Identifying trends in nurse retention via routinely collected operational data (the NuRS Study)– a retrospective multi-centre exploratory analysis of voluntary turnover across Acute Hospital and Mental Health settings

Background
Internationally there is a shortage of registered healthcare professionals. Previous studies have identified several operational factors that contribute to staff retention, but methods for the proactive identification of these factors are limited. This study explores if currently collected operational data can identify if and how operational factors influence staff retention.
Methods
The study is a secondary analysis of routinely collected operational data from two acute trusts and two mental health trusts operating in the English National Health Service via a knowledge discovery in databases methodology. The data were analysed via binomial regression using the lasso subset selection techniques and dimensional reduction via sparse principal component analysis.
Results
In the acute setting, voluntary resignation levels were found to be higher following periods of non-voluntary resignation (e.g. retirement and end of fixed term contract) though the effect was reduced if wards experienced increased levels of staff sickness and higher acuity patients. The analysis also found an increase in voluntary turnover was associated with above average usage of agency staffing, and below average use of bank staffing. In the mental health setting, an association was found between increased voluntary turnover and above average levels of historical sick leave.
Conclusions
This study demonstrates a method for the identification and quantification of existing pressures on workplace turnover. Due to the reliance on pre-existing data the analysis represents a method for the passive profiling of operational factors that contribute to staff turnover, which could be used to generate localised insight and allow for tailored improvements. Due to the focus on the re-use of existing data the study methodology could be replicated across multiple National Health Service trusts, and embed as a repeatable approach to characterise emerging risks.