Cut-off Projection

• CitiusTech

CMS assigns stars at a measure level, that range between 1 to 5, for each numeric measure score. Better the measure performance, higher will be the assigned rating i.e., for most of the measures higher the compliance, better the rating. Whereas, for inverse measures (e.g., Complaints about health plan) lower the compliance better the rating. The following methods are used to decide the STAR cut-offs for each measure based on the methodology described by CMS: Clustering: This method is applied to operational measures, process-based measures, HEDIS measures and other clinical care measures. Conceptually, the clustering algorithm identifies the “gaps” among the scores and creates four cut points resulting in the creation of five levels (one for each Star Rating). Relative distribution and significance testing for CAHPS measures: This method is applied to determine valid star cut points for CAHPS measures. It evaluates the relative percentile distribution and considers significance testing. Fixed cut points: The ‘Beneficiary Access and Performance Problems’ measure is assigned Star Ratings using fixed cut points. This methodology causes the final contract scores to be either zero or a multiple of 20 (20, 40, 60, 80 or 100). MAE is used for Model Validation. Once we have the Projection of next year’s un-scaled measure-level compliance score and Projection of next year’s cut points, we can assign STAR ratings to all measures based on their performance scores and cut-points. A weighted average of these measure level star ratings, gives us the final scaled STAR rating.

Input variables : Projected compliance rates - to create cut-off estimations, measure-level predictions made for contracts for the prediction year
Output Variables : cut offs - cut point 1, cut point 2, cut point 3, and cut point 4

Visit Model : medictiv.citiustech.com