Social Service Staffing and Quality of Care Outcomes in US Nursing Homes

Examine the association between social service staffing intensity and resident quality of care outcomes in US nursing homes.

National retrospective panel study linking facility-level staffing and quality data from 2021-2023.

A total of 38,493 facility-years representing 13,766 unique Medicare- and Medicaid-certified nursing homes nationwide.

Social service staffing (qualified social workers + other social service staff) was calculated from Payroll-Based Journal data as minutes per resident day. Outcomes were the quality measure star rating and 4 claims-based indicators: long-stay emergency department (ED) visits and hospitalizations per 1000 resident days, as well as short-stay outpatient ED visits and 30-day rehospitalizations. Random-effects ordinal logistic, Poisson, and linear regressions were conducted controlling for resident acuity, nursing home- and county-level characteristics, and state and year fixed effects.

Facilities averaged 5.84 ± 3.97 social service minutes per resident day. Each additional minute was associated with the odds of a higher quality star rating (odds ratio, 1.03; P < .001); a 1 standard deviation increase (∼4 min) raised the probability of earning a 5-star rating by 1.4 percentage points. For long-stay residents, each additional minute corresponded to 0.2% fewer ED visits [incidence rate ratio (IRR), 0.998; P = .011] and 0.6% fewer hospitalizations (IRR, 0.994; P < .001). Associations were nonsignificant for short-stay outcomes.

Modest increases in social service staffing are linked to reductions in long-stay acute transfers and higher quality ratings. Regulators and payers should consider expanding staffing standards and public reporting for social service personnel. Administrators could improve quality at relatively low cost by adding social workers or other social service staff. Future research should examine causal pathways, optimal staffing composition, and cost-effectiveness across resident subgroups.