Changes in Medicare's Chronic Condition Files and Disparities in Dementia Identification

In 2017, the Chronic Condition Warehouse released a 30-condition Chronic Condition file (CC30), which fully replaced the prior 27-condition file (CC27) in 2022. CC30 shortened the look-back period for dementia identification from 3 to 2 years and raised the required outpatient/carrier claims from 1 to 2. This change may disproportionately affect individuals with limited access to health care. This study aimed to quantify how the CC27-CC30 transition affects dementia identification across population subgroups.

Observational study.

Medicare beneficiaries with dementia, including 12,951,241 person-years between 2018 and 2021.

We linked 2018-2021 Medicare Beneficiary Summary Files (base file, CC27, and CC30) with publicly available data. Beneficiaries with Alzheimer's disease (AD) or Alzheimer's disease and related dementias (ADRD) in either CC27 or CC30 were identified. Disagreement was defined as having dementia in one file but not the other. Logistic regression was used to examine patient-level (eg, race/ethnicity, dual status, age, gender) and area-level (eg, mental health provider shortages, social deprivation) factors associated with the disagreement.

CC30 identified approximately 70% to 72% of AD and ADRD cases in CC27, and added very few additional ADRD cases (0.04%). Disagreement (mainly driven by the exclusion of individuals previously identified by CC27) varied by individual and community characteristics. For example, Asian and Hispanic beneficiaries had 51% to 55% and 20% to 26% higher odds of disagreement for ADRD, respectively (P < .01). Rural residence was associated with an 11% to 13% increase in the odds of disagreement of ADRD (P < .01). Residents in communities with higher socioeconomic deprivation and lower proportions of White residents also had a higher likelihood of disagreement in ADRD dementia identification.

Transitioning from CC27 to CC30 reduces identified AD and ADRD dementia cases, disproportionately affecting certain subpopulations. Researchers and policymakers should consider these changes when interpreting trends and designing dementia care interventions.