519Innovative use of MedicineInsight GP database for applied research on antibiotic prescribing in Western Australia

Abstract

Abstract Focus of Presentation This project showcases the use of existing general practice (GP) datasets to gain access to large-scale data for epidemiological research. There is limited quantitative research on antibiotic prescribing in Australian GP. The research question involves identifying predictors of antibiotic prescribing not in accordance with national therapeutic guidelines for urinary tract infection (UTI) in Western Australian (WA) GP. This requires large-scale, individual-level data, which is difficult to access in Australia, and in this case could not be achieved without collaboration. Findings Longitudinal records from 52 practices in WA participating in MedicineInsight were analyzed. GP datasets present challenges regarding cleaning, processing and storage. Substantial broad-spectrum antibiotic prescribing was identified in initial episodes of UTI, particularly for children, and not supported by laboratory pathology. Children received broad-spectrum antibiotics 85% of the time, compared to 58% in women. Mixed-effects logistic regression was used to elucidate patient- and consultation-related factors associated with antibiotic prescribing outside of the guidelines. Conclusions/Implications There is evidence of antibiotic prescribing outside of the guidelines for initial presentations of UTI in WA GP. Whilst GP datasets provide insightful information on antibiotic prescribing, it is important to explore these findings with practising clinicians to further understand the drivers behind this behaviour, and this is the next stage of this research. Key messages Collaborative partnerships can facilitate access to rich, large-scale datasets capable of answering important research questions in Australian primary care.