Oxford/UNIC-led study develops breakthrough models to predict blood pressure medication side effects

A study led by Constantinos Koshiaris, Assistant Professor of Medical Statistics at the Medical School and James P. Sheppard, Professor of Applied Health Data Science at the University of Oxford, has developed new models to predict serious side effects from antihypertensive medication.

The study, published in Nature Communications, used electronic health records from over 5.5 million UK primary-care patients to build and validate three risk-prediction models – called the STRATIFY models – for hypotension, syncope, and fractures in people undergoing antihypertensive therapy.

The statistical models estimate a patient’s medium and long-term risk for these side effects using routine clinical information such as age, blood pressure, medical history, and current medications. The models reliably differentiated between patients likely to develop side effects and those unlikely to do so, and showed meaningful clinical utility when assessed for their influence on real-world medical decision-making.

These tools could help clinicians and patients have more informed discussions about the potential harms of antihypertensive treatment by identifying individuals at particularly high risk of adverse events, whilst reassuring the majority whose risk is low.

‘Blood pressure-lowering with antihypertensive treatment has been shown to be very effective at reducing the risk of cardiovascular events across all age groups. However, blood pressure lowering is not without harm. To enable informed decision-making, clinicians need to understand an individual’s underlying risk of adverse events, so that this can be weighed against a patient’s likelihood of benefit from new or continued treatment. Further work is needed to evaluate how best to integrate these statistical models into clinical decision-making and digital health systems’, said Dr Koshiaris.