Evaluation of the effect of a Nursing System Framework on Nurse Sensitive Indicators, mortality and readmission in an NHS Trust.

The study was observational and could not assign causality for any observed difference in mortality or readmissions to the NSF. Only one year after NSF introduction was analysed because of the pandemic, not the originally planned two years. The pre- and post-NSF cohorts differed in age, gender and deprivation, requiring adjustment. No comorbidity variables were included and end-of-life patients were not excluded. Analysis was based on admissions rather than individuals, so a patient could appear in both cohorts and multiple admissions could not be identified. The NSF was still being prepared in the unexposed period and may not have been fully implemented just after launch. The statistical models fitted poorly, with pseudo-R-squared values of 0.12 for deaths and 0.008 for readmissions. Other changes, such as new drugs, devices, interventions or personnel, were not measured and could have affected outcomes. Nurse-sensitive indicator results were descriptive and not all audit findings were reported. The NSF contained many components, so the contribution of any single component could not be isolated.

The authors declared that there is no conflict of interest. The data extraction was done by Trust Information team members and research team members who are employees or associates of the Trust.

Do not read the lower percentages in the NSF period as proof the framework improved patient outcomes. In-hospital deaths were 5.0% before and 4.7% after, and emergency readmissions were 12.4% before and 11.9% after, but the study found insufficient evidence of an association. The NSF was a complex framework, and other changes were not measured, so no single component can be credited.