Poster Paper: Family, Income, & Medicaid Policy: Multinomial Logistic Model of Long-Term Care Decisions

Saturday, November 10, 2018
Exhibit Hall C - Exhibit Level (Marriott Wardman Park)

*Names in bold indicate Presenter

Selena Caldera, University of Texas, Austin


Providing care for older family members is a difficult and costly decision; in 2015, 34 million Americans provided care to an elder loved one (National Alliance for Caregiving & AARP Public Policy Institute, 2015). Formal long-term care (LTC) in a facility outside the home is expensive and caregivers that provide informal LTC at home may incur high opportunity costs, including limiting work hours or forgoing employment outside the home. Existing research focuses on this supply issue and ignores how characteristics of elders and their family members and Medicaid LTC policy jointly factor into the decision between formal and informal LTC. In this paper, I examine informal LTC in the context of all other alternatives for care (no care, informal LTC, formal LTC) and include demographic and economic measures of the elder care recipient and potential family caregivers as explanatory variables.

Incorporating characteristics of both the elder care recipient and potential family caregivers produces a more comprehensive model of the LTC decision that considers the linked lives of parents and adult children. I also can test the relative importance of economic and demographic factors of the elder and potential caregivers, and the role of public LTC policy in individual LTC decisions. This nuanced understanding of the individual care decision-making process provides evidence useful in designing future LTC policy that better responds to the needs of elders and their families. Responsive LTC policy ultimately enables improved economic and health outcomes for care recipients and their families.

I use Andersen’s behavioral model of health services utilization as a framework for identifying how different demographic, economic, and family characteristics enter the LTC decision as predisposing characteristics and enabling resources. The analytic sample includes Health and Retirement Study (HRS) respondents age 55 and older that experience new functional limitations between 2000 and 2010 and have a living partner or child. I draw from the 1998 to 2012 waves of the HRS RAND Version P, the RAND HRS Family data, and the RAND HRS Enhanced Fat Files. By pairing a discrete time multinomial logistic regression with individual-level longitudinal data on elders experiencing a functional limitation, I analyze the decision between formal LTC, informal LTC, and no care for elders at risk.

Medicaid eligibility is the strongest enabling characteristic indicating that state-level Medicaid LTC policies strongly influence individual choices of formal LTC over informal LTC. Specifically, I find that Medicaid-eligible respondents are 104 percent more likely to use formal LTC over forgoing care and 70 percent more likely to choose formal care over informal care. These findings demonstrate that LTC choices are strongly influenced by state and federal LTC policies. The next step in this research is to expand the model to include state identifiers and detailed measures of the relative generosity of state Medicaid guidelines and the availability of options that facilitate informal LTC. The wide variation in Medicaid LTC policies across states and increased focus on home and community-based options offer a ripe setting for further analyses.