Performance associated effect variations of public reporting in promoting antibiotic prescribing practice: a cluster randomized-controlled trial in primary healthcare settings

Aim: To evaluate the variations in effect of public reporting in antibiotic prescribing practice among physicians with different performance in primary healthcare settings. Background: Overprovision of antibiotics is a major public health concern. Public reporting has been adopted to encourage good antibiotic prescribing practices. However, which group, for instance, high, average or low antibiotic prescribers, accounted for antibiotic prescription reduction has not been fully understood. Methods: A cluster randomized-controlled trial was conducted. In total, 20 primary healthcare institutions in Qianjiang city were paired through a six indicators-synthesized score. Coin flipping was used to assign control-intervention status; 10 were then subjected to intervention where prescription indicators were publicly reported monthly over a one-year period. Prescriptions for upper respiratory tract infections (URTIs) before and after the intervention were collected. Physicians were divided into high, average and low antibiotic prescribers based on their antibiotic prescribing rates last month, which were publicly reported in intervention arm. Multilevel difference-in-differences logit regressions were performed to estimate intervention effect in each physician group on three outcome indicators: prescriptions containing antibiotics, two or more antibiotics and antibiotic injections. Findings: In total, 31 460 URTI prescriptions were collected (16 170 in intervention arm and 15 290 in control arm). Reduction in antibiotic prescription attributed to intervention was 2.82% [95% confidence intervals (CI): -4.09, -1.54%, P< 0.001], least significant in low prescribers (-1.41%, 95% CI: -3.81, 0.99%, P= 0.249) and most significant in average prescribers (-5.01%, 95% CI: -6.94, -3.07%, P< 0.001). Reduction in combined antibiotics prescriptions attributed to intervention was 3.81% (95% CI: -5.23, -2.39%, P< 0.001), least significant in low prescribers (-2.42%, 95% CI: -4.39, -0.45%, P= 0.016) and most significant in average prescribers (-5.01%, 95% CI: -7.47, -2.56%, P< 0.001). Conclusion: Public reporting can positively influence antibiotic prescribing patterns of physicians for URTIs in primary care settings, with reduction in antibiotic and combined antibiotic prescriptions. The reduction was mainly attributed to average and high antibiotic prescribers.