Abstract
PURPOSE
Alcohol use disorders are common and poorly detected in most primary care settings. We implemented a new method of screening for alcohol use disorders among primary care patients.
METHODS
A health screening Interactive Voice Response (IVR) interviewing system (respondents answer pre-recorded questions via touch tone telephones) that includes the Alcohol Use Disorders Identification Test (AUDIT) was developed and administered in an urban primary care practice housed in a county hospital. Patients were contacted for screening approximately two weeks prior to their next scheduled visit. At the conclusion of the interview, patients heard on-line feedback advising them on safe-drinking behaviors (appropriate to their AUDIT score). Each patient's primary care physician was faxed a report of the screening results, clinical management recommendations, and referral resources.
RESULTS
Of 6035 eligible patients contacted, 86% (5174) agreed to begin the interview, resulting in 2997 interviews in which all necessary alcohol-related questions were answered (58% completion rate). Seventy-one percent of respondents were female; 58% white, 32% black, and 6% Hispanic; the mean age was 49 ± 14.8; insurance status was: 26% uninsured, 24% private, 17% managed care, 17% Medicaid, 16% Medicare. Screening results for 7.9% of respondents were within the hazardous range (AUDIT = 8–10) and 14.1% screened in the harmful range (AUDIT = *11). As expected, increased AUDIT scores (*8) were more prevalent among men (29.3% of men and 10.8% for women). After controlling for gender, the Cochran-Mantel-Haenszel test was used to compare AUDIT level across age groups (p = 0.001) and racial groups (p = 0.018). Hazardous drinking rates were similar across age groups (range: 4.9–7%) and racial groups (range; 4.1–7.9%). A larger proportion of respondents with AUDIT scores indicating harmful use of alcohol were between the ages of 26–55 (range: 14.4–19.8%), with the highest proportion falling within the 36–45 age group. Hispanic and Black respondents were most likely to report harmful drinking/dependence—16.6% and 16.4%, respectively. However, when rates of hazardous and harmful drinking behaviors are considered together, there was little difference among races (18.5 to 22.9%).
CONCLUSION
Using a computerized interview allowed us to screen many more patients for alcohol use disorders than would normally be screened in most primary care settings. The finding that 20% of our patients reported hazardous or harmful drinking behaviors alerts us to how common this stigmatized problem is. IVR systems should be considered as effective tools to screen patients for a prevalent, sensitive issue such as alcohol misuse. We are in the process of examining the effects of novel treatment approaches for alcohol use disorders in primary care.