Injections are commonly overused in Indonesia. More than 60% of patients attending public health facilities receive at least one injection, which increases clinical risk and has adverse economic impact. This study assesses the efficacy of an innovative behavioral intervention, the Interactional Group Discussion (IGD), for reducing the overuse of injections. This study was a controlled trial in a single district with 24 public health centers randomized to intervention and control groups. Prescribers in the intervention group were invited to one IGD, each of which consisted of 6 prescribers and 6 patients; a total of 24 IGDs were held in a 4-week period, and all invited prescribers participated. The groups, which lasted 90-120 minutes, were facilitated by a behavioral scientist and a clinician, who also served as a scientific resource person. The hypothesized mechanism of behavior change involved reality testing prescribers' assumptions about patient beliefs, imparting scientific information about injection efficacy, and establishing peer norms about correct behavior. Outcomes were measured by a retrospective prescribing survey covering the periods 3 months before and 3 months after the intervention, with samples of 100 prescriptions per center per month. Rates of injection and average number of drugs per prescription were computed separately for each center, and t-tests were used to compare pre-post changes in outcomes in both groups. Results showed a significant decrease in injection use from 69.5 to 42.3% in the intervention group, compared to a decrease from 75.6 to 67.1% among controls [-18.7.0% intervention vs control, 95% CI = (-31.1%, -6.4%), P < 0.025]. There was also a significant reduction in average number of drugs per prescription [-0.37 drugs prescribed per patient, 95% CI = (-0.04, -0.52), P < 0.05], indicating that injections were not substituted with other drugs. We conclude that the IGD significantly reduces the overuse of injections. It is suggested to try out other behavioral interventions to improve the rational use of drugs.