Even though a growing number of Americans are dying with dementia, several receive suboptimal care near the end of life. Reported problems include under-diagnosis and poor treatment of pain, painful and unnecessary investigations and inappropriate use of aggressive treatments. Some studies have reported that capitated Medicare-Advantage (MA) plans, that are paid on a per-person (rather than a per-service) basis, may be better positioned than traditional fee-for-service Medicare to promote the use of recommended services at the end of life while discouraging unnecessary hospitalizations and invasive procedures. In addition, the hospice ?carve-out? also creates a strong financial incentive for MA plans to promote hospice enrollment among their terminally ill patients. On the other hand, capitated payments incentivize MA plans to selectively contract with health providers and restrict services and choice of providers in order to control costs. MA beneficiaries face higher cost-sharing if they receive care from providers outside of their network. Narrow networks can also influence quality of care directly if higher-quality providers are unwilling to accept low payment rates. Therefore, the relationship between MA enrollment on service use and quality of care is potentially a mixture of different effects that vary across outcomes and patients, necessitating robust empirical evidence on both the direction and magnitude of these effects. Much of existing literature in this area focuses on the use of a few health services, does not account for potential selection bias in MA enrollment, precedes enrollment changes accompanying the 2010 health reform, and does not specifically focus on dementia patients. Our study employs multiple waves of unique population- based mortality follow-back data associated with the Health and Retirement Study (HRS) (2000- 2014) and the National Health and Aging Trends Study (NHATS) (2012-2017) to examine the relationship between MA enrollment and three categories of end of life care outcomes: patterns of care (including site of death), out-of-pocket expenditures and perceived quality of end of life care, among dementia decedents. It also harnesses the core HRS and NHATS files to account for longitudinal patterns of insurance coverage among dementia decedents. We also propose several sensitivity and heterogeneity analyses to enhance the robustness of our findings.
The results will indicate the extent to which Medicare Advantage meets the preferences and needs of terminal dementia patients. This is significant given a rapidly aging population and a forecasted growth in MA enrollment to 41% of all Medicare beneficiaries by 2027.
