Physicians in the general medical setting commonly encounter but rarely counsel patients with dependent or harmful drinking behaviors. We tested whether providing physicians with their patients' results on the alcohol module of the Diagnostic Interview Schedule and counseling directives would prompt them to counsel these patients.
We randomly assigned 83 first-, second-, and third-year medical residents to receive or not to receive diagnostic information and counseling directives on 214 patients who reported at least one symptom of alcohol impairment as defined in the Diagnostic and Statistical Manual of Mental Disorders, Third Edition. Using binary logistic regression, we examined the effect of specific covariables on rates of physician counseling. These variables included physician information status, patient gender, and drinking disorder severity and recency. We also examined the effect of physician prompting on counseling of female patients, patients with inactive disorders, and nondependent but harmful drinkers. We determined counseling by postvisit patient interviews.
Physician prompting, dependent drinking, and recent disorder activity were significant correlates of physician counseling (P<.05), while male gender was a marginally significant correlate (P=.08). Informed physicians counseled female patients, harmful but nondependent drinkers, and patients with inactive disorders more often than their uninformed colleagues, although only the last variable achieved statistical significance.
Providing physicians with the results of the Diagnostic Interview Schedule and counseling directives resulted in short-term improvement in their rates of counseling patients with a history of dependent or nondependent but harmful drinking. Further research is necessary to determine long-term gains in rates of physician counseling and improvements in the course of these patients.(Arch Intern Med. 1993;153:1573-1577)