Facility and county effects on racial differences in nursing home quality indicators.

This study's goal was to examine the effects of nursing home (NH) and county racial mix on quality of care in NHs. We examined quality indicator (QI) outcomes for residents in 408 urban New York NHs in July through September, 1995. The QI outcomes studied were restraint and antipsychotic drug use (for low and high-risk residents), and at study commencement, these QIs were being used by the Centers for Medicare and Medicaid Services to monitor the quality of care in USA Medicare and/or Medicaid-certified NHs. A hierarchical modeling approach was used to properly reflect the nesting of both residents within NHs and NHs within counties. Separate regression models were fit to the two strata of interest (Urban Non-Hispanic Whites and Urban African Americans) to test, for each race group, the effect on quality of residing in NHs and counties with higher proportions of African Americans (than state medians). Descriptive analyses found that, compared to Whites, the unadjusted restraint rate was lower for African Americans while the antipsychotic drug rate was higher. For both race groups, multi-level analyses showed residence in for-profit NHs was associated with higher likelihoods of being restrained, and of receiving antipsychotic drugs. Also, for both race groups, residence in NHs with higher proportions of African-Americans was associated with lower likelihoods of being restrained and with higher, statistically nonsignificant, likelihoods of receiving antipsychotic drugs. Higher NH nurse staffing ratios were associated with higher likelihoods of being restrained and with lower likelihoods of antipsychotic drug use (statistically significant for low-risk African-Americans). Findings support the notion that differential care is provided in USA NHs caring for higher proportions of African-American residents and thereby suggest intervention at the organizational level is warranted to improve QI outcomes for both race groups.