First Annual Meeting of the NBER Coordinating Center on the Economics of AD/ADRD Prevention, Treatment, and Care
The National Bureau of Economic Research (NBER) Coordinating Center on the Economics of Alzheimer’s Disease and Alzheimer’s Disease-Related Dementias (AD/ADRD) Prevention, Treatment, and Care convened its first annual meeting in Rockville, Maryland, on June 12, 2025.
John Phillips, acting deputy director of the Division of Behavioral and Social Research at the National Institute on Aging (NIA), delivered welcoming remarks and described the NIA’s rationale for funding a Coordinating Center on the Economics of AD/ADRD. AD/ADRD has been an NIA priority for a long time due to its status as a leading cause of death and substantial economic burden. The NIA’s AD/ADRD Research Implementation Milestones offer unique opportunities for economic research to inform actions that address prevention and treatment, as well as improve care for people with dementia.
The Coordinating Center has three principal investigators: Julie Bynum, University of Michigan, a physician; Rhoda Au, Boston University, a neuropsychologist; and Kathleen McGarry, University of California, Los Angeles and NBER, an economist. Their combination of experience demonstrates the Coordinating Center’s intention, as Bynum explained in her opening remarks, “to be truly interdisciplinary and multidisciplinary, informing the economics research through people who know the disease and impacted health systems well.” The Coordinating Center will fund short-term and early career research projects, and strive to create opportunities for individuals with public policy and industry experience to interact with academics to inform the Coordinating Center’s research directions.
The introductory keynote by Courtney Van Houtven, Duke University, The Caregiving Challenges to Individuals and Families Dealing with Alzheimer’s Disease and Related Dementias, centered around caregiving-related challenges for individuals living with AD/ADRD and their families. Even though the National Health and Aging Trends Study and the National Study of Caregiving estimate that the United States has over 24 million caregivers, there are few federal supports for them; Medicaid waiver programs are one important example. Van Houtven pointed out that in contrast to the common economic assumption that choices are made by rational agents in an environment with full information, patients and caregivers do not have full information on the costs, risks, and alternatives involved with long-term care (LTC). This impairs their ability to make perfectly rational decisions and suggests an important role for behavioral economics, which incorporates considerations such as loss aversion, time-varying preferences, selfishness, and limited willpower.
Van Houtven presented the results of two studies. In the first, the investigators found that having private LTC insurance helps the caregiver but does not change their likelihood of taking on caregiver responsibilities. The second study investigated the effects of family versus formal care, finding that receiving both led to improved outcomes for those receiving LTC. Van Houtven proposed some future directions for caregiving research, highlighting “models in economics that better incorporate the interdependence of caregiver and care recipient utility and account for the uniqueness of dementia in particular.” This type of research could inform policy reform and help design cost-effective policies.
Jay Bhattacharya, director of the National Institutes of Health (NIH), in his remarks on Perspectives from the NIH, emphasized the important role that social science research plays in ensuring that patients receive the benefits of biomedical research, including AD/ADRD research. Bhattacharya’s talk is available online. He described several elements of his vision for NIH’s direction. The first is the Make America Healthy Again movement, which focuses on addressing chronic conditions with the greatest effects on Americans, like Alzheimer’s disease. “We have to focus on the mission, which is to support the science that advances the health and longevity of the American people.” The second is to improve the replicability of scientific studies; he noted that “the process by which we find truth in science is by replication, by independent teams looking at the same answer.” He described increased funding for replication studies and a potential journal to publish negative results to reduce the consequences of failure. The last element he discussed was the need to “encourage scientists to think big” and to empower young researchers to try out innovative ideas.
Research Panel 1: The Economics of Caregiving for People Living with Alzheimer’s Disease was moderated by Kathleen McGarry.
Gopi Shah Goda, Brookings Institution and NBER, presented “Long-Term Care in the United States and Around the World: Caring for Those with Low Cognition.”
