Medicare Advantage Star Ratings

CMS uses a hierarchical clustering algorithm that analyzes the distribution of all contracts’ scores on a measure and identifies natural groupings. The algorithm creates four cut points that divide contracts into five clusters, each corresponding to a star level (1 through 5). The advantage: cut points reflect actual performance distributions rather than arbitrary thresholds. The risk: small shifts in the distribution can cause large rating changes.

Survey-based measures use a different method: relative distribution with significance testing. CMS calculates national averages, then tests whether each contract’s score is statistically significantly above or below the average to assign ratings. This method is more stable but less sensitive to genuine improvement.