Staffing Patterns in US Nursing Homes During COVID-19 Outbreaks

Abstract

Importance: Staff absences and departures at nursing homes may put residents at risk and present operational challenges.

Objective: To quantify changes in nursing home facility staffing during and after a severe COVID-19 outbreak.

Design setting and participants: In this cohort study, daily staffing payroll data were used to construct weekly measures of facility staffing, absences, departures, and use of overtime and contract staff among US nursing homes experiencing a severe COVID-19 outbreak that started between June 14, 2020, and January 1, 2021. Facility outbreaks were identified using COVID-19 case data. An event-study design with facility and week fixed effects was used to investigate the association of severe outbreaks with staffing measures.

Exposures: Weeks since the beginning of a severe COVID-19 outbreak (4 weeks prior to 16 weeks after).

Main outcomes and measures: Total weekly staffing hours, staff counts, staff absences, departures, new hires, overtime and contract staff hours measured for all nursing staff and separately by staff type (registered nurses, licensed practical nurses, certified nursing assistants), facility self-reported staff shortages, and resident deaths.

Results: Of the included 2967 nursing homes experiencing severe COVID-19 outbreaks, severe outbreaks were associated with a statistically significant drop in nursing staffing levels owing to elevated absences and departures. Four weeks after an outbreak's start, around when average new cases peaked, staffing hours were 2.6% (95% CI, 2.1%-3.2%) of the mean below preoutbreak levels, despite facilities taking substantial measures to bolster staffing through increased hiring and the use of contract staff and overtime. Because these measures were mostly temporary, staffing declined further in later weeks; 16 weeks after an outbreak's start, staffing hours were 5.5% (95% CI, 4.5%-6.5%) of the mean below preoutbreak levels. Staffing declines were greatest among certified nursing assistants, primarily owing to smaller increases in new hires of this staff type compared with licensed practical nurses and registered nurses.

Conclusions and relevance: In this cohort study of nursing homes experiencing severe COVID-19 outbreaks, facilities experienced considerable staffing challenges during and after outbreaks. These results suggest the need for policy action to ensure facilities' abilities to maintain adequate staffing levels during and after infectious disease outbreaks.

Figure 1.. Nursing Home Staffing Measures and Resident Census During a Severe COVID-19 Outbreak

Coefficients and 95% CIs (shaded areas) are shown from facility-week regressions for an analysis sample of 2967 facilities experiencing a severe COVID-19 outbreak that started between June 14, 2020, and January 1, 2021. There are 456 029 facility-weeks between January 1, 2017, and March 31, 2021, included. Primary independent variables are event-time indicator variables relative to the outbreak start. All regressions also contain facility and week fixed effects. Dependent variables are facility COVID-19 cases reported in the National Healthcare Safety Network data (A), total hours worked by nursing staff at a given facility in a given week from the Centers for Medicare & Medicaid Services’ Payroll Based Journal data (B), resident census in a given week from the National Healthcare Safety Network data (C), and nursing staff hours per resident (ie, the measure from panel B divided by the measure from panel C) (D).

Figure 2.. Additional Nursing Staff Absences, Departures, and New Hires During a Severe COVID-19 Outbreak and Association With Weekly Nursing Staffing Levels

Coefficients and 95% CIs (shaded areas) are shown from facility-week regressions for an analysis sample of 2967 facilities experiencing a severe COVID-19 outbreak that started between June 14, 2020, and January 1, 2021. There are 456 029 facility-weeks between January 1, 2017, and March 31, 2021, included. Primary independent variables are event-time indicator variables relative to the outbreak start. All regressions also contain facility and week fixed effects. A, Dependent variables represent counts of nursing employees who were absent, departed, or hired in each week, scaled by each facility’s average weekly staffing level across all weeks. B, The dependent variable for the dashed line is the weekly nursing staffing level, the blue area uses current absences, the green area uses cumulative (past 12 weeks) departures, and the orange area uses cumulative (past 12 weeks) new hires, again all scaled by the average weekly staffing level. The right y-axis of each panel scales these effect sizes by the mean weekly staffing level across all facilities that experienced a severe outbreak.

Figure 3.. Use of Contract and Overtime Hours During a Severe COVID-19 Outbreak and Association With Weekly Nursing Home Staffing Hours

Coefficient estimates and 95% CIs (shaded areas) are shown from facility-week regressions for an analysis sample of 2967 facilities experiencing a severe COVID-19 outbreak that started between June 14, 2020, and January 1, 2021. There are 456 029 facility-weeks between January 1, 2017, and March 31, 2021, that are included. Primary independent variables are event-time indicator variables relative to the outbreak start. All regressions also contain facility and week fixed effects. Dependent variables are weekly hours worked by nursing staff, scaled by each facility’s average across all weeks. The dependent variable for the dashed line is total weekly hours worked. The blue area uses regular-time hours (ie, hours worked by noncontract employees up to 40 hours per week), the orange area uses overtime hours (ie, hours worked by noncontract employees that are more than 40 hours in a week), and the gray area uses all hours worked by contract workers. Together, these areas sum to the dashed line shown. The right y-axis multiplies the effect size by the mean average nursing home weekly hours across all facilities that experienced a large outbreak.

Figure 4.. Summary of Average Weekly Change in Nursing Home Staffing Counts and Hours During a Severe COVID-19 Outbreak by Staff Type

Coefficients and 95% CIs (error bars) are shown from facility-week regressions for an analysis sample of 2967 facilities experiencing a severe COVID-19 outbreak that started between June 14, 2020, and January 1, 2021. There are 456 029 facility-weeks between January 1, 2017, and March 31, 2021, included. Primary independent variables are event-time indicator variables relative to the outbreak start. Each bar represents the sum of these coefficients across the first 16 weeks after an outbreak’s start for a different dependent variable. A, The dependent variables are total staffing counts and hours for each of 3 staff types: registered nurses (RNs), licensed practical nurses (LPNs), and certified nursing assistants (CNAs). B, Dependent variables are absences, departures, and new hires for each staff type. C, Dependent variables are regular-time hours, overtime hours, and contract hours. All dependent variables are expressed as a percentage of the facility’s average staffing level (or hours worked) for each staff type. All regressions also contain facility and week fixed effects.

Figure 5.. Reported Nursing Home Staff Shortages and Resident Deaths During a Severe COVID-19 Outbreak

Coefficients and 95% CIs (shaded areas) are shown from facility-week regressions for an analysis sample of 2967 facilities experiencing a severe COVID-19 outbreak that started between June 14, 2020, and January 1, 2021. There are 456 029 facility-weeks between January 1, 2017, and March 31, 2021, included. Primary independent variables are event-time indicator variables relative to the outbreak start. All regressions also contain facility and week fixed effects. A, The dependent variable is an indicator for a facility reporting a shortage of nursing staff and/or aides in that week. B, The dependent variables are facility resident COVID-19–related deaths and resident non-COVID-19–related deaths in that week, as reported to the National Healthcare Safety Network.