Examining differences in death rates for medicaid and non-medicaid nursing home residents

The purpose of this work was to examine differences in the probability of death for Medicaid and privately funded nursing home residents, controlling for differences in facility, market, and resident characteristics.

Given that Medicaid residents are more likely to die in nursing facilities, the probability of dying within 1 or 2 years is estimated using a series of probit models, controlling for whether a resident is Medicaid-funded or privately-funded. As resident characteristics, market attributes, facility characteristics, and facility fixed effects are sequentially added to the specifications, the gap in the probability of death between Medicaid and private-pay residents is considered.

The overall mortality rate for Medicaid residents was 14.8% points higher than the death rate for privately funded residents. When considering death within 1 year and 2 years, Medicaid resident death rates were 4.2% and 7.8% higher, respectively. The apparent difference in mortality declines as one controls for resident, market, and facility characteristics. Both facility characteristics and facility fixed effects are relatively important in explaining differences in death rates between Medicaid and private-pay residents.

Differences in death rates between Medicaid and private-pay residents are relatively small, suggesting that residents with different payment sources are not treated differently in nursing homes in ways that impact the probability of death. However, policymakers may want to look more closely at whether Medicaid residents are segregated into lower-quality facilities.