Preferences for Long-term Care Coverage among Elderly Medicare Beneficiaries: The Role of Cognitive Ability

Long-term care (LTC) provides help with basic personal tasks of everyday life. LTC expenditure constitutes a large and growing component of health care expenditure among the elderly in the United States. Since Medicare does not provide coverage for LTC, the ability to assess and plan one’s LTC needs is particularly important for the well-being of the aging population.

We examine preferences for LTC coverage among elderly Medicare beneficiaries and their determinants. We use data from a unique module of the 2006 Health and Retirement Study (HRS) that asked elderly Medicare beneficiaries (age 66 and above) to choose among hypothetical alternative Medicare plans that include those with LTC coverage (as opposed to intensive hospital care coverage and cash back options). Of 1,730 respondents who completed the module, about 58% indicated some preferences for LTC coverage in Medicare program.

We find that, contrary to expectations, having difficulties with one or more activities of daily living (ADL) strongly and negatively predicts preferences for LTC coverage. This relationship persists after conditioning on the beneficiary’s sociodemographic characteristics, existing LTC and Medicaid coverage, availability of friends/relative for help with LTC, out-of-pocket medical care cost and a number of chronic and acute health conditions. We hypothesize that, since ADL difficulty is a strong marker for dementia and other cognitive problems, this relationship may be explained by the inability of cognitively impaired elderly to understand and predict their LTC needs. Our preliminary analysis finds supportive evidence for this as the negative relationship between LTC preferences and ADL difficulty is stronger among respondents who did poorly on the 10-word learning and recall test, a widely used instrument for dementia diagnosis.

Our findings raise concerns about the ability of cognitively impaired elderly to seek LTC help, especially given their high likelihood of needing LTC. In continued analysis, we plan to 1) construct more refined measure of dementia using all available cognitive tests in HRS data, and 2) use stated preferences for LTC coverage to predict future ADL needs using more recent waves of HRS, as well as examine potential heterogeneity in the predictive power by cognitive ability.