Background
Global developing antibiotic resistance necessitates research to identify and limit driving factors for imprudent antibiotic prescribing. However, in Australia this research has to date been limited by the paucity of access to large-scale, individual-level, patient-prescriber data from primary care.
Aim
This project will utilise administrative data from Australian primary care to identify factors predictive of antibiotic prescribing not conforming with the national therapeutic guidelines, with a focus on upper respiratory and urinary tract infections.
Methods
De-identified patient-prescriber data from 2012-2016 inclusive was obtained from GP software by NPS MedicineWise using the MedicineInsight program in 52 de-identified, consenting practices in Western Australia (WA). This project involves a strategic partnership between the WA Primary Health Alliance and Curtin University. Guideline adherence is assessed using algorithms developed based on the national therapeutic guidelines for antibiotic prescribing, and incorporating indications, clinical observationsand pathology results. Levels of adherence assessed include whether a prescription was likely indicated, and choice of agent. Descriptive statistics and generalized linear mixed methods are used to identify patient- and practice-related factors predictive of non-conforming prescribing.
Results
Descriptive results will be presented highlighting the usefulness of administrative data for research, and its challenges and limitations. The strategic partnership approach facilitates vital links to GPs for support in understanding the strengths and weaknesses of administrative data, and its implications for interpretation. This partnership will be key in the translation of findings to practice and supporting national antimicrobial stewardship efforts.
Conclusion
Routine administrative data can support research in Australian primary care.