Computerized administration of alcoholism screening tests in a primary care setting

This study compared the use of computerized versus paper-and-pencil administration of three commonly employed alcohol screening tests--the Short Michigan Alcoholism Screening Test (SMAST-13), CAGE (an acronym for four questions to discern problem drinking), and a quantity-frequency scale. These instruments were administered to 280 adults receiving health care in three primary care clinics in south-central Wisconsin. One hundred forty patients were randomly assigned to complete these instruments on a Macintosh SE, and 140 were assigned to the paper-and-pencil versions. Patients were classified as alcoholic based on responses to the Diagnostic Interview Schedule and DSM-III criteria. Results indicated the sensitivity and specificity of these instruments were similar for the two methods of administration. The sensitivity of the SMAST-13 was 0.56 for computer administration and 0.58 for the pencil-and-paper form. The findings suggest that computer administration of these instruments is at least as effective as use of the standard pencil-and-paper method. The data show that computers can be used for direct entry of information by patients, avoiding separate coding of paper-and-pencil information into a computerized format for clinical systems that use system-wide computerization of medical information.