Screening for substance use & misuse in the emergency department: Differences in computer prompted vs. physician face-to-face interview

Background: Data on sensitivity and validity of face-to-face substance abuse screening methods are mixed. Objective: To examine sensitivity of physician-elicited patient endorsements of substance use and abuse compared to computer-elicited responses. Methods: Participants were given a computerized interview regarding drug and alcohol use, including questions from the SCID-validated, DSM-based, Shedler Quick Psycho-Diagnostics Panel and standard NIAAA quantity-frequency questions. ED physicians used drug and alcohol questions in NIAAA format to interview participants face-to-face. Physician interview data were collected via structured chart review. Heavy drinkers were defined according to NIAAA cut-offs. Frequencies and proportions were calculated to examine discordance between computer interview and physician interviews. Results: 768 patients were enrolled during random time blocks between June and July of 2006 from the emergency department at Parkland Hospital in Dallas, TX. DSM-IV criteria for drug abuse and/or report of illicit drug use was met or endorsed by 175 (22.8%) participants in the computerized interview, while only by 43 (6.3%) participants in physician interview. 24 participants screened positive for substance use and/or abuse via both the computer and the physician. In computerized interview, 94 (12.2%) met NIAAA criteria for heavy drinking, while 154 (20.1%) met these same criteria in physician interview. 30 participants screened positive for heavy drinking in both computerized and physician interviews. Conclusions: Both computerized questioning and physician questioning are imperfect methods of capturing drug and alcohol data. In this sample the computerized questionnaire captured a significantly greater proportion of positive endorsements for drugs, while physician interview captured more participants for alcohol.