PROJECT SUMMARY/ABSTRACT
Alzheimer's disease and related dementias (ADRD) affect as many as 5.7 million Americans and will cost the
U.S. economy between $379 and $500 billion per year by 2040. Goal 5 of the 2020 update to the National Plan
to Address Alzheimer's, "Improve data to track progress", recognizes the need for increased surveillance of
ADRD. The goal states, "Data and surveillance efforts are paramount to tracking the burden of ADRD … and
assist with understanding health disparities among populations such as racial and ethnic minorities, low-
income populations, rural residents, and sexual and gender minorities." No data or surveillance system
currently exists for ADRD surveillance. However, Medicare and Medicaid data offer a promising source of
information to track ADRD diagnoses. Medicare covers most Americans over the age of 65, and Medicaid
covers long-term care services for people with ADRD. In 2016, the Data Infrastructure Committee of the
National Institute on Aging (NIA) identified "ways to reduce the significant obstacles faced by researchers in
using medical care claims and encounter data" as a primary NIA objective, highlighting the importance and
value of Medicare and Medicaid data. We propose to create a publicly accessible system to (1) track ADRD
diagnoses in Medicare and Medicaid; (2) conduct research to understand the correlation between diagnosed
and assessed dementia; and (3) demonstrate the value of the new system by using its data to evaluate an
important research question. In Aim 1, we propose to leverage an established set of partners and
methodologies to develop a publicly accessible system to monitor and surveil ADRD diagnoses in Medicare
and Medicaid, providing information at the national, state and county level and by age, sex, race/ethnicity,
comorbidity, eligibility and other characteristics and stratification factors. Medical claims do not identify all
persons living with dementia, so additional research is needed to understand the biases embedded in payment
claims information. In Aim 2, we will evaluate the predictive power of Medicare claims in identifying persons
with assessed dementia using linked data from Medicare and the nationally representative Health and
Retirement Study (HRS). Aim 2 will estimate the sensitivity and specificity of Medicare claims in detecting HRS
assessed dementia and explore differences across sociodemographic, health, and cognitive factors. Finally, in
Aim 3, we propose to use the new surveillance platform to perform a demonstration analysis; estimating rates
of COVID-19 mortality among patients diagnosed with ADRD in Medicare. When completed, the system we
describe will provide a much-needed ADRD surveillance system, vastly expanding access to and
understanding of Medicare and Medicaid ADRD information for the research community and general public.