Patient vs. Provider Incentives in Long Term Care

Long-term care (LTC) expenditures are high and rising. In light of increasing demand for LTC services, it is critical that public policies align patient and provider incentives to achieve an efficient utilization of LTC. In this paper, we study how financial patient and provider incentives affect the mode and cost of LTC as well as their health consequences. This setting is of particular interest for at least two reasons. First, LTC services and, in particular, nursing home care are largely paid for using public funds but provided privately. Separating the role of patient and provider incentives is therefore key for the optimal design of Medicaid policies, which affect utilization of services through cost-sharing and reimbursement regulations. Second, and in addition to the significant spending implications, refining current policies may have important implications for the health of a particularly vulnerable elderly population.

We focus our analysis on the timing of nursing home discharges to the community. More than 40% of nursing home stays end with a discharge to the community, suggesting that community-based care is a feasible alternative for a significant fraction of residents. The precise timing of discharges is largely at the discretion of the nursing home discharge manager and the patient, so it is plausible that economic incentives affect LTC utilization at this margin. To estimate the role of patient and provider incentives, we exploit variation in out-of-pocket prices among residents who spend down their assets and transition to Medicaid during their stay. We also exploit variation in the fraction of occupied beds, which affects the nursing home’s incentive to discharge Medicaid beneficiaries. We investigate these effects using micro data from the Long-Term Care Minimum Data Set (MDS) combined with Medicaid and Medicare claims data.

We find that at low occupancies, when providers have little incentive to discharge Medicaid or private patients, weekly Medicaid discharge rates are less than half as large as private discharge rates. This suggests that patient financial incentives affect the length of stay. As the occupancy rate increases towards full capacity, we see an economically and statistically significant increase in the discharge probability for Medicaid beneficiaries but no effect for residents who pay out-of-pocket, suggesting that provider incentives also influence the length of stay. Moreover, we find no evidence that shorter nursing home stays (on the margin) lead to increases in hospitalization or mortality rates or worsened health status at discharge.

We then develop and estimate a dynamic structural model of nursing home discharges. The model estimates imply a provider elasticity of 1.6 and a patient elasticity of 0.5, indicating that providers react more elastically to financial incentives than patients. Our counterfactual policy simulations based on the estimated structural parameters show that changing the timing of reimbursements is very effective is shortening Medicaid stays and lowering public spending. Specifically, transitioning 10.5% of current Medicaid per-diem reimbursements to an episode-based (up-front) reimbursement reduces the length of Medicaid stays to the length of private stays and yields annual total cost savings of about $0.18 billion.