Segmentation of High-Need, High-Cost Medicare Subpopulations using ICD 10 Data

The two primary aims of this project are to: 1) update the segmentation algorithms using ICD-10 codes that builds upon the prior segmentation framework that was used in the NAM report as well as the high-need, high-cost (HNHC) subsegments that came out of Dr. Jose Figueroa’s current grant working with ACOs on subsegments, and 2) apply and test new algorithms in the CMS Virtual Research Data Center (VRDC) by the Institute for Accountable Care Team (IAC). Once Dr. Figueroa and his team create the new ICD-10 algorithms for the segmentation and subsegment cohorts, they will share the code with the IAC led by Jennifer Perloff. The IAC team will test the adaptation of population segmentation algorithms for use with ICD-10 Medicare claims. More specifically, IAC will import the SAS code into the VRDC and run it with ICD-10 claims data for 2018 in five health care markets, comparing results for accountable care organizations in those markets with the market-area averages. IAC staff will collaborate with the Harvard team on devising validation tests and sharing the results with experts. The ultimate product will be an updated algorithm of new patient segments using ICD-10 codes that is publicly and widely available.