Abstract
OBJECTIVE: To examine racial/ethnic differences in risk factors, and their associations with COVID-19-related outcomes among older adults with Alzheimer's disease and related dementias (ADRD). DESIGN: Observational study. SETTING AND PARTICIPANTS: National Medicare claims data and the Minimum Data Set 3.0 from April 1, 2020, to December 31, 2020, were linked in this study. We included community-dwelling fee-for-service Medicare beneficiaries with ADRD, diagnosed with COVID-19 between April 1, 2020, and December 1, 2020 (N = 138,533). METHODS: Two outcome variables were defined: hospitalization within 14 days and death within 30 days of COVID-19 diagnosis. We obtained information on individual sociodemographic characteristics, chronic conditions, and prior health care utilization based on the Medicare claims and the Minimum Dataset. Machine learning methods, including lasso regression and discriminative pattern mining, were used to identify risk factors in racial/ethnic subgroups (ie, White, Black, and Hispanic individuals). The associations between identified risk factors and outcomes were evaluated using logistic regression and compared across racial/ethnic subgroups using the coefficient comparison approach. RESULTS: We found higher risks of COVID-19-related outcomes among Black and Hispanic individuals. The areas under the curve of the models with identified risk factors were 0.65 to 0.68 for mortality and 0.61 to 0.62 for hospitalization across racial/ethnic subgroups. Although some identified risk factors (eg, age, gender) for COVID-19-related outcomes were common among all racial/ethnic subgroups, other risk factors (eg, hypertension, obesity) varied by racial/ethnic subgroups. Furthermore, the associations between some common risk factors and COVID-19-related outcomes also varied by race/ethnicity. Being male was related to 138.2% (95% CI: 1.996-2.841), 64.7% (95% CI: 1.546-1.755), and 37.1% (95% CI: 1.192-1.578) increased odds of death among Hispanic, White, and Black individuals, respectively. In addition, the racial/ethnic disparity in COVID-19-related outcomes could not be completely explained by the identified risk factors. CONCLUSIONS AND IMPLICATIONS: Racial/ethnic differences were detected in the likelihood of having COVID-19-related outcomes, specific risk factors, and relationships between specific risk factors and COVID-19-related outcomes. Future research is needed to elucidate the reasons for these differences.