Table 1.

Summary of All Articles Used in This Systematic Review on Measuring NH Segregation and Health-Related or Quality Subsequent Outcomes

				Sample size
Reference	Design, years	Location	Population	Residents	Facilities*	Measure of segregation	Health or quality outcome measure	Summary of findings
Chang and colleagues (2012)	Cross-sectional, 2009 (2007–2008: MDS data)	Missouri (MO; United States)	All long-term and short-term care Medicare/ Medicaid NH residents; black and white residents alone for the Dissimilarity Index calculation	N/A	n = 511	Standardized indices	Facility healthcare quality measures (high-risk pressure ulcer, physical restraint, influenza vaccination, pneumococcal vaccination, disparities quality index)	A high percentage of black residents are concentrated in few MO NHs; more segregated NHs had more observed disparities and poorer outcomes in quality measures. The percentage of black residents was the strongest predictor variable of poor quality outcomes, Disparities in quality appeared in three of the four measures: high-risk pressure ulcers, influenza vaccines, and pneumococcal vaccines.
Davis and colleagues (2014)	Cross-sectional, 2004 (2000: Census data)	National (USA)	All Medicare/ Medicaid certified NHs	N/A	n = 8,950	Standardized indices	NH market competition/ quality (segregation, bed occupancy)	The mean distribution of racial/ethnic minorities in U.S. NHs is less than that in the community (MSAs). NH markets appear to be halfway between integration and segregation (DI mean ± SD = 0.51 ± 0.11). NHs in more segregated regions were less racially/ethnically diverse. Organizational NH factors (e.g., ownership) and competition (e.g., occupancy) were also associated with racial/ ethnic diversity. NHs with more empty beds (higher excess capacity) had more diversity.
Feng, Lepore, and colleagues (2011)	Longitudinal, 1999–2008, (2000: Census data)	National (USA + Washington DC)	All Medicare/ Medicaid NHs	N/A	n = 18,259	Standardized measure	NH closures (concentration of and clustering patterns) and changes in bed supply	A higher rate of NH closures and greater amount of net beds lost was observed in zip codes with a higher percentage black/Hispanic population or higher rate of poverty. The weighted Gini coefficient indicated more inequalities on the local zip code level (0.71) and a more equal distribution of closures when evaluating disparities on the larger MSA scale (0.55).
Miller and colleagues (2006)	Cross-section, 1995, (1989/1990: Area Resource File)	New York (urban regions)	Black, white NH residents	n = 60,932 (restraint use outcome); n = 58,763 (antipsychotic use outcome)	n = 716 (restraint use outcome); n = 806 (antipsychotic use outcome)	Non-standardized model	Quality indicators (QIs; daily restraint use and antipsychotic drug use in the absence of psychotic conditions among low and high-risk patients)	NH care is different among facilities with higher proportions of black residents. Both black and white NHs residents in higher proportion black facilities were shown to have a higher likelihood of antipsychotic drug use and lower likelihood of restraint use (characteristics also associated with lower NH staff ratios) compared to facilities with lower proportions of black residents. Additionally, NHs with higher proportions of black residents also had higher percentages of Medicaid beneficiaries, were more likely to be for-profit, and were in facilities in counties that also had relatively high proportion of black residents.
Smith and colleagues (2007)	Cross-sectional, 2000	National (United States)	Black, white NH residents	n = 1,466,471	n = 14,374	Standardized index	Facility quality measures (inspection deficiencies, staffing, financial viability)	NHs appear to be segregated with a DI that averages up to 0.73 in the Midwest. Black NH residents are more likely to be in lower quality NHs (more deficiencies, lower staffing ratios, more financial strain) than white residents.
Strully (2011)	Cross-sectional, 2004	National (United States)	Black, white NH residents	n = 12,501	n = 1,140	Standardized index, proxy measure	Receipt of influenza vaccination	Segregation of black NH residents across all sample facilities is relatively high (DI = 0.71 from 0 [lowest] to 1 [highest]). Black residents are less likely to receive an influenza vaccine by a ~20% margin than white residents, but after adjusting for racial composition, this effect becomes nonsignificant.
								Black residents are also more likely to be in larger capacity NHs, for-profit NHs, and in facilities in metropolitan areas. For every one-point increase in percentage of black residents in each NH, vaccination odds decrease by 1.3%.
Troyer and McAuley (2006)	Cross-sectional, 1996	National (United States)	Black, white NHs residents	n = 2,665	n = 730	Non-standardized model	Documentation of advanced directives	White NH residents are more than twice as likely to have a documented advanced directive compared to black residents with a 37% gap between them. Almost half of the gap could be attributed to county-facility characteristics with black residents more likely to be in facilities in metropolitan areas (vs rural) or in counties with high rates of poverty than white residents. Residents in facilities of those characteristics were predicted to be less likely to have advanced directives.
Rahman and Foster (2015)	Cross-sectional, 2004–2005, (2002: Medicare enrollment), (2000: Census data)	National (United States)	Black, white Medicare free-for-service beneficiaries entered NH for SNF care as new admits post-hospitalization	n = 646,494	N/A	Non-standardized model	Facility quality measures (number of nursing staff / bed)	Blacks from mostly white neighborhoods travel longer distances to NHs with a higher percentage of black residents and lower care quality than whites from the same neighborhoods; “strong evidence of race-based sorting,” (Rahman & Foster, 2015) yet parameters for distance and quality preference are similar across race.

Notes: NH = Nursing homes; SNF = skilled nursing facility; MSA = Metropolitan Statistical Area; DI = Dissimilarity Index; N/A = Not available.

*Facilities are Nursing homes unless otherwise noted.