Min Hee Kim, Lisandro D Colantonio, M Maria Glymour, Rita Hamad, Julia Adler-Milstein, Kiarri N Kershaw, John Boscardin, Zhixin Wang, Ayse Sert Oti, Virginia J Howard
The sample included 45,191 observations (23% Black) of 8368 participants. Among those with incident cognitive impairment, 30.8% received a claim-based diagnosis, and while 8.9% of those without incident cognitive impairment received a claim-based diagnosis. ADRD diagnosis rates remained stable, whereas rates among observations with incident cognitive impairment rose through the early 2010s, then plateaued and declined. The black race and education below the college level were associated with lower odds of diagnosis in this group; female sex was associated with higher odds. Among observations without cognitive impairment, older age and female sex were associated with higher odds; Midwest or West/Northeast residence and less than high school education were associated with lower odds.
INTRODUCTION: Understanding the concordance between incident cognitive impairment status and the diagnosis of Alzheimer disease and related dementias (ADRD) is clinically important, as a diagnosis serves as the gateway for treatment, care planning, and supportive services.
METHODS: Using the REGARDS cohort (2006 to 2021) linked to Medicare claims, we examined the discordance between incident cognitive impairment and claims-based ADRD diagnoses. Directly standardized diagnosis rates were estimated by year, overall, and race, and stratified by cognitive impairment status. Logistic regression models were used to examine sociodemographic correlates of diagnosis.
RESULTS: The sample included 45,191 observations (23% Black) of 8368 participants. Among those with incident cognitive impairment, 30.8% received a claim-based diagnosis, and while 8.9% of those without incident cognitive impairment received a claim-based diagnosis. ADRD diagnosis rates remained stable, whereas rates among observations with incident cognitive impairment rose through the early 2010s, then plateaued and declined. The black race and education below the college level were associated with lower odds of diagnosis in this group; female sex was associated with higher odds. Among observations without cognitive impairment, older age and female sex were associated with higher odds; Midwest or West/Northeast residence and less than high school education were associated with lower odds.
DISCUSSION: Differences observed in the study may potentially represent underlying disparities in access to care and dementia diagnosis in clinical practice.