Dose‐response association between nurse staffing and patient outcomes following major cancer surgeries using a nationwide inpatient database in Japan

Abstract

To examine the non-linear dose-response associations between nurse staffing levels and patient outcomes using a nationwide inpatient database in Japan. Previous studies showed that higher nurse staffing levels were associated with better patient outcomes. However, it remains unclear whether there are thresholds for the associations between higher nurse staffing levels and improved patient outcomes. Retrospective observational study design following the STROBE guideline. We identified all patients aged ≥20years who underwent one of six major cancer surgeries between July 2010 and March 2018 using data from the Diagnosis Procedure Combination database, a nationwide database for acute-care inpatients in Japan. Restricted cubic spline regression analyses, the statistical method that allows non-linear functional form, were performed with several scenarios of cut-off points to examine the dose-response associations between patient-to-nurse ratio per shift and failure to rescue, 30-day in-hospital mortality and postoperative complications. Among 645,687 patients, restricted cubic spline regression analyses showed insignificant associations of patient-to-nurse ratio with failure to rescue and 30-day in-hospital mortality with no threshold, but a reverse J-shaped association with postoperative complications with a threshold of patient-to-nurse ratio per shift of 5.4. In terms of postoperative complications, additional registered nurses were associated with decreased postoperative complications. However, this incremental benefit of additional registered nurses may disappear if hospitals allocate five to six number of registered nurses in general wards. This study suggested that additional registered nurses over one per five to six patients may not bring the incremental benefit to decrease postoperative complications.