Financial incentives and antibiotic prescribing patterns: Evidence from dispensing physicians in a public healthcare system

To ensure sufficient access to healthcare in remote areas, some countries allow physicians to directly dispense prescribed drugs through on-site pharmacies. Depending on the medication prescribed, this may pose a significant financial incentive for physicians to over-prescribe. This study, therefore, explored the effect of on-site pharmacies on antibiotic dispensing in a social health insurance system. Investigating physicians’ prescribing behavior is especially relevant in the case of antibiotics, as over-utilization expedites antimicrobial resistance, leading to the development of untreatable bacterial infections. The empirical analysis was based on comprehensive administrative data on 13,741 antibiotic prescriptions issued by all 4044 public general practitioners (GPs) in Austria between 2016 and 2019. Switches from dispensing to non-dispensing status (and vice versa) were exploited in a difference-in-difference framework to mitigate a potential selection bias. GPs with the right to dispense over the entire observed period were used as the control group, and those who had either lost or gained the right to dispense as the treatment group. The results from a log-linear mixed model show that not currently operating an on-site pharmacy is associated with a 9.2% lower dispensing rate (i.e., antibiotics per 1000 yearly consultations). The results are robust to potential differences between GPs who switch from dispensing to non-dispensing and those who switch from non-dispensing to dispensing, to potential patient sorting, and to different functional forms. A prescribing effect interpretation (i.e., financial incentives give rise to more prescriptions for antibiotics) explains the observed volume effect provided that the share of unfilled antibiotic prescriptions issued by non-dispensing physicians does not exceed 4%.