Patient Risk Models: Out With the Old, In With the New

Move out of the way traditional (linear) methods for identifying high-risk patients, there is a new sheriff in town: Machine Learning.

Photo Source: Milliman MedInsight


Move out of the way traditional (linear) methods for identifying high-risk patients, there is a new sheriff in town: Machine Learning.


With the entire healthcare landscape shifting towards population-based health outcomes, healthcare organizations serving Medicare, Commercial and Medicaid patients are jumping on the ‘value-based bandwagon’, which will require more sophisticated machine learning risk identification models to improve patient outcomes and reduce costs for both payers and providers alike.


Machine learning algorithms allow as many input factors as possible without needing to explicitly define relationships, which is essential in identifying patient risk amongst the sheer variety, volume and complexity of healthcare data.