The study, modeled after the NBER’s International Social Security Project, seeks to learn about LTC systems by standardizing measurement and analysis techniques across several countries (Canada, Denmark, England, Germany, Italy, Japan, Netherlands, Singapore, Spain, and the United States). The project used data from each country; the Health and Retirement Study (HRS) was used for the United States, and HRS sister surveys were used for the other countries. The data were used to analyze long-term care and the cost of care for individuals with cognitive and/or physical limitations as measured by reported difficulties with activities of daily living (ADLs). The study found that in the United States, aggregate healthcare spending on individuals with low cognition (regardless of whether they had limitations with respect to ADLs) was higher than that for individuals with two or more ADL limitations but higher scores on cognitive tests, even though the population with low cognition is smaller than that with multiple ADL limitations. Much of the cost for those with low cognitive scores was driven by the need for 24-hour care, either in a nursing home or at home. Cross-country comparisons showed that the United States and England spend more on long-term care for those with cognitive limitations, while Spain, Denmark, and Singapore spend more on those with two or more ADL limitations but without low cognitive scores.
Katherine Miller, Johns Hopkins University, discussed work on “Market Changes in the Stability of Home Health Care Services: Evidence of Minimum Wage Stability and Overtime Protections.”
In 2015, home healthcare workers became eligible for minimum wage and overtime protection under federal policy. This study analyzed how this policy change impacted turnover rates, income, and earnings in home healthcare by comparing outcomes for the home healthcare workforce before and after the policy change with those for workers in animal slaughter and processing, who have higher job stability and similar levels of education, and childcare workers, who have similar demographics to home healthcare workers. The study found that after 2015, home healthcare workers had decreased turnover compared to the control groups, indicating that the federal policy had the desired effect on improving home healthcare workforce stability.
The last presentation of the panel was by Yulya Truskinovsky, Syracuse University and NBER, on “Family Spillovers of AD/ADRD.”
Using administrative data from Denmark, this study investigated the impact of AD/ADRD on adult children. Children were linked to a parent and tracked for 10 years before and 7 years after the AD/ADRD-related death of the parent. In Denmark, because the government provides generous funding for home-based care coupled with robust support for family caregivers, the effects on children are not due to direct primary caregiving. To isolate the effect that an AD/ADRD-related death had on adult children, the study compared parent-child pairs in which the parent died of AD/ADRD to those in which the parent died of an unrelated cause. The study found no significant effect on labor market outcomes and only a very small effect on physical health outcomes around the parent’s death. The adult children whose parents died from AD/ADRD had increased utilization of mental healthcare and psychotropics, peaking around the time of the death but ramping up in the years prior. This effect was more pronounced for only children, daughters, and those who lost their parent early in life. This study concluded that the effects of AD/ADRD on adult children are mitigated, but not eliminated, by a robust social welfare state.
Research Panel 2: The Economics of Medical Care for People with Alzheimer’s Disease was moderated by Julie Bynum.
Cyrus Kosar, Brown University, discussed research on the “Impact of Institutional Special Needs Plans on Hospitalization Rates for Nursing Home Residents with Dementia.”
The institutional special needs plan (I-SNP) model, an alternative financing and delivery-of-care model that may be promising for nursing home residents with AD/ADRD, has two key features. First, the institution receives fixed per-beneficiary-per-month payments from Medicare to finance Part A and Part B benefits, which incentivizes more efficient utilization but could also lead to reduction in necessary care. Second, I-SNPs are required to develop and administer a Medicare-approved model of care. This research examines the impact of I-SNP enrollment on hospitalizations and other outcomes among nursing home residents with AD/ADRD using Medicare administrative data, the Minimum Data Set (a nursing home quality assessment), and public skilled nursing facility enrollment reports. Initial results indicate that I-SNP enrollees are healthier than other beneficiaries. I-SNP enrollment is around 25 percent two to three years after initiation in a nursing home, with higher uptake for people who had Medicare Advantage than for those who had traditional Medicare at admission. I-SNP enrollment is associated with about a 3 percent reduction in the chance of being hospitalized.
