Physician characteristics.

Objective
To evaluate the impact of personalized prescribing portraits on antibiotic prescribing for treating uncomplicated acute cystitis (UAC) by Family Physicians (FPs).
Design
Cluster randomized control trial.
Setting
The intervention was conducted in the primary care setting in the province of BC between December 2010 and February 2012.
Participants
We randomized 4 833 FPs by geographic location into an Early intervention arm (n = 2 417) and a Delayed control arm (n = 2 416).
Intervention
The Education for Quality Improvement in Patient Care (EQIP) program mailed to each FP in BC, a ‘portrait’ of their individual prescribing of antibiotics to women with UAC, plus therapeutic recommendations and a chart of trends in antibiotic resistance.
Main outcome measures
Antibiotic prescribing preference to treat UAC.
Results
Implementing exclusion criteria before and after a data system change in the Ministry of Health caused the arms to be unequal in size–intervention arm (1 026 FPs, 17 637 UAC cases); control arm (1 352 FPs, 25 566 UAC cases)–but they were well balanced by age, sex and prior rates of prescribing antibiotics for UAC. In the early intervention group probability of prescribing nitrofurantoin increased from 28% in 2010 to 38% in 2011, a difference of 9.9% (95% confidence interval [CI], 9.1% to 10.7. Ciprofloxacin decreased by 6.2% (95% CI: 5.6% to 6.9%) and TMP-SMX by 3.7% (95% CI: 3.1% to 4.2%). Among 295 FPs who completed reflective surveys, 52% said they were surprized by the E. coli resistance statistics and 57% said they planned to change their treatment of UAC.
Conclusion
The EQIP intervention demonstrated that feedback of personal data to FPs on their prescribing, plus population data on antibiotic resistance, with a simple therapeutic recommendation, can significantly improve prescribing of antibiotics.
Trial registration:ISRCTN 16938907.