Publications

Combining capitation with a cap on health care providers’ average revenue can reduce allocative inefficiency. But the cap may be undercut by health care providers who churn their patient censuses. We investigate this possibility in the U.S. hospice industry, where Medicare pays hospice programs fixed daily rates but caps their average annual revenue. By leveraging variation generated by the cap’s nonlinear design and the transition between fiscal years, we find that programs on track to exceed the cap raise enrollment rates by 5.8 % and live discharge rates by 4.3 % in the fourth quarter. But this churning falls far short of eliminating their financial penalties: it amounts to 10 % of an average program’s excess revenue at most. Marginal enrollees have longer remaining lifetimes and more fragmented hospice spells on average, suggesting weaker intrinsic demand for hospice care. We discuss the cap’s implications for market structure.

The voluntary Pennsylvania Rural Health Model (PARHM) provided participating rural hospitals with global budgets rather than traditional fee-for-service payments to improve financial stability. Whether participation in PARHM was associated with financial improvements is unknown. In this synthetic difference-in-differences study from the period 2014-23, we estimated changes in rural hospital finances associated with PARHM participation among seventeen PARHM participants and forty nonparticipating Pennsylvania comparison hospitals. In unadjusted models, participation was associated with a 4.5-percentage-point differential increase in operating margins and a 4.7-percentage-point differential increase in total margins; however, these changes were nonsignificant in adjusted models (3.0 and 3.2 percentage points, respectively). Results were similar when we compared PARHM participants with 160 border-state comparison hospitals and when we used alternative difference-in-differences estimators. These findings offer mixed evidence that global budgets may help stabilize rural hospital finances in the short term.

Alternative payment models, such as bundled payment, have been proposed as a solution to the high costs of health care. While these models are typically effective at constraining spending on post-acute care, the decrease in consumption of formal post-acute care may result in a compensatory increase in the need for and use of informal or family caregiving. We estimate the effect of a large, randomized experiment with Medicare bundled payment on the need for and receipt of caregiving. Using data on over 2 million Medicare beneficiaries undergoing knee or hip replacement and a difference-in-differences approach, we find that the mandatory bundled payment caused a 1 to 2 percentage point absolute increase (a 9 percent to 14–15 percent relative increase) in both the need for and receipt of help with activities of daily living at the end of a home health episode, help that was likely provided by family caregivers. This increased caregiver burden was corroborated by a large shift away from nursing-home-based post-acute care (or care in a skilled nursing facility or SNF) after knee or hip replacement, a shift toward home health care, and an accompanying decline in the intensity of home care.

Although long-term care remains one of the largest uninsured risks facing older Americans, demand for insurance remains low. While there is a long literature estimating a variety of factors that contribute to this low demand, much of it has overlooked the fact that most private long-term care insurance (LTCI) purchases are made within couples, adding a host of additional reasons for low demand. This paper examines the role of financial decision-making power within the couple and the association with LTCI purchase decisions. We document LTCI purchase patterns among married couples and find that, among couples who ever purchase LTCI, they are roughly equally likely to purchase for the woman exclusively (10.0%), the man exclusively (11%), or both (11%). However, among couples where women have more bargaining power, LTCI purchases are more likely overall (40% vs. 33%), and more likely to cover the woman, either exclusively (16% vs. 11%) or as part of both members of the couple (14% vs. 11%), than among couples with more traditional gender roles. In adjusted analyses, we find that women are more likely to be insured when they have more bargaining power. These findings suggest that intra-household bargaining power may be another potential explanation for the particularly low LTCI take-up, especially in the time period in which policies were unisex-priced.

Long-term services in the home are predominately provided by family or friends, with a growing proportion of individuals receiving formal care, or paid care by a professional, or a combination of both. However, the relative benefits to the care recipient of who provides the care are largely unknown. A person's use of formal and family care is affected by factors that also may affect their outcomes, complicating the estimation of any causal relationship. Using the 2002-2018 Health and Retirement Study (HRS), we examine three types of home-based care combinations: family only, formal only, and both formal and family care. We use an instrumental variables strategy, using family structure as instruments for both formal care and the combination of formal and family care, to estimate the plausibly causal impact of the care provider on self-reported mental and physical health outcomes. We find that, once the endogeneity of the care provider is accounted for, having both formal and family care leads to better self-rated health, mobility and lower depression compared to people receiving family care only. Receiving formal care only does not affect care recipient outcomes compared to receiving family care only. These results are robust to several sensitivity analyses, including different instrument specifications, subsamples of care recipients that do not have a spouse/partner, among women care recipients, and changing the timing of the measurement of the outcomes. These findings are important to consider as we strive to best meet the growing demand for person-centered, high-quality long-term care in the least restrictive setting possible.

