Consistency of a Dementia Diagnosis Between Systems for Patients Receiving Care Through the VA, Traditional Medicare, and Medicare Advantage

BACKGROUND: Older Veterans receive care through the Department of Veterans Affairs (VA) and Medicare, including Medicare Advantage (MA). We assessed the extent to which a dementia diagnosis recorded in one system is reflected in the other and factors associated with discordance among dual users. DESIGN: Parallel cohort study of patients with dementia identified in the VA or Medicare, both overall and separated into traditional Medicare (TM) and MA. PARTICIPANTS: Patients aged 65 years or older enrolled in VA and Medicare with a 2018 face-to-face dementia diagnosis encounter in the VA (N=68,092; VA dementia cohort) or Medicare (N=76,622; Medicare dementia cohort). MEASURES: In both cohorts, our primary outcome was whether the dementia diagnosis was absent at each professional encounter in the other system during 1-year follow-up. Characteristics considered included patient sex, age, race, ethnicity, rurality, priority group, dementia type, and clinician specialty during follow-up. RESULTS: For VA and Medicare cohorts, dementia diagnosis was absent in 88.7% and 80.5%, respectively, of follow-up encounters in the other system. For the VA cohort, the diagnosis was absent in 89.2% and 71.8% of TM and MA encounters, respectively. Encounters with a psychiatrist or neurologist (vs. primary care physician) were linked to lower absence during follow-up (eg, 25.3 and 20.3 percentage points lower for the VA cohort during Medicare follow-up; P<0.001 for both); those in TM (vs. MA) were 14.5 percentage points more likely to have an absent diagnosis. CONCLUSIONS: Absent dementia diagnoses across VA and Medicare among dual users was common, associated with clinician specialty and Medicare type.