instruments
The AUDIT instrument consists of 10 questions, which have a possible maximum score of 40. The questions 18 may score 04 points, and questions 9 and 10 may score 0, 2, or 4 points. The English version of AUDIT was translated in to Vietnamese. Back-translation procedures were performed through several steps in order to keep the translated version in concordance with the original and to be adaptable to Vietnamese culture. We followed the procedures described by Room et al. (1996).
The Composite International Diagnostic Interviews (CIDIs) (12 months) were used as a validation instrument. The CIDI was developed by the World Health Organisation to assess different mental disorders according to the definition and criteria in the International Classification of Diseasesversion 10 (ICD-10) and to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). Section J consists of 23 questions regarding alcohol use, alcohol abuse, withdrawal, and alcohol dependence conditions. The English version of CIDI, 12-month core version 2.1, section J was translated in to Vietnamese, was checked and refined in several steps of back-translation in a similar process as for the Vietnamese version of AUDIT described above.
Sampling
A total of three communes representative of three main geographical areas (lowland, highland, and mountainous) of the Bavi district were randomly selected from 32 communes. In these communes 500 people aged from 18 to 60 years were randomly selected using an identification number list that is available from the demographic surveillance system. Of the 500 subjects invited to participate in the study, only fifteen were absent during the data collection. The distribution with regard to sex, age, and other demographic characteristics did not differ significantly from the total population.
Data collection
Interviewers who work for the longitudinal epidemiological surveillance system performed the data collection. The interviewers were inhabitants of the district and had high school education or finished secondary medical education. Six interviewers were trained to perform interviews with AUDIT and eight interviewers were trained to assess alcohol use and alcohol problems by interviews with the CIDI 2.1, section J. The CIDI interviews were performed shortly after the interviews with AUDIT, and the interviewers were blind to the preceding interview results. The interviewers used hand cards to facilitate the interviewee's estimation of alcohol intake. The hand card had pictures of common beverages in the setting and the corresponding units for standard drinks.
The data collection was performed during 3 months from December 2002 to February 2003.
This study was approved both by the Research Ethics Committee at Gothenburg University and by the Scientific and Ethical Committee at Hanoi Medical University. All selected interviewees were informed about the study and gave consent to participate.
Definitions
Standard drink refers to the amount of 12.6 g pure alcohol, which equals 330 ml of 5% beer, or 40 ml of 40% liquor, or 130 ml of 12% wine, etc.
Measures of alcohol problems
Non-drinkers were defined as having consumed less than a total of 12 standard drinks during the previous 12 months. Non-problem drinkers were people who used alcohol but did not meet the criteria for problem drinkers defined as at-risk drinkers, as harmful use/alcohol abuse, or as alcohol-dependent. According to the level of alcohol consumption in CIDI interviews, we defined at-risk drinkers as people who drank a daily average of >30 g alcohol (2.4 standard drinks) for a man and >20 g (1.6 standard drinks) for a woman. Harmful use was defined according to the criteria for harmful use in the ICD-10 and refers to alcohol consumption that results in consequences for physical and mental health. Alcohol abuse was defined according to the criteria in DSM-IV, which include also the social consequences caused by alcohol consumption. Alcohol dependence was classified according to the ICD-10 and DSM-IV criteria. These include a strong desire to drink alcohol, impaired control over its use, persistent drinking despite harmful consequences, a higher priority given to drinking over other activities, increased alcohol tolerance, and a physical withdrawal reaction when alcohol use was discontinued (Hasin, 2003).
Educational level was divided into three categories: primary consisted of people who had one to five schooling years; secondary included people who had six to nine schooling years; high school referred to individuals with ten or more than ten years of schooling. Seventy-one percent of the population was counted as farmers since they were mainly doing farming. Due to a few observations, the rest were classified as non-farmers, which included government staff, small traders, handicraft-makers, housewives, etc.
Statistical analyses
Data analyses were performed using STATA 8.0. First, we examined the validity of AUDIT in detecting at-risk drinkers. Then, the validity of AUDIT was assessed using ICD-10 diagnostic criteria for harmful use and alcohol dependence. We also performed a validation using DSM-IV criteria for alcohol abuse and alcohol dependence.
