Back

Performance of ICD10 Code-Based Dementia Case Definition in the Health and Retirement Study

Gianattasio, K. Z.; Steffan, M.; Reist, B.; Power, M. C.; Rein, D.

2026-01-13 epidemiology
10.64898/2026.01.12.26343791 medRxiv
Show abstract

STRUCTURED ABSTRACTO_ST_ABSINTRODUCTIONC_ST_ABSThe Dementia DataHub (DDH) reports U.S. dementia prevalence and incidence from Medicare data. Variation in sensitivity, specificity, and accuracy of diagnoses related to geography and participant characteristics complicates interpretation of these data. METHODSWe evaluated performance of DDH-defined Medicare claims diagnoses against linked HRS survey-based dementia classifications. RESULTSDDHs likely-or-higher dementia definition achieved 50% sensitivity, 97% specificity, and 91% accuracy relative to the HRS classification. Sensitivity was higher in urban settings, but patterns varied across census divisions. Respondents with dementia missed in claims were younger and healthier than those correctly identified. DISCUSSIONMedicare claims reflecting diagnoses of dementia provide valuable information about who may be receiving dementia treatment and where; however, they often miss cases in ways that differ by geography and patient characteristics. Variation in diagnosis in Medicare claims relative to HRS survey-based dementia classification can be used to improve the value of Medicare diagnoses for surveillance purposes.

Matching journals

The top 9 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.