Physicians' decisions to prescribe benzodiazepines for nervousness and insomnia.

OBJECTIVE: To assess the effects of particular clinical cues on decisions about prescribing benzodiazepines. DESIGN: A factorial survey based on social judgment theory. SETTING: A midwestern U.S. medical school. PARTICIPANTS: Physicians (n = 115) recruited from the staff by invitation and interview. MEASUREMENTS AND MAIN RESULTS: Physicians indicated their level of agreement with prescribing a benzodiazepine for 24 hypothetical cases of nervousness and insomnia. The cases stemmed from the same scenario but varied systematically with regard to psychiatric diagnosis, recent ability to work, and long-term social stability. A fourth cue, called "health status," covertly depicted the presence or absence of three common alcohol-related medical problems. One fourth of the physicians agreed with prescribing for 15 or more cases, and 15% disagreed for all of them. Agreement was cumulative and least common for major depression, more common for adjustment disorder, and most common for generalized anxiety. Agreement with prescribing for cases with alcohol-related medical problems was 14% less than that for cases without them. Over half the physicians agreed with prescribing for 4 or more of the 12 cases with alcohol-related medical problems. CONCLUSIONS: Prescribing decisions varied widely. Some physicians avoided benzodiazepines unnecessarily for some cases, while others agreed with prescribing for patients with a high probability of alcohol abuse. Blanket calls for more or less prescribing are overly simplistic; physicians should be able to recognize substance use disorders among anxious patients and make prescribing decisions based on relevant literature and clinical cues.