Previous studies suggest that physicians' fear of addiction and regulatory mechanisms may inhibit prescribing of opioids and benzodiazepines. The magnitude of such inhibition, the extent to which it disserves patients with legitimate need, and its responsiveness to educational interventions are unknown. Two separate, co-administered, three-phase studies - one on opioid prescribing and one on benzodiazepine prescribing - will investigate the determinants of physicians' prescribing decisions and their amenability to an educational intervention. Random stratified selection from a national physician list will proceed such that 40 physicians represent each of 48 combinations of six specialties, four geographic regions, and states that do and do not have multiple copy prescription requirements (MCPR's) for scheduled medications. In Phase 1, the subjects will indicate their prescribing decisions for clinical vignettes that vary systematically by medical diagnosis, alcohol disorder, gender, and race. A logit multilevel hierarchical analysis and a cluster analysis will discern the subjects' decisionmaking policies from associations of their prescribing decisions with the three physician-based and four vignette-based variables. In Phase 2, a random half of the subjects who exhibited each policy will receive written, individualized feedback on their responses from Phase I and a relevant monograph. In Phase 3, the subjects will respond to the same clinical vignettes as in Phase I. The intervention will be assessed by comparison of pre- and post- intervention prescribing decisions and process data. The results of the randomized controlled trial will be analyzed by a structural equation model that accounts for differential historical effects. Sub- studies nested in Phase I will assess physicians' self-reported predictions of how their prescribing would change if MCPR's would be instituted in states without them, or if MCPR' s would be repealed in states with them. Fifty dollars and free continuing medical education credit will help elicit participation. Panels of experts will guide the construction of the vignettes and the intervention. The decision analytic framework is based on Hammond's lens model. This study will enhance our understanding of the effect of MCPR's, physician specialty, geographic region, medical diagnoses, alcohol disorders, and patients' gender and race on physicians' decisions to prescribe opioids and benzodiazepines, and it will assess the effectiveness of a readily transferable intervention for improving these decisions.
