Routine Prescription Feedback and Peer Comparison to Lower Antibiotic Prescriptions in Primary Care

Background: Excessive use of antibiotics may lead to unnecessary adverse events and raise the emergence of bacterial resistance, an increasingly serious problem in Europe. In absolute terms most antibiotics are prescribed in primary care with considerable unexplained variation of antibiotic use indicating the need for further important improvement of prescription practice in Switzerland.

Aim: To evaluate the effect of a continuous postal and web-based feedback and peer comparison system of individual antibiotic prescription rates on the prescription behaviour of primary care physicians in Switzerland.

Design: Randomized, controlled, pragmatic intervention trial. Setting: Primary care providers of Switzerland above the median of antibiotic prescription rates.

Population: Primary care physicians caring for patients enlisted with social health insurance companies that provide invoice data to the SANTÉSUISSE DATENPOOL AND TARIFPOOL.

Endpoints: Primary endpoint: Prescription rate of antibiotics as defined daily doses (DDD) per 100 consultations at 12 and 24 months. Secondary endpoints: Costs-savings from the intervention, acceptability of the program, percentage of prescriptions fulfilling disease-specific quality indicators for outpatient antibiotic criteria by the European Surveillance of Antimicrobial Consumption (ESAC).

Intervention: Quarterly postal and electronic feedback for 24 months on the crude number of antibiotics prescribed, a population adjusted benchmark-profile in comparison to other primary care physicians, and evidence-based guidelines for the use of antibiotics in primary care. Physicians in the control group receive no information.

Variables and measurement: Rates of antibiotic prescriptions overall and per drug class (DDD), cost of prescribed drugs and the intervention program; number of web-application logins and participation cancellations, and - in a sub-sample - number of prescriptions fulfilling disease-specific quality indicators for outpatient antibiotic criteria (ESAC).

Expected results: We expect a 5% reduction of antibiotic prescription rates between the intervention and control groups after 12 months with an assumed participation rate (at least one web-access) and return of acceptance questionnaire of 30%. We expect that the intervention program will be cost-saving. ESAC quality indicators will be a useful tool for monitoring the quality of antibiotic prescription in ambulatory care in Switzerland.

Analyses: Analysis will be done by intention to treat principles. We will use linear regression analysis to determine the difference in antibiotic prescriptions between the intervention and the control group with appropriate adjustment for the case mix of patient populations and self-dispensation. Based on outpatient data of 2009 from the Helsana insurance and conservative effect estimates a sample-size of 1427 physicians each for the intervention and control group is planned.

Significance: This trial will investigate if a repeated feedback system results in a long-term reduction of antibiotic prescription practices. In addition, the feasibility of a web-based interface as communication tool to primary care physicians will be assessed. If effective, the system could be easily employed for other interventions as well.

Inclusion criteria

• Board certified physicians with a certificate from the Swiss Medical Association (Facharzttitel' General Internal Medicine, Subspecialities in Internal Medicine, Paediatrics) with an own ZAHLSTELLENREGISTER NUMBER (Konkordatsnummer).
• At least 100 patients enlisted with the specified social health insurance providers (SANTÉSUISSE DATENPOOL AND TARIFPOOL). (to avoid classifying physicians as high prescribers who only see a small number of patients and describe antibiotics to many of them)
• Prescription-rates of antibiotics (DDD per 100 consultations) - that are prescribed to patients enlisted with the specified social health insurance providers (SANTÉSUISSE TARIFPOOL)
• are in the upper median of the prescription number distribution in the year prior to the start of the trial.

Exclusion criteria

• Physicians with no own "Zentralstellenregister" (ZSR) number (practice fellows or substitutes, medical trainees in general practices).
• Physicians working in institutions without own but with institutional ZSR NUMBER (ambulatory care facilities of University Hospitals, permanence, practices or Health Maintenance Organisation (HMO) with group ZSR number).