# Routine data registries as a basis to analyse and improve the quality of antimicrobial prescription in primary care

## TL;DR

This study uses healthcare and socioeconomic data to analyze and improve the quality of antimicrobial prescriptions in primary care, especially for respiratory infections.

## Contribution

Combines routine healthcare and socioeconomic data to identify novel determinants of inappropriate antimicrobial prescribing.

## Key findings

- Respiratory tract infections had 14.5% inappropriate antibiotic prescriptions.
- 77.1% of macrolide prescriptions did not align with guideline recommendations.
- Patient characteristics and practice size were linked to poorer guideline adherence.

## Abstract

The overuse of antimicrobials is the main driver of the increasing antimicrobial resistance (AMR). Between 80 and 90% of antimicrobial prescriptions originate from primary care. The goals were to establish the quality of prescriptions through combining data from a primary healthcare registry and a national socioeconomic database for novel determinants.

Pseudonymized routine healthcare data from 269,547 patients (1,150,252 antimicrobial prescriptions) obtained between 2012 and 2020 from primary care practices (n = 101) in the region The Hague – Leiden were used. These data were linked with individual socioeconomic data from the Statistic Netherlands database to identify determinants of inappropriate antimicrobial prescribing. The quality of prescription was analysed using predefined criteria based on primary care guidelines. Multivariable logistic regression analyses were performed to identify associations with appropriateness.

Respiratory tract infections (RTI) were most commonly associated with inappropriate antibiotic prescribing, with 14.5% RTI prescriptions being inappropriate. For macrolide prescriptions, 77.1% did not correspond with 1st and 2nd guideline choices. Patient characteristics (migration backgrounds, female gender, comorbidities and age) and medium and large primary care practice size, a proxy for continuity of care and consultation time per patient, were associated with poorer guideline adherence.

Combined analyses of socioeconomic and routinely collected healthcare data does reveal relevant additional information to answer medical questions in a broader context, such as AMR. Most room for improvement was found for RTIs and macrolides, especially in specific risk groups. Assuring continuity of care and/or providing extended consultation time per patient might be essential elements to establish, before disseminating improvement strategies.

The online version contains supplementary material available at 10.1186/s12875-025-03008-4.