Patterns of care for clinically distinct segments of high cost Medicare beneficiaries.
BACKGROUND: Efforts to improve the efficiency of care for the Medicare population commonly target high cost beneficiaries. We describe and evaluate a novel management approach, population segmentation, for identifying and managing high cost beneficiaries. METHODS: A retrospective cross-sectional analysis of 6,919,439 Medicare fee-for-service beneficiaries in 2012. We defined and characterized eight distinct clinical population segments, and assessed heterogeneity in managing practitioners. RESULTS: The eight segments comprised 9.8% of the population and 47.6% of annual Medicare payments. The eight segments included 61% and 69% of the population in the top decile and top 5% of annual Medicare payments. The positive-predictive values within each segment for meeting thresholds of Medicare payments ranged from 72% to 100%, 30% to 83%, and 14% to 56% for the upper quartile, upper decile, and upper 5% of Medicare payments respectively. Sensitivity and positive-predictive values were substantially improved over predictive algorithms based on historical utilization patterns and comorbidities. The mean [95% confidence interval] number of unique practitioners and practices delivering E&M services ranged from 1.82 [1.79-1.84] to 6.94 [6.91-6.98] and 1.48 [1.46-1.50] to 4.98 [4.95-5.00] respectively. The percentage of cognitive services delivered by primary care practitioners ranged from 23.8% to 67.9% across segments, with significant variability among specialty types. CONCLUSIONS: Most high cost Medicare beneficiaries can be identified based on a single clinical reason and are managed by different practitioners. IMPLICATIONS: Population segmentation holds potential to improve efficiency in the Medicare population by identifying opportunities to improve care for specific populations and managing clinicians, and forecasting and evaluating the impact of specific interventions.
Clough, JD; Riley, GF; Cohen, M; Hanley, SM; Sanghavi, D; DeWalt, DA; Rajkumar, R; Conway, PH
Volume / Issue
Start / End Page
Pubmed Central ID
Electronic International Standard Serial Number (EISSN)
Digital Object Identifier (DOI)