Abstract

Background: Although nurse staffing is fundamental to nursing home quality, policymakers have generally focused on measuring and reporting only average staffing levels during a quarter or a year. Average staffing measures mask daily variation, which may also influence outcomes, and which could offer additional information about nursing home quality and relative ranking.

Objectives: The objective of this study was to examine daily staffing variation and to test the following hypotheses: 1) High daily staffing variation is negatively associated with quality; 2) Daily staffing variation provides information about quality ranking of nursing homes over and above the information provided by average staffing levels.

Methods: Retrospective analyses of 2017-2018 administrative data on all Medicare- and Medicaid-certified nursing homes in the United States. Data included the Payroll-Based Journal for daily staffing information, merged with Medicare Cost Reports and Nursing Home Care Compare data about other nursing home characteristics. Three measures of daily variation, adjusted for daily resident census, were constructed and calculated for registered nurses (RNs) and certified nurse assistants (CNAs): Coefficient of Variation (COV), Total Outlier Days (TOD), and Low Outlier Days (LOD). Association between the measures and other quality measures was estimated in a regression model. Agreement about ranking nursing homes into quality deciles by the average and the variation measures was assessed by weighted Kappa statistics.

Results: The sample included 13,339 (93.9%) of all nursing homes reporting complete data. Outcome measures varied substantially: COV for RNs averaged 0.5 (S.D.=0.6), 0.1 (S.D.=0.1) for CNAs, Total Outlier Days averaged 220 (S.D.=69) for RNs, 44 (S.D.=45) for CNAs, and Low Outlier Days averaged 115 (S.D.=45) for RNs and 22 (S.D.=24) for CNAs. Regressions revealed that higher daily variation was significantly associated with worse quality as captured by the 5-star quality measures (QMs) ratings: COV for RNs -0.014; COV for CNAs -0.004; Total Outlier Days for RNs -3.79; Total Outlier Days for CNAs -2.52; Low Outlier Days for RNs -2.46; Low Outlier Days for CNAs -1.29. Similarly significant and negative association with the 5-Star Survey quality measures ratings findings were found and will be reported. The Kappa for agreement in decile ranking between average staffing and the variation measures ranged from 0.23 to 0.63, indicating little agreement between the measures, and suggesting that the addition of daily variation measures would change the quality ranking of nursing homes relative to using average staffing alone.

Conclusions: Daily variation in staffing is partially correlated with other quality measures and adds information about quality over and above average staffing levels. These findings demonstrate the importance of measuring and reporting daily variation in staffing to improve understanding of the relationship between staffing and quality. Adding measures of daily staffing variation to the current average staffing measures in Nursing Home Care Compare may enhance its value for providers engaged in quality improvement and consumers searching for high quality nursing homes.

Slides