Determination of a brief AUDIT screening questionnaire to identify women at risk of harmful and hazardous alcohol consumption in primary care settings

Master Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title

University of Cape Town

Abstract Background: Alcohol consumption in South Africa is a major contributing factor to the quadruple burden of disease. Additionally, South Africa has the highest rate of Foetal Alcohol Syndrome in the world. As effective interventions are available for hazardous and harmful drinking, screening for this behaviour in primary care can potentially contribute to improving health outcomes and reducing the cost to society. Existing validated screening questionnaires are available to identify those who drink at hazardous or harmful levels. However, no tools have been validated for women of childbearing age in South Africa. Any screening tools recommended for implementation should be as brief as possible with high sensitivity and specificity to justify the time spent on screening and minimise time spent on false positive results. Objectives: To identify the most appropriate existing abbreviated version of the Alcohol Use Disorder Identification Test (AUDIT) questionnaire for South African women aged 18-44 and determine whether adjustments are required for urban/rural residence or patterns of consumption (binge drinking vs non-binge drinking). Additionally, to determine whether a single question can be used to identify possible dependent drinking. Methods: An existing dataset was used for the analysis, collected by means of household surveys conducted in 2006 in rural areas of the Western Cape and urban areas in Gauteng, South Africa. The Western Cape sample was selected by stratified random sampling of farm workers and the Gauteng sample was selected by cluster random sampling. AUROC analysis was used to compare the abbreviated questionnaires (AUDIT-C, AUDIT-3, AUDIT-4, AUDIT-PC, AUDIT-QF), CAGE and single questions to results of the full AUDIT screening questionnaire as the gold standard. Data was stratified in relation to binge drinking and the analyses repeated to determine any effect of drinking patterns on the results. Results: AUDIT-4 and AUDIT-PC were the best performing brief questionnaires in both rural and urban settings. AUDIT-4 had AUCs of 97.52% and 96.87% in Western Cape and Gauteng respectively. AUDIT-PC had scores of 97.52% and 93.98% in Western Cape and Gauteng respectively. Stratification by drinking pattern did not substantially influence the results. The single question method of identification of possible dependent drinking is not supported by the results of this study, with AUCs of < 75%. Conclusion: AUDIT-4 appears to be the best brief questionnaire for the identification of hazardous and harmful drinking among women aged 18-44 in South Africa, with no adjustments required for urban/rural settings or drinking patterns.