Validation and Comparison of Alcohol-Screening Instruments for Identifying Hazardous Drinking in Hospitalized Patients in Taiwan

Alcohol & Alcoholism Vol. 43, No. 5, pp. 577–582, 2008 Advance Access publication 8 May 2008 doi: 10.1093/alcalc/agn036 ASSESSMENT AND DETECTION Validation and Comparison of Alcohol-Screening Instruments for Identifying Hazardous Drinking in Hospitalized Patients in Taiwan Shu-I Wu1,2 , Hui-Chun Huang2,3 , Shen-Ing Liu1,2,3,∗ , Chiu-Rong Huang3 , Fang-Ju Sun3 , Tse-Yun Chang2,4 , Shou-Chuan Shih5 and Kuo-Shyang Jeng6 1 2 3 Department of Psychiatry, Mackay Memorial Hospital, Taipei, Taiwan Mackay Medicine, Nursing and Management College, Taipei,5 Taiwan Department of 4 Medical Research, Mackay Memorial Hospital, Taipei, Taiwan Department of Nursing, Mackay Memorial Hospital, Taipei, Taiwan Department of Internal 6 Medicine, Mackay Memorial Hospital, Taipei, Taiwan Department of Surgery, Mackay Memorial Hospital, Taipei, Taiwan ∗ Corresponding author: Dr. Shen-Ing Liu, Department of Psychiatry, Mackay Memorial Hospital, 45 Ming-Sheng Road, Tamshui, Taipei County 251, Taiwan. Tel: +886-2-2809-4661 extension 2733; Fax: +886-2-2809-5679; E-mail: , (Received 8 October 2007; first review notified 12 December 2007; in revised form 9 April 2008; accepted 10 April 2008; advance access publication 8 May 2008) Abstract — Aim: The aim of this study was to validate the Mandarin Chinese version of different screening instruments and compare their performances for identifying hazardous drinkers in Taiwan. Methods: We compared the performance of the Mandarin Chinese versions of AUDIT, AUDIT-C (AUDIT items 1, 2 and 3), AUDIT-4 (AUDIT items 1, 2, 3 and 10), AUDIT-3 (AUDIT item 3), TWEAK, SMAST and CAGE to detect hazardous drinking in hospitalized patients in Taiwan. The results of the test instruments were blindly compared with the reference standard Schedule for Clinical Assessments in Neuropsychiatry (SCAN). Results: Of 404 patients evaluated, 100 were identified as having a hazardous drinking pattern. All screening instruments showed acceptable sensitivities (ranging from 85 to 93%) and specificities (ranging from 72 to 92%), but AUDIT and its short forms performed consistently better than the other instruments. Conclusions: The Mandarin Chinese versions of AUDIT and its derivatives perform well in screening hospitalized Taiwanese patients for hazardous drinking. INTRODUCTION The apparent increased prevalence over the past five decades of alcohol-use disorders, that is alcohol abuse and alcohol dependence, is an important public health issue in Taiwan (Liu and Cheng, 1998). Hazardous drinking (also referred to as problem, heavy, at-risk or excessive drinking) (Reid et al., 1999) is a pattern of drinking that puts people at increased risk for an overt alcohol-use disorder, with all the attendant physical, legal, psychosocial and economic costs (Dawson and Archer, 1993; National Institute on Alcohol Abuse and Alcoholism, 1995; Wechsler et al., 1994). The prevalence of alcohol-related problems identified in patients in general hospitals, ranging from 7.4% to 28% (Schneekloth et al., 2001), is higher than that in community surveys. This is not particularly surprising, as people with alcohol-use disorders often seek help only when they are ill, whether or not their illness is alcohol related. Hospitalization thus provides a window of opportunity for early identification of and brief interventions for people whose drinking may put them at risk. In the busy clinical setting, however, medical staff often do not recognize a hazardous drinking pattern unless there are obvious physical or psychosocial effects related to alcohol abuse. Many studies have found that less than a third of such individuals are identified in general hospitals (Wu et al., 2006), and we found that only 5–10% of patients were referred to the psychiatric service for further evaluation and treatment. (In Taiwan, individuals with alcohol-use disorders are generally referred to a psychiatrist for management.) Failure to identify these problems in hospitalized patients is exceedingly costly in terms of both morbidity and expense. It is increasingly recognized that people with hazardous drinking patterns may be responsive to early intervention before an overt alcohol-use disorder develops. As a consequence, screening, early recognition and brief interventions for hazardous drinking before irreversible complications set in is a potentially effective sec- ondary prevention (Adams et al., 1996; Cherpitel, 1995; Volk et al., 1997). A wealth of screening instruments have been validated for detecting alcohol-related problems in various clinical settings, including questionnaires that can be self-administered or used by an interviewer. The most widely used are the 4-item CAGE, 25-item Michigan Alcoholism Screening Test (MAST) or the shorter 13-item SMAST, TWEAK and the Alcohol Use Disorder Identification Test (AUDIT) (Fiellin et al., 2000). MAST, SMAST and CAGE are good for detecting lifetime alcohol problems but do not assess pre-symptomatic hazardous drinking patterns. TWEAK, a combination of selected items from the MAST and CAGE, was originally developed for the detection of hazardous drinking during pregnancy (Russell et al., 1994). AUDIT, designed by the World Health Organization (WHO), has good validity in detecting current hazardous drinking patterns as well as more severe alcohol-related problems (Fiellin et al., 2000; Reinert and Allen, 2002, 2007). These instruments were all developed in the West and therefore need to be translated and validated for use in other populations. The ethnic Chinese constitute a plurality of the world’s population, so valid alcohol-screening instruments are extremely important. Taiwan is situated 100 miles off the southeast coast of China and has a population of 23 million. Most Taiwanese are ethnic Chinese who emigrated from China many years ago. Mandarin (a dialect originating in northern China) is the official language in Taiwan (Liu et al., 2005). Thus far, Kuo et al. had validated the Mandarin Chinese version of CAGE for identifying alcohol abuse or dependence among hospital inpatients (Kuo et al., 1999). Tsai et al. recommended cutoff points for the full AUDIT and for the first three questions (AUDITConsumption or AUDIT-C) based on a study of identifying harmful or dependent drinkers among inpatients (Tsai et al., 2005). Nevertheless, no studies have compared the performance of the Mandarin Chinese versions of CAGE, TWEAK, SMAST,  C The Author 2008. Published by Oxford University Press on behalf of Medical Council on Alcohol. All rights reserved Wu et al . 578 AUDIT, AUDIT-C, AUDIT-4 (a combination of AUDIT questions 1, 2, 3, 10) and AUDIT-3 (AUDIT question 3 as a single screening tool) in a Chinese population. The aims of this study were to compare those tests in hospitalized patients in Taiwan and determine appropriate cutoff points for identifying hazardous drinking. SUBJECTS AND METHODS Subjects This research was conducted in a general hospital in Taipei, a city located i (...truncated)