Work

Can routine billing data enable audit and feedback for antibiotic stewardship? A primary care case study from Switzerland

Year: 2026
Type: conference-abstract
Abstract: Context Audit and feedback has been shown to reduce antibiotic prescribing in primary care but depends on individually attributable, benchmarkable prescribing data. In Switzerland, as in many countries without centralised prescribing registries, it remains unclear whether routine billing data can fill this gap. Objective To evaluate use of insurance billing data for generating physician-level stewardship metrics and peer comparisons in primary care. Study Design and Analysis Methodological case study embedded within a primary care cluster randomised trial (clusters: quality circles, group practices, and walk-in clinics) conducted on antibiotic prescribing. Billing data were linked, via unique billing identifiers, to physicians who consented to share their data. Prescribing variation was characterised using descriptive statistics. Dataset Billing data from primary care physicians in French-speaking Switzerland, from a Swiss health insurance claims database, 2023 to 2024. Population Studied 108 physicians across 17 clusters; 53 held individual billing numbers permitting physician-level analysis. Instrument Billing data as an instrument for individual prescribing metrics. Outcome Measures Antibiotic prescribing measured as annual and monthly defined daily doses (DDD) per 1000 consultations and absolute DDD. Patient characteristics for adjustment; diagnostic testing rates. Results 53/108 physicians (49%) had individual billing numbers; the remainder used shared group numbers. Data were available for 39 physicians in 2023 and 41 in 2024. Mean annual antibiotic prescribing was 677 DDD (SD 335) per 1000 consultations in 2023 and 750 (SD 492) in 2024. Monthly DDD per 1000 consultations showed a marked December spike, coinciding with a drop in consultation counts. Diagnostic testing rates correlated positively with prescribing. Billing data supported adjustment for selected patient characteristics (patient sex and age mix) but lacked clinical information for indication-specific quality indicators and treatment duration assessment. Conclusions Shared billing numbers constitute a structural barrier to individualised feedback. Between-physician variation was substantial. Annual rather than monthly indicators are recommended to avoid denominator artefacts. Billing data can capture diagnostic testing rates alongside prescribing, broadening feedback beyond antibiotic volume. These findings show the potential and pitfalls of billing-based antibiotic feedback.
Language: English
Field: Medicine
Open Access status: gold
OpenAlex ID: W7225909601
Cites:
Cited by:
Related to: 10