RESUMEN
Overprescribing of antibiotics and injectable medicines is common in ambulatory care in many low- and middleincome countries. We evaluated the effects of three different interventions in improving physician prescribing. We conducted a four-armed randomized controlled trial with one-month and three- months follow-up. General physicians, pediatricians, and infectious disease specialists were included in this study if they had an outpatient office in Tehran, Iran. The study involved two behaviorally guided interventions: "new-design audit and feedback (NA&F)"; "printed educational material (PEM)" and an existing intervention of "routinely conducted audit and feedback (RA&F)". The theoretical framework underpinning the intervention was the theory of planned behavior. Main outcome measures were the percentage change in the proportion of prescriptions containing injectable dexamethasone; oral amoxicillin and cefixime. NA&F reduced the proportion of prescriptions particularly those containing dexamethasone injectable and cefixime (1.64, 0.99 absolute percentage change, p = 0.006, p = 0.01 respectively). PEM reduced the proportion of prescriptions containing cefixime (0.93 absolute percentage change p = 0.04). Other primary outcomes had no significant differences. A secondary outcome measure showed overall prescribing of injectables also reduced (absolute risk reduction: 3%). Overally, the study provides strong evidence that using theoretical insights in the development of the intervention improved prescribing behavior that lasted at least three months after the intervention. The design, format, and presentation of messages in feedback forms significantly influence the impact of audit and feedback on physician prescribing. While the interventions were effective, the impacts on inappropriate prescribing were modest and limited. In settings with rampant problems of overprescribing, intensive interventions are required to substantially improve prescribing patterns. Graphical abstract Graphical abstract.
