Safe nurse staffing for the NHS: what does the evidence say?

Author Liliana Cunha
Posted 2018.12.05 Comments 0

Safe staffing remains a major concern for patients, practitioners and policy makers, even more so in the face of unprecedented workforce shortages, with the number of vacant posts doubling over the last three years.


At HSRUK and the Health Foundation’s recent evidence and policy roundtable on safe staffing on 21 November, Professor Peter Griffiths, Chair of Health Services Research, University of Southampton explained what research tells us about safe staffing, from the right skills mix to how tools influence planning and outcomes. His guest blog post for HSRUK looks in more detail at the evidence around safe staffing numbers. 

Read Professor Griffiths’ new paper “Nurse staffing, nursing assistants and hospital mortality: retrospective longitudinal cohort study” in BMJ Quality and Safety.


Although nurses are the largest professional group working in the health services, the value of their contribution is more contested than most. While few doubt that doctors are essential to safe and effective healthcare, nurse teams are often considered as part of the ‘hotel services’ at NHS hospitals rather than an essential service line. While there is ample evidence establishing links between nurse staffing and important patient safety outcomes, the significance of this evidence is often challenged; and it is not always clear how it can be used to inform decisions about staffing levels, which can also have huge economic implication under conditions of substantial uncertainty.

I recently spoke on the subject of safe nurse staffing at a roundtable run by Health Services UK (HSRUK) and the Health Foundation in London, to an audience of policymakers, hospital directors of nursing and workforce, health and social care workforce representatives and fellow academics. This blog post delves further into the evidence around staffing levels. There is much more to safe staffing than numbers and no single magic number to be found. Plus we simply do not have enough registered nurses, with shortages in the UK both longstanding and getting worse. But we have to start somewhere. Having the right number and mix of staff is a necessary (if not sufficient) prerequisite for delivering care that is safe, meets the standards of care and compassion that professionals aspire to and that all of us would expect for our family, friends and indeed ourselves.


Don’t we know all this already?

A link between nurse staffing and patient safety was recognised well before the events that led to the inquiries into poor care and patient outcomes at the Mid Staffordshire NHS Trust from 2005-2008. Reading the Mid Staffordshire inquiry reports makes it clear that this link was simply not recognised by many in the Trust, or if it was, senior managers felt able or compelled to ignore it. One of the positive outcomes of the inquiries is that safety and nurse staffing has become inextricably linked in the discourse of the NHS. So much so, that sometimes the production of new evidence in this field evokes cynicism – simply dismissed as statements of the obvious. A recent report of one of my own conference presentations linking low nurse staffing and poor quality care was greeted with a wry riposte “In other news, Pope comes out as Catholic”; and my personal favourite “no shit Sherlock…”, in the comments sections of the Nursing Times. 

So what does the evidence say exactly?

As a result of the Francis inquiry, the Department of Health commissioned NICE to develop guidance on safe staffing. NICE in turn commissioned our team at Southampton to review the evidence. There are hundreds of studies exploring associations between nurse staffing, skill mix and outcomes. Our review concluded that higher registered nursing (RN) staffing levels were associated with lower hospital related mortality. Higher RN staffing was also associated with fewer falls, shorter hospital stays and nurses reporting less of “missed nursing care”. For pressure ulcers and drug errors there was a mixed picture, although it is likely this arises from problems with risk adjustment for the outcomes. Lower staffing levels were also linked to worse staff outcomes, with increased burnout and intention to leave.

The evidence also supported an association between better outcomes and a skill mix that is richer in RNs. We concluded that higher levels of healthcare assistant staffing did not improve patient safety outcomes – and might actually make things worse. In some studies, higher assistant staffing levels were associated with higher rates of falls, pressure ulcers, readmission rates, medication errors, use of physical restraints and lower patient satisfaction.

Yet to the surprise of many, NICE concluded that There is a lack of high-quality studies exploring and quantifying the relationship between registered nurse and healthcare assistant staffing levels and skill mix and any outcomes”.

