Effect of payer source on hospitalization risks among nursing home residents in New York State

Hospitalizations of nursing home (NH) residents are a marker for quality and efficiency. We examine whether payer type (Medicaid vs. private-pay) and bed-hold status impact NH hospitalizations.

Data include: the 2003 Minimum Data Set (MDS), Medicaid and Medicare claim files, Medicaid cost reports. Individual observations are defined to start from a non-Medicare period. Hospitalization is the dependent, dichotomous variable. Payer status is the main independent variable of interest. Individual-level risk factors are controlled.

Facilities are stratified into 4 groups based on ownership and eligibility for bed-hold payment (determined by occupancy-rate). Base and conditional fixed-effects logit models are fit for these subgroups.

The analytical sample includes 69,307 residents (18 % private-pay & 82% Medicaid) in 564 NY NHs. 28% of residents have >=1 hospitalization during non-Medicare periods. Overall Medicaid residents are more likely to be hospitalized than private-pay residents, and this difference declines significantly after accounting for facility fixed-effects.

Compared with private-pay, Medicaid residents are more likely to be hospitalized in bed-hold-payment eligible facilities (OR=1.34, P<0.001, in for-profits with an average payer-mix level; OR=1.17, P<0.001, in nonprofits). This difference in ORs is significant. Similar effect is detected in bed-hold-payment ineligible nonprofit facilities (OR=1.25, P<0.01), but not in for-profits (OR=1.02, P=0.764) NHs.

Hospitalization risk for Medicaid and private-pay residents, adjusted for individual characteristics, is significantly different. The impact of payer status can be ascribed to both within-facility disparities and to across-facility differences. To reduce hospitalization risks for Medicaid residents, efforts should be focused on reducing both within-facility disparities and across-facility differences.