The agreements between ICD-10 and DSM-IV in giving diagnosis of harmful use/alcohol abuse and alcohol dependence were examined by the Kappa test. We calculated sensitivity, specificity, and correct classification rate of AUDIT compared to ICD-10 and DSM-IV. We estimated the means of AUDIT scores with 95% confidence interval, adjusted for age. We compared results of AUDIT in different categories of alcohol use, and tested the differences in means of AUDIT score by analysis of covariance (ANOVA test), adjusting for age, and socio-economic and demographic groups.
Receiver operating characteristic (ROC) analysis was applied to assess the ability of AUDIT in discriminating cases and non-cases of at-risk drinking, harmful use, alcohol abuse, and alcohol dependence. ROC curves were obtained by plotting sensitivity against the false positive (1-specificity) at each cut-off point. The area under the ROC curve (AUC) was used as an indicator of test performance. Values of this area range from 0.5 to 1. A value of 1 indicates that the instrument gives a perfect discrimination between case and non-case, and a value of 0.5 implies an ability to discriminate no better than chance (Hajian-Tilaki et al., 1997). We calculated the AUC of AUDIT for detecting at-risk drinking, harmful use, alcohol abuse, and alcohol dependence. Firstly, we evaluated the ability of AUDIT to discriminate between non-problem drinkers (including non-drinkers and non-problem drinkers) and at-risk drinkers. Then, we assessed the ability of AUDIT in differentiating subjects with harmful use from subjects without harmful use and without alcohol dependence. We finally evaluated the ability of AUDIT to discriminate between alcohol dependence (according to ICD-10) and all other categories. Similar analyses were also applied for the use of AUDIT in detecting alcohol abuse and alcohol dependence using DSM-IV criteria.
We further compared the AUCs in different population groups with regard to educational level, occupation, marital status, and age. Since the prevalence of drinking among women was very low, we only report the validity of the instrument in detecting alcohol problems in men.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
Due to few cases of alcohol problems observed among women (1/282), we could only evaluate the validity of AUDIT in men. Table 3 shows crude and age-adjusted mean of AUDIT score for alcohol problems, as well as results from the covariance analysis. The AUDIT scores were higher for alcohol problems. Age-adjusted mean scores of AUDIT increased markedly by increasing level of alcohol problems. Similar results also were found when DSM-IV criteria were followed. The AUDIT scores seemed to be influenced by age since age-adjusted means differed from crude means. The analysis of covariance with age as covariate found only alcohol problem as the main effect. There was a statistically significant interaction between alcohol problem and age, while there was no other significant interaction effect. The analyses using DSM-IV classification gave almost the same results to the ICD-10.
|
|
|
Compared to DSM-IV criteria, the AUDIT had somewhat lower sensitivity but higher specificity for predicting alcohol dependence at all thresholds. AUDIT had a lower sensitivity for detecting alcohol abuse compared to that for harmful use.
Figure 1 presents the trade-off between sensitivity and false positive (1-specificity) at different cut-off points and the AUCs of AUDIT for at-risk drinkers, harmful use, alcohol abuse, and alcohol dependence. The smallest AUC was 0.82 (0.720.91) for alcohol abuse and the largest was 0.91(0.840.98) for harmful use. There was no significant variation in AUC by age, educational level, occupation, and marital status (data not shown).
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
According to our results, the AUDIT score significantly increased by increasing the level of alcohol problems (from at-risk drinking to harmful use/alcohol abuse, and to alcohol dependence). This suggests that the AUDIT is valuable instrument in recognizing alcohol problems in our setting. Robert et al. indicated significant interactions between alcohol problem and gender, and between alcohol problem and ethnicity, while we found an interaction between age and alcohol problem, and between occupation and age (Volk et al., 1997). Higher prevalence of alcohol problems among men, and more current drinking among farmers are possible explanations for our findings.
