Administrator Turnover in Oregon Assisted Living and Residential Care Communities, March 2020-February 2021.

The impact of the COVID-19 pandemic on residents and staff of assisted living and residential care (AL/RC) facilities has been sizable, yet little is known about how the pandemic affected their administrators and their job stability. Considering the crucial role that administrators play in the delivery of high-quality care and staffing in residential long-term settings,3, 4, 5, 6 this research letter describes the turnover experience among Oregon AL/RC administrators and its organizational and structural correlates during the first year of the COVID-19 pandemic (March 2020–February 2021).

Administrative records were retrieved from Oregon Department of Human Services (the licensing agency) for 549 AL/RC facilities licensed and operated during the study period (see Supplemental Material for details about AL/RC licensing in Oregon). The dependent variable was a community-level binary indicator that measured whether the administrator as of March 1, 2020, had left their position by February 28, 2021 (stayers = 0, leavers = 1). Organizational characteristics included in the analysis were the number of licensed beds, whether the AL/RC facility was endorsed for memory care (MC) (0 = non-MC, 1 = MC), whether the AL/RC facility had a contract to serve Medicaid residents (0 = non-Medicaid, 1 = Medicaid), nonprofit status (0 = for-profit, 1 = nonprofit), and the tenure of administrator as of March 1, 2020, in years. Structural characteristics included rural location (0 = urban, 1 = rural) and the number of AL/RC facilities licensed in the county (see Supplemental Material for detailed variable descriptions).

AL/RC facilities in the study had an average of 52 licensed beds; 38% were endorsed to provide memory care and 77% had a contract to serve Medicaid residents (Supplementary Table S1). Almost 10% were nonprofit, and 40% were in rural areas. At the beginning of the pandemic, in March 2020, the average administrator tenure was 2.3 years. Among one-third (36%) of AL/RC facilities, the administrator employed at the beginning of the pandemic left the position within the year.

Figure 1 shows odds ratios and 95% CIs of leaving the job during the study period from a logistic regression model that included all available organizational and structural characteristics. Each 100 beds and nonprofit status both halved the odds of the administrator employed at the beginning of the pandemic leaving that position. Each year of prepandemic tenure decreased the odds by 29%. None of the other covariates were significantly associated with likelihood of an administrator change during the study period.

Fig. 1

Exponentiated coefficients (odds ratios) and 95% CIs from a logistic regression model that includes all covariates listed (N = 549). AL/RC, assisted living and residential care. ∗P < .05, †P < .01, ‡P < .001.

The high rate of administrator turnover and its accompanying correlates (ie, size, nonprofit status, and prepandemic tenure) have implications for interventions aimed at improving quality of care in this setting as well as indirect, negative effects on quality of care. For instance, administrator turnover can lead to loss of institutional support and memory, an integral part of sustainable quality improvement interventions. Furthermore, high management turnover can result in a workforce environment unsupportive of direct care workers, leading to high turnover among the latter. Finally, differential administrator turnover can exacerbate the existing inequalities between groups of AL/RCs (eg, small vs large; nonprofit vs for-profit).

This study has several limitations. First, it is observational and hence cannot rule out unobserved heterogeneity or reverse causality. Second, because of data limitations, the administrators who left their position and those who were fired were combined in the outcome measure, potentially confounding some of the observed associations. Finally, although Oregon has a regulatory requirement that AL/RCs report changes to administrators, and state surveyors can cite for violations of this requirement, data accuracy was not independently confirmed. However, to address this issue and allow more time for any late reports of administrator changes, the study focused on an earlier part of the pandemic.