Performance of Chronic Conditions Data Warehouse Algorithms in Medicare Advantage and Fee-For-Service Populations.

BACKGROUND

The Centers for Medicare & Medicaid Services (CMS) Chronic Conditions Data Warehouse (CCW) algorithms are widely used to identify chronic conditions in Medicare fee-for-service (FFS) claims. However, their validity in Medicare Advantage (MA) encounters remains uncertain.

METHODS

We conducted a cross-sectional study using the National Health and Aging Trends Study linked to Medicare data. We analyzed NHATS Rounds 9 and 11, aged≥65 years, with continuous 24-month enrollment in MA or FFS. CCW algorithms were applied to identify six conditions: arthritis, chronic lung disease, diabetes, heart disease, hypertension, and osteoporosis. Self-reported, physician-diagnosed conditions from NHATS interviews served as the reference standard. Survey-weighted sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and balanced accuracy were compared between MA and FFS.

RESULTS

We included 1,708 MA and 2,239 FFS beneficiaries in Round 9 and 1,472 MA and 1,566 FFS beneficiaries in Round 11. Across both rounds, balanced accuracy was similar between MA and FFS for all six conditions. Diabetes and hypertension demonstrated high sensitivity (0.88-0.94), while arthritis and osteoporosis showed lower sensitivity but high specificity. Although statistically significant MA-FFS differences were observed for select metrics (diabetes PPV in Round 9; chronic lung disease specificity in Round 11), absolute differences were minor and not clinically meaningful.

CONCLUSIONS

CCW algorithms perform comparably in MA and FFS Medicare data. These findings support the validity of using CCW algorithms in MA encounter data for epidemiologic research for risk adjustment, facilitating more generalizable studies of the growing MA population.