In our study, the cut off point of 7/8 gave an acceptable sensitivity and specificity for detecting alcohol problems. AUDIT had a lower sensitivity for prediction of alcohol abuse, but had high sensitivity for detecting harmful use. Difficulty in defining alcohol as cause of medical problems might be the explanation for this. In several studies from Australia, Bulgaria, Kenya, Mexico, Norway, USA (Saunders et al., 1993b; Bradley et al., 1998
; Daeppen et al., 2000
; Fiellin et al., 2000
; Reinert and Allen, 2002
) Hong Kong (Leung and Arthur, 2000
), the cut-off point of 7/8 had a sensitivity and specificity of around 80%. Some studies even reported such high sensitivity and specificity at a cut-off of 4/5 (Piccinelli et al., 1997
; Daeppen et al., 2000
). For instance, Piccinelli et al. (1997)
reported a sensitivity of 84% and a specificity of 90%. However, Hans et al. in Germany found a low sensitivity of 33%; 37% for at-risk drinking and for current alcohol misuse at the cut-off point of 7/8 (Rumpf et al., 2002
). The cut-off point of 7/8 has been found and used as an optimal threshold in several studies, and was called "standard cut-off point" elsewhere (Conigrave et al., 1995a
). The cut-off of 7/8 has been suggested for use in primary care (Babor et al., 2001
).
Choosing a cut-off point normally depends on the purpose of each specific work. Lower cut-off points are preferably used in general population surveys and for two-stage surveys. Higher cut-off points can be used when the aim is to identify severe cases who need further diagnostic evaluation and treatment (Babor et al., 2001).
We used both ICD-10 and DSM-IV to validate AUDIT. The Kappa test showed a high level of agreement between two sets of criteria for dependence (0.98) but lower agreement for alcohol abuse (0.68). This is understandable since ICD-10 just limits harmful use to the medical problems caused by alcohol, while DSM-IV includes social problems resulting from alcohol (Hasin, 2003).
A limitation was the small number of women (n = 282), among whom the prevalence of alcohol problems was low. We could thus analyse performance of AUDIT only in men. Also, the study was conducted in a typical rural community in the North of Vietnam that might not be representative for other communities, e.g. urban, in Vietnam.
In conclusion, we confirmed that AUDIT is feasible for use in a rural community in a developing country. Different cut-off points are appropriate for different purposes, but for general population screening of at-risk drinking we found a cut-off point of 7/8 to be optimal, and this will be used for further prevalence studies in this population.
![]() |
ACKNOWLEDGEMENTS |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Babor, T. F., Higgins-Biddle, J. C., Saunders, J. B. et al. (2001) AUDITThe Alcohol Use Disorders Identification Test-Guidelines for Use in Primary Care. World Health Organisation, Department of Mental Health and Substance Dependence, Geneva.
Bradley, K. A., Bush, K. R., McDonell, M. B. et al. (1998) Screening for problem drinking: comparison of CAGE and AUDIT. Ambulatory Care Quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification Test. Journal of General Internal Medicine 13, 379388.[CrossRef][ISI][Medline]
Cherpitel, C. J. (1998) Performance of screening instruments for identifying alcohol dependence in the general population, compared with clinical populations. Alcoholism: Clinical and Experimental Research 22, 13991404.[ISI][Medline]
Chuc, N. T. and Diwan, V. (2003) FilaBavi, a demographic surveillance site, an epidemiological field laboratory in Vietnam. Scandinavian Journal of Public Health Supplement 62, 37.[Medline]
Conigrave, K. M., Hall, W. D. and Saunders, J. B. (1995a) The AUDIT questionnaire: choosing a cut-off score. Alcohol Use Disorder Identification Test. Addiction 90, 13491356.[CrossRef][ISI][Medline]
Conigrave, K. M., Saunders, J. B. and Reznik, R. B. (1995b) Predictive capacity of the AUDIT questionnaire for alcohol-related harm. Addiction 90, 14791485.[CrossRef][ISI][Medline]
