Type of publication:
*Walsh S.; Haroon S.; Nirantharakumar K.; Bhala N.
The Lancet; Nov 2017; vol. 390
Background Alcohol misuse is common among patients accessing secondary care and an important cause of premature disability and death. The objective of this review was to summarise approaches to alcohol screening in secondary care, including the diagnostic accuracy, uptake, yield, and implementation challenges. Methods Search terms for alcohol, screening, and secondary care were combined in Medline, EMBASE, and other bibliographic databases for English language studies published from Jan 1, 2000, to Sept 4, 2015. We included studies and reviews of any design that evaluated alcohol screening with questionnaires or biochemical tests among adolescents and adults in secondary care. The primary outcomes were the diagnostic accuracy, uptake, and yield of alcohol screening. A random-effects proportion meta-analysis summarised screening uptake and yield, stratified by clinical setting. Findings 97 articles met the inclusion criteria and were included, with data from 1 213 761 screened patients. The Alcohol Use Disorders Identification Test (AUDIT) and AUDITConsumption (AUDIT-C) were the most widely validated screening tests and demonstrated high diagnostic accuracy and uptake. Overall, uptake for alcohol screening in secondary care was 79% (95% CI 74-84; n=45 studies, 477 533 of 604 471 screened patients) and the highest uptake was in outpatient clinics at 91% (82-96; n=8, 208 245 of 228 841). Overall, the proportion of patients screening positive for alcohol misuse was 28% (23-32, n=44; 135 741 of 484 788) with the highest yield in outpatient clinics of 31% (22-41; n=7, 72 270 of 233 128). However, there was significant heterogeneity in estimates of both uptake and yield of alcohol screening (I2>90%). Interpretation Alcohol screening in secondary care is likely to have a high uptake and yield, particularly in outpatient clinics. AUDIT and AUDIT-C are the most widely validated screening tools for alcohol misuse in secondary care and have high diagnostic accuracy and uptake. The review included a large number of studies, and a range of clinical settings and patient groups, strengthening the generalisability of the findings. However, a systematic assessment of risk of bias was not conducted and study selection was performed by one reviewer. Further research is needed to evaluate the cost-effectiveness of alcohol screening in secondary care.