Amresh Hanchate, Wake Forest University, presented findings on “Telemedicine: A Substitute or Complement for Home-Based Medical Care? The Experience of Medicare Enrollees with Alzheimer’s Disease and Related Dementias.”
The federal government expanded telemedicine availability in March of 2020 due to the COVID-19 pandemic. This project examined how telemedicine expansion affected the use of home-based medical care (i.e., house calls). A large share of those diagnosed with AD/ADRD are homebound with advancing disease and worsening impairment, and house calls are an important part of their care. Medicare provides partial coverage for house calls for homebound enrollees, less than for office-based visits, but the volume of these calls is generally low. AD/ADRD-related visits account for two-thirds of total house calls in traditional Medicare. However, the availability of providers varies widely by region, with some regions having as few as 1.5 providers per 10,000 traditional Medicare enrollees. Telemedicine expansion was associated with increased house calls in regions with high home-based medical care availability.
The panel concluded with Lauren Nicholas, University of Colorado, reviewing results from “How Do Surrogates Make Treatment Decisions for Patients with Dementia? An Experimental Survey Study.”
Most adults have some cognitive impairment by the time they need end-of-life care, so they need to have a surrogate decision-maker. End-of-life care is a major driver of Medicare spending and an area where recipients report preferences do not match well with the care provided. To investigate how these decisions are made, the researchers conducted a large, online experiment asking what the surrogates would do for a hypothetical 85-year-old patient who was currently unconscious and hospitalized with a sudden, serious illness. Physically invasive treatments to sustain life were available but would not restore health. Survey respondents were asked about their own treatment preferences and the factors that influenced them. Similar to real-life data, the results showed that surrogates chose life-sustaining treatments less often for those with dementia than for those without, even when the patient wanted life-sustaining treatments. The respondents recommended overruling treatment preferences in about 30 percent of all scenarios; respondents whose personal treatment preferences matched the hypothetical patient’s preferences were more likely to follow the patient’s preferences. The findings suggest that older adults may wish to choose surrogate decision-makers whose preferences align with their own.
Research Panel 3: Early Detection, Causes, and Classification of Dementia Symptoms was moderated by James Poterba, NBER.
Brian Beach, Vanderbilt University and NBER, presented initial findings from “Lead Exposure and Cognitive Impairment Throughout the Lifecycle.”
The prevalence of dementia in younger cohorts is decreasing, which might be explained by decreases in childhood lead exposure. Because lead has similar chemical properties to calcium, it lodges in the synapses where calcium is needed, lowering cognition, prefrontal cortex functioning, and impulse control. Many of the studies that link childhood lead exposure to lower cognition are correlational, and childhood lead exposure is confounded with several other potential risk factors for dementia, such as living close to a highway, low socioeconomic status, and low educational attainment. To disentangle these risk factors, this study investigated the link between individuals’ childhood exposure to waterborne lead through service lines and subsequent cognitive decline. This source of lead exposure is not highly correlated with other risk factors for dementia. Initial results demonstrate a gap in early life cognitive difficulties for individuals in areas with lead pipes compared to individuals in other areas; this gap disappears by age 70. The project will continue to use variation in lead solubility to isolate the causal effect of lead exposure on cognition, using Medicare data to explore actual AD/ADRD incidence.
Michelle Mielke, Wake Forest University, shifted the panel discussion from dementia risk factors to detection, discussing research on “Experience with Implementing Cognitive Testing and Blood-Based AD Biomarkers in Primary Care Practices.”
Cognitive impairment (CI) is very common in older adults and should be diagnosed during primary care visits due to long wait times for neurologists. However, patients have very limited time with their primary care physician, and those at greatest risk for AD/ADRD are less likely to have wellness visits. Diagnosing CI early is important for patients as treatments and strategies exist to mitigate the impairment. Mindcare365 is a reimbursable pathway to improve CI detection and follow-up. It places a limited burden on primary care physicians and could work in both urban and rural areas within current health system workflows. Patients are referred to the pathway by family members, self-referral, or physicians, and then screened for CI using a telephone-based test. If positive, the patient will be tested for blood-based biomarkers and then be referred to a specialist if desired. This pathway for finding CI has gotten positive initial feedback from primary care physicians, and preliminary results indicate that blood-based biomarkers can help assess etiology but are not entirely accurate.
