Risk prediction models in perioperative medicine: methodological considerations
Collins GS., Ma J., Gerry S., Ohuma E., Odondi L., Trivella M., De Beyer JA., Vazquez-Montes MDLA.
Identifying individuals who are at increased risk of mortality or major morbidity following surgical procedures is an important challenge in patient management. Multivariable risk prediction models (e.g., EuroSCORE, POSSUM, and the STS Risk Calculator) have been developed to help surgeons calculate a patient’s risk of mortality using a combination of risk factors. These prediction models have transformed preoperative risk assessment. Studies evaluating the performance of risk prediction models in this and other areas of medicine have, however, been characterized by poor design, methodological conduct, and reporting. We discuss the main methodological considerations behind risk prediction models and critically discuss issues in their design, validation, and transparency.