Objective
High nursing home staffing instability (ie, daily variation) is associated with poor resident outcomes. This study examined changes in nursing home staffing instability during the COVID-19 pandemic and investigated whether instability patterns differed by nursing home characteristics.

Design
National 2017–2023 nursing home data were used. We used staffing data from the Payroll Based Journal (PBJ), nursing home characteristic data from the LTCFocus, and Centers for Medicare and Medicaid Services (CMS) Provider Data Catalog datasets.

Setting and Participants
A total of 13,219 nursing homes reporting PBJ data.

Methods
We estimated mixed effects models predicting staffing instability (percent of days 20% below average) with nursing home random effects. The dependent variable was nursing home staffing instability. Fixed effects included % Black residents, % Hispanic residents, % Medicaid payer residents, ownership type, chain status, and staffing level.

Results
Staffing instability increased during the COVID-19 pandemic and has since declined but has not returned to pre-pandemic levels. Nursing homes with the highest proportion of Black residents had higher levels of staffing instability, with a larger increase in staffing instability during COVID-19, compared with nursing homes with the fewest Black residents. Nursing homes with a high proportion of Hispanic residents had lower instability. Government-owned nursing homes had higher levels of instability than either for-profit or nonprofit nursing homes, and chain nursing homes had lower instability.

Conclusion and Implications
Previous studies have examined the impacts of the COVID-19 pandemic on the average level of nursing home staffing. This study adds to that body of work by demonstrating an increase in staffing instability as well, and one that varies according to certain nursing home characteristics. As these metrics are associated with different aspects of resident outcomes, policies being developed to prepare for future pandemics should consider mechanisms and plans that will address both.