UCSF

Background: Registered nurses (RNs) take care of patients at the bedside and play a critical role in protecting them from adverse events through monitoring and surveillance. Researchers have examined the effect of nurse staffing on patient outcomes, but studies have produced inconsistent findings. The mixed findings might have resulted from various staffing measurements and insufficient adjustment for the effect of high patient turnover on nursing workload.

Purpose: This dissertation was to refine evidence of the relationship between nurse staffing and patient outcomes while minimizing measurement errors. The specific aims were: (1) to compare various staffing measures and select the best methods; (2) to examine the association of nurse staffing with patient outcomes for surgical care; and (3) to evaluate the effect of patient turnover on the relationship between nurse staffing and patient outcomes.

Methods: Three studies were performed using data obtained from the University HealthSystem Consortium (UHC) for the year 2005. Multivariate regressions were conducted while controlling for patient and hospital characteristics. For the first aim, the Akaike Information Criterion (AIC) was used to determine which nurse staffing measure performed the best. For the second aim, surgical intensive care unit (ICU) and non-ICU staffing levels were included in the regression models to examine the association with surgical patient outcomes. The interaction between patient turnover and RN staffing was examined for the third aim.

Findings: Measuring nurse staffing as productive hours by both direct and indirect patient care providers was best. Higher RN staffing in surgical non-ICUs was related to lower rates of failure to rescue and post-operative sepsis. We found that there was a significant interaction between patient turnover and RN staffing on general units.

Conclusions: The findings show that higher RN staffing contributes to patient safety by decreasing failure to rescue. However, using only surgical staffing did not reveal more different effects on surgical patient outcomes than previous work. While productive hours of direct and indirect providers were better than other measures of nurse staffing, the differences were small. Further work is needed to determine the reasons for the inconsistency in research findings.