From claims to care: machine learning algorithm to classify urinary tract infection cases using nationwide health insurance data

To evaluate whether machine learning (ML) applied to nationwide health insurance claims data without diagnostic codes can distinguish urinary tract infection (UTI)–related antibiotic prescriptions from non-UTI episodes. Such approaches could strengthen antimicrobial stewardship when diagnosis-linked datasets are unavailable. Outpatient antibiotic prescription claims from three major Swiss insurers (2017–2020; covering ∼40% of the Swiss population) were analysed. Based on clinical input, specific constellations of claims codes (e.g., positive urine culture plus typical antibiotic) were a priori assigned as indicating UTI episodes, providing the reference classification. Predictors included sex, age group, comorbidity, and diagnostic tests ordered during the episode. Four ML classifiers were tested; performance and interpretability were evaluated, with Extreme Gradient Boosting prioritized. After data cleaning and balancing, 38 982 records (19 491 UTIs; 19 491 non-UTIs) were included. Extreme Gradient Boosting achieved an area under the receiver operating characteristic curve of 0.94, accuracy of 87.6%, sensitivity of 79.2%, and specificity of 96.1%. Misclassification was asymmetric: 11% of non-UTI cases were labelled UTI, while 2% of UTI cases were misclassified as non-UTI. Diagnostics ordered were the strongest predictors, followed by female sex and older age. Even in the absence of diagnosis codes, ML applied to claims data can reliably identify UTI-related prescriptions. This supports the feasibility of claims-based surveillance tools for stewardship while highlighting the need for scalable, low-burden approaches to improve direct diagnostic coding in routine data.