Prediction models - development, evaluation and clinical application

New England Journal of Medicine

23 April 2020 - When national lipid guidelines first incorporated a model based on data from the Framingham Heart Study — a turning point for the role of risk prediction in health care — that Massachusetts city was an anomaly: a community with extensive, available, longitudinal health data. 

Today, U.S. health care systems have amassed large, local data sets through adoption of electronic health records and the standardisation associated with provider consolidation. 

More recently, payers have moved toward capitation and other value-based models. This shift places a higher premium on avoiding costly conditions altogether. These trends create greater demand for prediction models, since prevention is difficult without accurate identification of who specifically is at risk.

Read New England Journal of Medicine Perspective

Michael Wonder

Posted by:

Michael Wonder