Chief Investigator: Paul Meredith, Senior Research Fellow, University of Southampton

Team:

Christina Saville, Senior Research Fellow, University of Southampton

Chiara Dall’Ora, Associate Professor in Health Workforce, University of Southampton

Zlatko Zlatev, Senior Enterprise Fellow, University of Southampton

Peter Griffiths, Chair in Health Sciences Research, University of Southampton

Ian Dickerson, PPI Representative

Tom Weeks - E Systems Implementation Manager Tom.Weeks@porthosp.nhs.uk

Sue Wierzbicki - Lead Nurse - Workforce Sue.Wierzbicki@porthosp.nhs.uk

Partners: Hampshire and Isle of Wight Integrated Care Board, Hampshire Hospitals NHS Foundation Trust, Portsmouth Hospitals University NHS Trust, Salisbury NHS Foundation Trust.

Start: 1 October 2024

End: 30 September 2025

Our aim

We aim to show that a computer algorithm we have developed which uses information that is already collected about patients can provide good estimates of the number of nurses needed on hospital wards to provide safe care for the patients.

Background information

It is important to have enough nurses to care for patients on hospital wards. If there are too few nurses, patients may take longer to recover, suffer complications, or die, and the capacity of the hospital to cope with new admissions is reduced. Also staff well-being is affected by high workloads and there is more staff sickness. Many hospitals use the Safer Nursing Care Tool (SNCT) to help them manage staffing levels. This involves surveying all the patients in a ward perhaps three times a day. Assessing each patient in this way is an extra nursing task and in itself adds to the workload.

We have developed a computer algorithm using data from one hospital which can provide similar estimates of nursing staff requirements to SNCT but we need to check that these estimates would keep patients safe if they were followed.

What we will do

We shall use data collected for a previous study as input to the algorithm to produce estimates of nurse staffing requirements for each ward shift using information which could be known at the time. We will compare actual staffing with the algorithm’s estimate to see if there is a deficit or surplus of staff. For each admission we will examine how these deficits and surpluses relate to patient outcomes. We will compare using the algorithm to set a threshold for safe staffing with using the SNCT estimates as a threshold. Our comparisons will include looking at how good the methods are for wards with higher numbers of under-served groups such as the over 75s, those with learning disabilities and those with mental health conditions. We will measure the effect of staff shortfalls on the number of staff sickness absences. The performance of the algorithm will be checked using data from a second hospital in the database.

We shall work with our partners to find out what tools are used to determine nurse and other staffing requirements on a day-to-day basis in community and mental health settings and what data on care requirements and outcomes is electronically recorded.

We will discuss with partners what opportunities, potential benefits and practical considerations there are to implementing a predictive tool of staffing requirements.  

Communicating results

We will write an academic paper, produce an article for the Nursing Times, create a poster for display at conferences, and publicise the results on social media.

Involving the public

We will involve local PPIE group members in evaluating and commenting on the possible uses of a predictive tool to support decisions in the day-to-day management of nurse staffing levels on wards.