Home
Research

The Bulletin on the Economics of Alzheimer's Disease and Related Dementias

A summary of research funded through the NBER Coordinating Center on the Economics of Alzheimer’s Disease and Alzheimer’s Disease-Related Dementias Prevention, Treatment, and Care, as well as other related work. Produced twice a year.

X LinkedIn Facebook Bluesky Threads Email Link

Probable Dementia and Health Insurance Choices

10/15/2026

Summary of working paper 35606

This figure is a line chart titled “Medicare Coverage Among People with Probable Dementia (PWPD).” It compares the share of people enrolled in traditional Medicare, with and without supplemental coverage, among people with probable dementia and a control group, in the years before and after a dementia diagnosis. The y-axis is labeled “Percent enrolled in traditional Medicare” and ranges from 10% to 50% in increments of 10 percentage points. The x-axis is labeled “Years relative to dementia diagnosis” and ranges from −10 to 6 in increments of 2 years. The chart contains four lines. A solid blue line represents traditional Medicare with supplemental coverage for people with probable dementia, and a dashed blue line represents traditional Medicare with supplemental coverage for the control group. A solid dark gray line represents traditional Medicare without supplemental coverage for people with probable dementia, and a dashed dark gray line represents traditional Medicare without supplemental coverage for the control group. A vertical dashed line at year 0 is labeled “Dementia diagnosis.” For people with probable dementia, enrollment in traditional Medicare with supplemental coverage declines from about 48.5% at 10 years before diagnosis to about 45.5% at 4 years before diagnosis. It then falls more steeply to about 37% at diagnosis and to about 26% at 6 years after diagnosis. For the control group, enrollment with supplemental coverage declines gradually from about 47% at year −10 to about 45.5% at year 0 and about 43% at year 6. For people with probable dementia, enrollment in traditional Medicare without supplemental coverage rises from about 17% at year −10 to about 29.5% at year −2, dips slightly to about 28% at diagnosis, peaks at about 33% at year 2, and ends at about 30% at year 6. For the control group, enrollment without supplemental coverage stays nearly flat, rising slightly from about 22.5% to about 23.5% over the period. Among people with probable dementia, the two lines cross shortly before year 2, after which more people are enrolled in traditional Medicare without supplemental coverage than with it. A note on the figure reads: Control refers to respondents never assessed with probable dementia during the sample period. The source line reads: Researchers’ calculations using 2000–2016 data from the Health and Retirement Study.

Nearly one in ten older Americans lives with dementia, a condition that gradually erodes memory and the cognitive skills needed for complex decision-making. Even though Medicare coverage is nearly universal among this age group, traditional Medicare alone leaves significant gaps as it does not cover dental, vision, or hearing care, and it carries substantial cost-sharing. To fill these gaps, beneficiaries must actively choose among several insurance options—enrolling in Medicare Advantage, adding private supplemental insurance to traditional Medicare, or, for those with sufficiently low income and assets, applying for Medicaid. Beneficiaries may make changes during open enrollment each year and must re-enroll in private supplemental coverage or Medicare Advantage if their plan is no longer offered or they fail to pay premiums. Those who are eligible for Medicaid must recertify each year.

This annual decision-making process, which may involve interacting with multiple agencies, may be especially difficult for individuals whose cognitive abilities are declining. This may leave them with incomplete insurance protection at a time when they need it most. In Supplemental Medicare Insurance Coverage for Older Adults with Probable Dementia (NBER Working Paper 35606), Melissa McInerney, Lorena Moreno, and Jennifer M. Mellor examine whether previously documented gaps in supplemental insurance coverage among those with cognitive impairment actually begin to emerge before individuals meet the criteria for probable dementia.

Individuals who develop probable dementia are more likely to have Medicare coverage with no supplemental insurance beginning four years before they first meet the criteria for probable dementia.

The researchers use the 2000–2016 Health and Retirement Study, a nationally representative longitudinal survey of Americans age 50 and older. Probable dementia is identified using the Langa-Weir algorithm, a standard method that scores cognitive function based on tests of memory and other skills. The study sample includes 3,001 Medicare-enrolled adults age 65 and older who eventually meet criteria for probable dementia and 14,055 people who never do. For each respondent with probable dementia, the researchers identify the wave in which they first meet the algorithm's criteria for probable dementia, referred_to_hereafter as onset, and track insurance outcomes from ten years before onset to six years after.

Comparison groups are constructed separately for each time period using propensity score weighting, which adjusts for differences in age, sex, education, and marital status so that people with and without probable dementia are compared to similarly situated peers. Beginning four years before onset, those later classified as having probable dementia are 3 percentage points more likely than matched peers to have no supplemental Medicare coverage. This gap is even larger after onset, peaking at 9.5 percentage points two years later.

This gap in "bare" Medicare coverage reflects the net effect of lower enrollment in Medicare Advantage and traditional Medicare with private supplemental coverage, partly offset by higher Medicaid enrollment.

Individuals who develop probable dementia also become significantly more likely to have income at or below the federal poverty level—a standard proxy for Medicaid eligibility. This gap emerges six years prior to onset and continues to widen. However, among those who meet this income-based eligibility threshold, the rate of Medicaid enrollment does not differ significantly between the two groups in the years before or immediately after onset.

The researchers acknowledge support from the Tufts University Faculty Research Awards Committee.