In the US, approximately 27% of deaths due to coronavirus disease 2019 (COVID-19) have occurred among residents of nursing homes (NHs).1 However, why some facilities have been more successful at limiting the spread of infection than others is unclear. For example, those with greater staffing or higher performance on quality measures may be better at containing the spread of COVID-19 among staff and residents.
We evaluated whether NHs rated highly by the Centers for Medicare & Medicaid Services (CMS) across 3 unique domains—health inspections, quality measures, and nurse staffing—had lower COVID-19 cases than facilities with lower ratings.
We used data from 8 state health departments (California, Connecticut, Florida, Illinois, Maryland, Massachusetts, New Jersey, and Pennsylvania) to determine the total number of COVID-19 cases occurring in NHs between January 1, 2020, and June 30, 2020. We linked these data with CMS Nursing Home Compare, which includes star ratings (range, 1 [low] to 5 [high]) that characterize performance across the 3 domains.2 The health inspection rating is based on the number of deficiencies identified in the 3 most recent state surveys across several areas, including staff-resident interactions and adequate infection control protocols. The quality measures rating is based on the weighted mean of performance across 15 quality measures (eg, avoidable hospitalizations, pressure ulcers, urinary tract infections). The nurse staffing domain is based on the mean staffing hours per resident by qualified nursing staff.
Given how COVID-19 data are publicly reported across some states, we were limited to grouping NHs into 3 categories: those with 10 or fewer, 11 to 30, or more than 30 COVID-19 cases. We performed 3 separate ordinal logistic regression models to assess the odds of high-performing facilities (4- or 5-star facilities) having more than 30 cases vs 11 to 30 cases vs 10 cases or fewer relative to low-performing facilities (1- to 3-star facilities), adjusting for the number of certified beds and including county fixed effects. The study was conducted using SAS version 9.4 (SAS Institute Inc). Two-sided P values were considered significant at the P < .05 level. The Harvard T. H. Chan School of Public Health Institutional Review Board waived the need for informed consent.
Of the 4254 NHs across the 8 states, 4254 (100%) had star ratings for health inspection; 4241 (99.7%), quality measures; and 4225 (99.3%), nurse staffing domains. Within each domain, 1451 (34.1%) were considered high performing for health inspection; 2974 (70.1%) for quality measures; and 1517 (35.9%) for nurse staffing (Table 1). High-performing NHs were less likely to have had more than 30 COVID-19 cases than were low-performing facilities across each domain (health inspections, 348 [24.0%] vs 948 [33.8%]; quality measures, 897 [30.2%] vs 397 [31.3%]; nurse staffing, 382 [25.2%] vs 907 [33.5%]). High-performing NHs had a lower median number of certified beds. After adjustment, NHs with high ratings on nurse staffing were less likely to have more than 30 COVID-19 cases vs facilities with 11 to 30 and vs facilities with 10 or fewer cases than were low-performing NHs (OR, 0.82; 95% CI, 0.70-0.95; P = .01) (Table 2). There was no significant association between high- vs low-performing NHs in the health inspections or quality measures domains with COVID-19 cases.
Across 8 states, high-performing NHs for nurse staffing had fewer COVID-19 cases than low-performing NHs. In contrast, there was no significant difference in the burden of COVID-19 cases between high- vs low-performing NHs for health inspection or quality measure ratings. These findings suggest that poorly resourced NHs with nurse staffing shortages may be more susceptible to the spread of COVID-19.3,4 Although guidance on best practices on infection control are important, which has been the primary strategy used by CMS to date, policies that provide immediate staffing support may be more effective at mitigating the spread of COVID-19.5,6
This study has limitations. It included data from only 8 states; however, these states rank among those with the highest COVID-19 burden. The state-reported data used are also more reliable than the national COVID-19 data set recently released by CMS, which reports suggest is incomplete and inaccurate. In addition, high-performing NHs may have greater capacity to test and diagnose cases, which may lead to an underestimate of the association between low performance on the staffing domain and higher COVID-19 cases.
Corresponding Author: Jose F. Figueroa, MD, MPH, Department of Health Policy and Management, Harvard T. H. Chan School of Public Health, 677 Huntington Ave, Kresge Room 413, Boston, MA 02115 (jfigueroa@hsph.harvard.edu).
Accepted for Publication: July 22, 2020.
Published Online: August 10, 2020. doi:10.1001/jama.2020.14709
Author Contributions: Drs Figueroa and Zheng had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Figueroa, Wadhera, Papanicolas, Jha.
Acquisition, analysis, or interpretation of data: Figueroa, Wadhera, Riley, Zheng, Orav.
Drafting of the manuscript: Figueroa, Riley.
Critical revision of the manuscript for important intellectual content: Figueroa, Wadhera, Papanicolas, Zheng, Orav, Jha.
Statistical analysis: Papanicolas, Zheng, Orav.
Obtained funding: Figueroa.
Administrative, technical, or material support: Figueroa, Riley, Jha.
Supervision: Figueroa, Jha.
Conflict of Interest Disclosures: Dr Figueroa reported receiving research support for other work not related to this topic by the Commonwealth Fund, the Robert Wood Johnson Foundation, and the Harvard Center for AIDS Research. Dr Wadhera reported receiving research grant K23HL148525-1 from the National Heart, Lung, and Blood Institute and previously serving as a consultant for Regeneron. Dr Jha reported receiving funding from the Commonwealth Fund, the Robert Wood Johnson Foundation, and the Bill and Melinda Gates Foundation for other work. No other disclosures were reported.