The problem is clearly not the number of studies, nor is it their size. Several studies included hundreds of hospitals and millions of patients. But the studies were observational and most were cross sectional, linking average staffing with patient outcomes over a period of time (typically a year), generally at a hospital level, although one important US study used a longitudinal approach and showed that individual patient experiences of low staffing preceded poor outcomes. There was essentially no economic evidence from the UK. In simple terms this evidence was the wrong type of evidence for NICE to use to make concrete recommendations about staffing levels.


That was then, this is now….

Evidence is starting to emerge that addresses a number of important limitations and uncertainties that arose from the NICE reviews. Our own NIHR funded study just published in BMJ Quality & Safety shows the effects of individual patient exposure to variation in nurse staffing levels, confirming existing findings but using a longitudinal design, with data collected over three years at a large NHS trust across 32 wards and 138,000 patients. Economic modelling based on the results suggests that increases in RN staffing or the skill mix of the nursing team might prove to be cost effective, consistent with conclusions from studies in other countries including the USA. While increases in RN staffing are associated with increased staff costs, in some scenarios these increases were offset by savings from reduced hospital stays. Increases in numbers of support staff above current levels were associated with increased mortality and longer hospital stays.


What is the solution?

You might reasonably argue that we need to invest more in the community, not hospitals. I agree that we need to invest more in community services, but these economic findings suggest that investment in better nurse staffing in hospitals, particularly a richer skill mix, might be a way of reducing demand for inpatient beds by making hospitals more efficient.

There is great faith in some sectors in the potential for new technology to increase the productivity of health services and potentially reduce the need for health workers. The brave new world of care robots is not yet with us to any great extent, but I am not aware of any evidence of their safety, ability to substitute for staff or their cost effectiveness. Computerised care records and other electronic documentation have been around for some time, yet research evidence is decidedly mixed. It may well be that technology generate some efficiency savings, but technology in healthcare is notoriously expensive and there is little evidence to suggest it reduces the need for professionally qualified staff rather than the support services it often replaces.

The NHS has also invested in ‘evidence-based’ staffing tools to give a more efficient allocation of nursing staff.  But evidence amounts to little beyond demonstration that they provide some measure of nursing workload because they are associated with some other measure of nursing workload. There is essentially no evidence of the costs or benefits from staffing according to the tools prescription.

Let’s also look at another proposed solution to the registered nurse shortage in England, the nursing associate: a new grade of registered practitioner with less training than a RN. Of course, we don’t have any direct evidence, because nursing associates are too new. But we know that substituting lesser-qualified staff for RNs is associated with worse outcomes. Although there are savings on training and staff pay, total costs may increase due to additional resource use. Although the risk of adverse outcomes might be reduced by providing more staff training, potential cost savings are lower because training and salary costs are now higher. If, on the other hand, nursing associates don’t substitute for RNs but instead simply increase the skills in the support workforce then – while increased costs seem certain – improved outcomes are possible.


And finally

There is so much more to safe staffing than numbers. Leadership, positive work environments and cultures are all vitally important, but hard to maintain in the face of staff shortages.

There has been little serious and sustained commitment to meeting the demand for professionally qualified nurses in UK policy. Alternative routes into nursing for those who lack the qualifications for a degree course are being widely championed, yet the replacement of bursaries with student loans has created a prohibitive obstacle for well-qualified candidates who seek to enter nursing at a later stage yet have student debt from a first degree.

There may be residual uncertainties in the evidence, but that is not a reason to pursue alternatives that are supported by no evidence at all. The solution to a shortage of registered nurses is to recruit, train and retain more registered nurses. Safe nurse staffing may well come at substantial cost, but this should not divert us from the goal of a full complement of professionally qualified registered nurses on our hospital wards.

Whichever way you look at it there is risk, yet all the evidence points toward needing larger and better educated nursing workforce. Only a net upskilling of the health workforce presents any realistic prospect of improved outcomes.


© Peter Griffiths 2018

Published under a creative commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) license.