2020-2021 Pilot Awardees Have Been Selected!
We are delighted to announce the five projects selected for the Center for Advancing Sociodemographic and Economic Study of Alzheimer’s Disease (CeASES-ADRD) 2020-2021 pilot grant awards. These five projects were selected from a number of competitive applications. Please join us in congratulating our newest cohort of pilot grant recipients. Information about applying for the 2021-2022 CeASES-ADRD Pilot Award can be found here.
National Longitudinal Study, 1972 (NLS-72) Record Matching to Centers for Medicare and Medicaid Services (CMS) records.
Investigator: Chandra Muller, PhD, University of Texas, Austin
Education is among the most important determinants of later life cognitive health. Over the life course, education leads to better cognitive functioning and healthier aging through access to better jobs, health care, residential neighborhood, wealth accumulation, and economic security. But, research on early-life educational processes that affect cognitive functioning has been severely limited by a lack of prospective longitudinal data. We do not have data sufficient to identify the specific (and potentially manipulability) organizational, curricular, social, or skill aspects of education that shape cognitive functioning. The proposed project will: link information from sample members of the National Longitudinal Study of 1972 NLS-72 cohort to CMS records; use information gained from matching to provide valuable data on health for analysis of educational precursors (adolescence and early adulthood) of health and cognitive aging among individuals in their mid 60’s as they enter old age; and establish a proof of concept and results for potential re-interviewing of the NLS-72 sample members. The proof of concept will involve not only the possibility of locating respondents but also the value of the education data as a resource for understanding cognitive aging. This study would create new data for analysis of the role of education in cognitive aging among individuals currently age 65 and older. It will also make possible comparisons to the HS&B cohort that is approximately a decade younger.
The Impact of Cognitive Screening on Early ADRD Diagnosis, Treatment and Spending
Investigators: Mireille Jacobson, PhD, University of Southern California; Jason Doctor, PhD, University of Southern California
As part of a new benefit created under the Affordable Care Act (ACA), Medicare now covers an Annual Wellness Visit (AWV) that requires, among other things, an assessment to detect cognitive impairment. In newly collected data on about 1000 seniors ages 65 and over, we find systematic differences by Medicare plan type (traditional fee-for-service vs. commercial Medicare Advantage (MA)) in the likelihood that seniors ever received either an annual wellness visit or a cognitive assessment. Seniors enrolled in MA plans are about 20 percentage points (p<0.001) more likely to have had an annual wellness visit than those in traditional Medicare (traditional group mean = 47%) and about 13 percentage points more likely to have had a cognitive assessment (traditional group mean = 30%). These findings hold even after controlling flexibly for age, gender, race, education, retiree status, general health and mental health status. The specific aims of the broad project are to: (1) Assess whether, consistent with higher self-reported screening rates, MA plan enrollees are more likely to receive an early ADRD diagnosis, where early is captured by age or an MCI diagnosis; (2) Describe health care service use before, during, and after a beneficiary screens positive for cognitive impairment both overall and by plan type (traditional versus MA); (3) Assess the impact of early diagnosis on the trajectory of health care use and spending. In this pilot, we will leverage the difference in MA plan cognitive screening to study the implications for early diagnosis, treatment and spending. Combining both Medicare Fee- for-service claims and MA plan encounter files, we will study how plan type affects the likelihood of an early ADRD diagnosis, where early is defined as prior to and including age 70. As an alternative definition of early diagnosis, we will consider diagnosis of mild cognitive impairment (MCI). Conditional on a diagnosis, we will describe health care utilization and spending and the impact of early diagnosis on them. These aims will help fill an enormous gap in our understanding of the impact of early ADRD diagnosis on health care utilization and spending in both the short and long term.
ADRD Care in New Delivery and Payment Models
Investigator: Alice Chen, PhD, University of Southern California
Over the last few decades, US health policy has focused on new payment and delivery models to improve the organization and financing of care. Within the Medicare population, there has been a shift away from traditional fee-for-service payments toward alternative payment models that emphasize value-based care. However, there is little empirical evidence on the efficacy of these programs in managing the care of patients with ADRD. In this pilot, we will estimate the impact of new payment and delivery models—including accountable care organization (ACO) models, patient centered medical care homes (PCMH), value-based payment modifiers, and merit-based incentive payments (MIPS)—on the health care and drug use and cost for ADRD patients. Data on health care and drug use will come from Medicare claims data from 2008 to 2019. Data on provider (a) ACO participation will come from the Centers for Medicare and Medicaid Services’ (CMS) Medicare Shared Savings Program Files from 2013 onward and the Next Generation ACO variable in the Carrier Claims data from 2016 onward, (b) PCMH participation will come from the National Committee for Quality Assurance’s PCMH practice location files from 2012 onward; and (c) provider value-based payments from CMS’ Value Modifier NPI-Practice file from 2013-2016. We anticipate that data on MIPS incentives, which affected Medicare Part B reimbursements in 2019, will be available from administrative claims data. For each delivery model, we will employ a differences-in-differences framework that compares providers and beneficiaries participating in the model (treatment) to those unaffected (control), pre- and post-model implementation. We will focus on responses by primary care providers of ADRD patients and consequential changes in ED reliance, hospital lengths of stay, and access to new technologies among ADRD patients. Results from this research will identify which delivery and payment models are most effective in improving dementia care.