Péter Hudomiet, RAND Corporation, concluded the panel with a presentation on “Disparities in the Accuracy of Claims-Based Dementia Classifications in Medicare Data.”
In administrative datasets, claims-based dementia classifications introduce errors due to undiagnosed cases and false positives. Using Health and Retirement Study data linked to Medicare fee-for-service data and restricting the sample to those 65 years of age or older, this project evaluated the accuracy of claims-based dementia classification, specifically the 27 Chronic Conditions Data Warehouse (CCW) algorithms, and stratified disparities by age, sex, race, ethnicity, and other dimensions. Investigators recreated the CCW algorithms from raw data and then applied them to those interviewed in the HRS, checking for a dementia diagnosis, before calculating the algorithm’s sensitivity, specificity, and prediction accuracy. The CCW underestimates dementia prevalence by 4 percent compared to the HRS Aging, Demographics, and Memory Study data, an accurate but small clinical dataset. The sensitivity of the CCW algorithms is significantly higher for women than men, probably because women with dementia are more likely to seek medical care and receive an official diagnosis. Hispanic and Black individuals are more likely to have dementia compared to White individuals, but the Langa-Weir measure, the classification typically used, appears to overpredict the gap. Overall, the 27 CCW algorithms have moderate sensitivity, high specificity, and decent accuracy with relatively little variation across subgroups.
The conference concluded with a keynote from Amitabh Chandra, Harvard University and NBER, on Transformational Medicines: Lessons from Social Science.
This presentation considered how economists might expedite the translation of a transformative therapeutic for AD/ADRD into patient treatment. Chandra started with a case study: the timeline for the development and approval of Casgevy, a cure for sickle cell disease. In this scenario, which he considered a best-case outcome with a direct trajectory from basic science research to patient care, treatment development still took 75 years. He noted that the AD/ADRD field is not well organized and that there remain open questions about pathophysiology and resulting drug targets. The incentives to develop a transformational intervention are weak, especially for preventative interventions. Chandra described the steps in the value chain for drug discovery: foundational science to determine pathophysiology and drug targets, an intellectual property regime that fits the discovery (i.e., longer patents for preventative molecules to give enough time for a full clinical trial), capital markets to finance drug development, high-quality independent regulators to approve these technologies, and patient access through health insurance and cost sharing.
Chandra noted that the foundation of this process is fundamental scientific research. He reported on a recent study in which he analyzed over 500,000 research articles published in the 15 leading life-science journals between 1945 and 2022 to investigate where fundamental discoveries in life science occurred. The United States produces around 50 percent of the world’s basic science research in his study, with Boston specifically producing around 10 percent and the Bay Area another 8 percent. These cities’ productivity is driven by individual institutions producing large shares of the world’s total discovery research. Chandra’s study estimates that close to 60 percent of a scientist’s productivity may be the result of that scientist’s institution’s productivity. This may be driven by resources, facilities, graduate students, and postdoctoral associates at the university. He concluded that “in an era of declining NIH budgets … these ideas of spreading the money around are very well-intentioned… but if you're trying to produce science, that's not one of the reasons to spread it around.”
The conference described in this summary was a public meeting hosted by the National Bureau of Economic Research (NBER) Coordinating Center on the Economics of AD/ADRD Prevention, Treatment, and Care, funded by the National Institute on Aging pursuant to grant U54 AG090084. The opinions and conclusions expressed are solely those of the meeting participants and do not represent the opinions or policy of any agency of the Federal Government or NBER. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of the contents of this report. References herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply endorsement, recommendation, or favoring by the United States Government or any agency thereof.
