Electronic Health Record-Based Prediction Models to Inform Decisions about HIV Pre-exposure Prophylaxis: A Systematic Review

BackgroundSeveral clinical prediction models have been developed using electronic health records data to help inform decisions about HIV pre-exposure prophylaxis (PrEP) prescribing, but the characteristics and quality of these models have not been systematically assessed. We identified and critically appraised the characteristics and quality of studies reporting the development of electronic health records (EHR)-based models predicting HIV risk to inform decisions about PrEP prescribing. MethodsWe searched PubMed and the CINAHL databases between January 1, 2013 and June 19, 2023, with keywords related to EHR, HIV, and clinical prediction. We extracted data using the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) checklist and assessed risk of bias using the Prediction model Risk Of Bias Assessment Tool (PROBAST) short form. We used narrative synthesis to describe characteristics and quality of eligible models. ResultsWe identified 324 studies, of which 7 studies (resulting in 7 models) were eligible for our review. Several studies inadequately reported key components of the corresponding model. Most models were developed in the United States and used machine learning methods. The area under the receiver operating characteristic curve was reported for six models, which ranged between 0.77 and 0.89. All models had high risk of bias, primarily because of low events per variable and risk of overfitting. ConclusionsWe observed inadequate reporting of key components and high risk of bias across all EHR-based models. Future studies would benefit from following standard reporting guidelines and best practices for developing prediction models, which may strengthen the validity and applicability of EHR-based prediction models for informing decisions about HIV PrEP prescribing. Trial registrationThe review protocol was registered and published in PROSPERO (CRD42023428057)