Trends and Projections of ADRD in Asia
Investigators: Jay Bhattacharya, MD, PhD, Stanford University; Karen Eggleston, PhD, Stanford University
Dementia prevalence is increasing at an even faster rate in Asia than many other parts of the world. The study will focus on the social and economic burden of ADRD over the life course in heterogeneous Asian populations, with three specific aims: (1) To develop innovative methods for estimating cognitive decline and dementia incidence and prevalence in Asian economies within pseudo-panel-based microsimulation models such as the Japanese Future Elderly Model (JFEM); (2) To study heterogeneity in the future disability burden associated with dementias and the implications for labor supply, and caregiver burden, using microsimulation models; and (3) To project the implications of cognitive decline and ADRD for health spending and long-term care spending in East Asia. We will work with members of the Japanese Future Elderly Model (JFEM) team to extend the current microsimulation model built from repeated cross-sectional data to include cognitive function and dementia. We will use this expanded JFEM to study the social and economic burden of ADRD in Japan, and compare to other FEM for North America and Europe. Existing Asia microsimulation models include those for South Korea and Singapore, with a model under development for Taiwan, and in the planning stages for Hong Kong, the rest of China, and India. We will project the health and cost implications of cognitive decline and ADRD and implications of long-term care insurance (LTCI) programs in Japan and South Korea. Changing educational attainment within and across countries, is one important element of heterogeneity for understanding cognitive decline and ADRD burden. Our research will be important for understanding these important changes in Asia and their interaction with rapid demographic and economic change.
Harmonized Cognitive Measures for International Microsimulation Modeling of ADRD
Rebeca Wong PhD University of Texas Medical Branch; Bryan Tysinger, PhD(c), University of Southern California
Projecting the future burden of ADRD is of critical importance to understand implications for individuals, families, health systems, and government programs, but doing so is a complex calculation. The first step is an accurate projection of the prevalence of ADRD. Ideally, these projections incorporate trends in sociodemographic factors, health risk factors, competing health risks, medical innovations, and more to better reflect the future population. Recent advancements in data collection potentially enable better modeling of ADRD incidence and prevalence into the future. In 2016, the Health and Retirement Study (HRS) for the United States, the Mexico Health and Aging Study (MHAS), and others (surveys in England, India, China, and South Africa) incorporated an extensive cognitive battery for some respondents. This battery, called the Harmonized Cognitive Assessment Protocol (HRS-HCAP for the U.S., Mex-Cog for Mexico) includes an hour-long cognitive assessment that incorporates several clinical cognitive tests, as well as 20-minute interviews of informants who provide insight regarding the respondents’ level of function and changes in that function. The respondent battery includes many validated measures, including MMSE, TICS, 10/66, Constructional Praxis, and CES-D. Combining the HCAP with data from the surveys will potentially allow for more nuanced modeling of ADRD prevalence, including for subpopulations of interest. This pilot project will extend two existing microsimulation models, the United States Future Elderly Model (US-FEM, based on the HRS) and the FEM-Mexico (based on the MHAS) to incorporate measures from the HCAP supplement. This project has the following aims: (1) Develop the US-FEM and FEM-Mexico to incorporate information from HCAP. Produce population and subpopulation (gender, education, and race) projections; and (2) Recommend best practices for projecting ADRD using HCAP to guide other researchers using the HRS and MHAS, as well as those utilizing data for England, India, China, and South Africa. The United States and Mexico FEMs are dynamic microsimulations based on the HRS and MHAS, respectively. They combine sociodemographic factors and competing health risks to project chronic disease, functional limitations, mortality, and economic outcomes at the individual level. The AD- FEM version of the United States FEM has been used to model ADRD from two perspectives: the first used information from a subset of HRS respondents who underwent extensive assessment, the second uses the TICS responses to project cognitive function. Both models will be extended to include prevalence models of cognitive decline using information from the HCAP. This research will produce projections of ADRD prevalence in the United States and Mexico, for the population and for subgroups of interest.