Prediction of the Risk of Bleeding in People Living with Hemophilia

Abstract

A tool allowing the prediction of the risk of bleeding in patients with hemophilia would be relevant for patients, stakeholders, and policymakers. We performed a systematic review of the literature searching for available risk assessment models to predict the risk of bleeding in people living with hemophilia, and to determine the key risk factors that the ideal model should include. We also systematically review the literature to determine the acceptability and accuracy of wrist-wearable devices to measure physical activity in the general population. Finally, we validated the performance of a risk assessment model for the prediction of the risk for bleeding in people living with hemophilia. We identified the following risk factors for bleeding in people living with hemophilia: plasma factor levels, history of bleeds, physical activity, antithrombotic treatment, and obesity. The FitBit Charge and FitBit Charge HR are the most accurate devices for measuring steps, and the Apple Watch is the most accurate for measuring heart rate. No device proved to be accurate in measuring energy expenditure. The predictive accuracy of the risk assessment model that we validated does not endorse its use to drive decision making on treatment strategies based on the predicted number of bleeds. This might in part be explained by the methods used in the derivation phase. The need for an accurate risk assessment model to predict the risk of bleeding in people living with hemophilia is still unmet. This should be done by including the relevant risk factors identified through our work, with data on physical activity possibly collected using an accurate wrist-wearable device, and through the application of rigorous methods in the derivation and validation phases.