Daeppen, J. B., Yersin, B., Landry, U. et al. (2000) Reliability and validity of the Alcohol Use Disorders Identification Test (AUDIT) imbedded within a general health risk screening questionnaire: results of a survey in 332 primary care patients. Alcoholism: Clinical and Experimental Research 24, 659665.[CrossRef][ISI][Medline]
Fiellin, D. A., Reid, M. C. and O'Connor, P. G. (2000) Screening for alcohol problems in primary care: a systematic review. Archives Internal Medicine 160, 19771989.[CrossRef][ISI]
Hajian-Tilaki, K. O., Hanley, J. A., Joseph, L. et al. (1997) A comparison of parametric and nonparametric approaches to ROC analysis of quantitative diagnostic tests. Medical Decision Making 17, 94102.[ISI][Medline]
Hasin, D. (2003) Classification of alcohol use disorders. Alcohol Research and Health 27, 517.[ISI][Medline]
Leung, S. F. and Arthur, D. (2000) The alcohol use disorders identification test (AUDIT): validation of an instrument for enhancing nursing practice in Hong Kong. International Journal of Nursing Studies 37, 5764.[CrossRef][ISI][Medline]
McPherson, T. L. and Hersch, R. K. (2000) Brief substance use screening instruments for primary care settings: a review. Journal of Substance Abuse Treatment 18, 193202.[CrossRef][ISI][Medline]
Mei Chu Tsai, Y.-F. T., Ching-yen Chen, and Chia-Yih Liu (2005) Alcohol Use Disorders Identification Test (AUDIT). Alcoholism: Clinical and Experimental Research 29, 5357.[CrossRef][ISI][Medline]
National Commission of Social Affaires of Vietnam (2003) Record of Workshop on Policies for Alcohol Abuse Control (In Vietnamese). Vietnamese National Assembly, Hanoi.
National Institute on Alcohol Abuse and Alcoholism (1995) Assessing Alcohol ProblemsA guide for clinicians and researchers. Treatment Handbook. NIAAA, Bethesda, MD.
Piccinelli, M., Tessari, E., Bortolomasi, M. et al. (1997) Efficacy of the alcohol use disorders identification test as a screening tool for hazardous alcohol intake and related disorders in primary care: a validity study. British Medical Journal 314, 420424.
Reid, M. C., Fiellin, D. A. and O'Connor, P. G. (1999) Hazardous and harmful alcohol consumption in primary care. Archives Internal Medicine 159, 16811689.[CrossRef][ISI]
Reinert, D. F. and Allen, J. P. (2002) The Alcohol Use Disorders Identification Test (AUDIT): a review of recent research. Alcoholism: Clinical and Experimental Research 26, 272279.[ISI][Medline]
Room, R., Babor, T. and Rehm, J. (2005) Alcohol and public health. Lancet 365, 519530.[ISI][Medline]
Room, R., Janca, A., Bennett, L. A. et al. (1996) WHO cross-cultural applicability research on diagnosis and assessment of substance use disorders: an overview of methods and selected results. Addiction 91, 199220, discussion 221230.
Rumpf, H. J., Hapke, U., Meyer, C. et al. (2002) Screening for alcohol use disorders and at-risk drinking in the general population: psychometric performance of three questionnaires. Alcohol and Alcoholism 37, 261268.
Saunders, J. B., Aasland, O. G., Amundsen, A. et al. (1993a) Alcohol consumption and related problems among primary health care patients: WHO collaborative project on early detection of persons with harmful alcohol consumption-I. Addiction 88, 349362.[ISI][Medline]
Saunders, J. B., Aasland, O. G., Babor, T. F. et al. (1993b) Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection of Persons with Harmful Alcohol Consumption-II. Addiction 88, 791804.[ISI][Medline]
Thiem, N. V. (2004) Alcohol Abuse and Mental Health. National Workshop on Mental Health Care and Suicide Prevention (In Vietnamese). Ministry of Health of Vietnam, Hue.
Volk, R. J., Steinbauer, J. R., Cantor, S. B. et al. (1997) The Alcohol Use Disorders Identification Test (AUDIT) as a screen for at-risk drinking in primary care patients of different racial/ethnic backgrounds. Addiction 92, 197206.[CrossRef][ISI][Medline]
World Health Organisation (2002a) Alcohol in developing societies summary. Available at http://www.who.int/substance_abuse/publications/en/APDSSummary.pdf.
World Health Organisation (2002b) The World Health Report 2002: Reducing Risks, Promoting Healthy Life. WHO, Geneva.
|