Consumer credit has shown increasing relevance to the health of older adults; however, studies have not been able to assess the extent to which creditworthiness influences future health or health influences future creditworthiness. We assessed the relationships between 4-year pre and postmorbid consumer credit history and self-rated physical and mental health outcomes among older adults. Generalized estimating equations models assessed pre and postmorbid credit history (credit scores, derogatory accounts, and unpaid accounts in collections) and the onset of poor self-rated health (SF-36 score <50) among 1,740 participants aged 65+ in the Advanced Cognitive Training for Independent and Vital Elderly study from 2001 to 2017, linked to TransUnion consumer credit data. In any given year, up to 1/4 of participants had a major derogatory, unpaid, or collections account, and up to 13% of the sample had poor health. Each 50-point increase in credit score trended toward a 5% lower odds of poor health in the next 1 year, a 6% lower odds in the next 2 years, and a statistically significant finding of 13% lower odds by 3 years. A drop in credit score was associated with a 10% greater odds of poor health in the next year, and having a major derogatory account was associated with an 86% greater odds of poor health in the next 3 years. After poor health onset, credit scores continued to see significant losses up to the 3 years, with larger decrements over time. Having a major derogatory account or a sudden loss in credit may be a time to monitor older adults for changes in health. After a downturn in health, supporting older adults to manage their debt may help stabilize their credit.

No abstract available

Although neurodegenerative diseases are a leading cause of death, little is known about health care utilization and cost during the end-of-life (EoL) period or how it compares with that of other life-limiting conditions. We aimed to describe and compare resource utilization among US Medicare decedents with neurodegenerative diseases with decedents with cancer. We conducted a retrospective study of Medicare Part A and B beneficiaries with Alzheimer disease (AD), Parkinson disease (PD), or amyotrophic lateral sclerosis (ALS) who died in 2018. Decedents diagnosed with malignant brain tumors or pancreatic cancer served as non-neurodegenerative comparators. Descriptive analyses examined demographic and clinical characteristics in the last year of life. The probabilities and associated costs of emergency department (ED), inpatient, skilled nursing facility (SNF), and hospice utilization during the last 12 and 6 months of life were also compared between persons with neurodegenerative diseases and cancer, adjusting for sociodemographic factors and comorbidity burden. A total of 1,126,799 Medicare beneficiaries died in 2018, of which 357,926 had a qualifying diagnosis. Persons with neurodegenerative diseases were older and more frequently received Medicaid assistance than persons with brain or pancreatic cancer. In all groups, health care service utilization increased over the last year of life, and total costs were predominantly attributable to inpatient care. In the last 6 months of life, neurologist care was infrequent among patients with neurodegenerative disease (AD: 1.5%; PD: 8.6%; ALS: 32.0%). Persons with neurodegenerative diseases as compared to persons with malignant brain tumors also had greater odds of ED use (AD: adjusted odds ratio [aOR] 1.17, 95% CI 1.11-1.23; PD: aOR 1.18, 95% CI 1.11-1.25; ALS: aOR 1.11, 95% CI 1.01-1.23), lower odds of hospitalization (AD: aOR 0.64, 95% CI 0.60-0.68; PD: aOR 0.65, 95% CI 0.61-0.69; ALS: aOR 0.33, 95% CI 0.30-0.37), and lower odds of hospice enrollment (AD: aOR 0.33, 95% CI 0.31-0.36; PD: aOR 0.33, 95% CI 0.31-0.36; ALS: aOR 0.41, 95% CI 0.36-0.46). The findings were similar in pancreatic cancer. Persons with neurodegenerative diseases in the United States are more likely to visit the ED and less likely to use inpatient and hospice services at EoL than persons with brain or pancreatic cancer. These group differences may stem from prognostic uncertainty and reflect inadequate EoL care practices, requiring further investigation to ensure more timely palliative care and hospice referrals.

Persons living with Alzheimer's and other related forms of dementia rely heavily on care from family and friends for assistance with daily activities ("family care"), but little is known about care transitions over time. We analyzed data from the Health and Retirement Study to describe caregiving patterns, from 2 years before dementia onset and up to 6 years after. Using sociodemographic data from the interview prior to dementia onset, we determined if there are significant factors that predict receipt of family care at dementia onset. We found that one-third (33%) of people living with dementia were receiving help with daily activities 2 years prior to their first positive dementia screen and this increased to 60% during the first positive screen. Nearly all of those receiving assistance received family care. We found multiple significant predictors of receiving family care at onset, including race, education, access to private health insurance, number of activities of daily living that were difficult, number of chronic conditions, and already receiving help. This demonstrates potential gaps in dementia care, and which subpopulations may benefit most from targeted interventions for household members who do not have adequate caregiving resources or programs that provide additional formal care.