Intended for healthcare professionals


Safe staffing: this is how many doctors we really need

BMJ 2018; 362 doi: (Published 18 July 2018) Cite this as: BMJ 2018;362:k3136
cropped thumbnail of infographic

Infographic available

Click here for a visual summary of safe medical staffing levels.

  1. Matthew Limb, freelance journalist, London, UK
  1. limb{at}

The Royal College of Physicians has pioneered benchmarks for medical staffing levels in UK hospitals and is urging NHS trusts to check how their wards measure up. Matthew Limb reports

The Royal College of Physicians (RCP) has developed the first recommendations for safe medical staffing levels in UK hospitals.1 It urges NHS trusts to measure their workforce against these “indicative” standards (see visual summary) to guard against shortages that pose risks to patients.

College registrar and president elect, Andrew Goddard, says there is “staggering” variation in the numbers of doctors per bed in the UK but no clear tally of how many doctors, nurses, or other healthcare professionals are needed in the acute medical setting at any point in time.

“Everybody’s been afraid to measure it, because they haven’t known what to compare it to. This is the first time that anybody’s actually put a standard down to say, ‘This is what we think we should have—let’s now measure against it to see what we do have,’” he says.

After Mid Staffs

Five years ago, in his final report on Mid Staffordshire NHS Foundation Trust, the chair of the public inquiry, Robert Francis, recommended that “minimum safe staffing and skill-mix levels should be drawn up by the National Institute for Health and Care Excellence [NICE] and policed by the Care Quality Commission [CQC].”

Since then, attempts to develop such levels have had mixed fortunes and have focused on nursing2; though both Wales and Scotland have introduced safe staffing legislation,3 the RCP work is the first to suggest hard figures.

Guidance on Safe Medical Staffing, published on 13 July,1 sets out the number of hours that different levels of clinicians (including physician associates and advanced nurse practitioners as well as doctors) need to be present in various hospital settings. When practicable, it also estimates the number of posts needed to deliver that type and volume of care safely.

The college says there is much variation in the way that individual medical wards function and the calculations provide indicative rather than absolute staffing numbers. It says staffing calculations should be based on “80% of maximum activity” and that the modelling assumptions take account of the estimated 30-70% of medical time spent on indirect care such as leadership and management.

It points out that the out-of-hours workload of the medical registrar on call is “inappropriately onerous, with implications for patient safety.” It argues that services must always support training and that an increase in consultant delivered care may be limiting opportunities for trainees to acquire experience in decision making.

The recommendations explained

These safe staffing benchmarks are a necessity, not a pipedream, say RCP president Jane Dacre, Andrew Goddard, and cardiologist Rhid Dowdle, who chaired its guidelines working group.

Why has the report been produced?

“We are heading for a real workforce crisis and we need to have national solutions,” says Dacre. The college says the number of doctors coming through the system has “flatlined” because of caps in medical student numbers and caps on visas.

“Patient care is now jeopardised by staff shortages and low morale,” she says. “We recognise that the complex issue of safe medical staffing will not be resolved at a stroke but this guidance is a significant step towards that objective.

“I think it’s a necessity rather than a pipe dream.”

How long has this work taken to complete?

Two years, but the seeds were sown much earlier, says Goddard. In 2010, an RCP snapshot survey indicated how many junior doctors there were on NHS wards. “What became clear was that nobody collects those data and that there’s a huge variation,” Goddard says.

Although NICE has produced guidance on safe nursing levels, it has not had a similar mandate for medical staffing. “We thought that’s a space we have to go into,” says Goddard. “This is ground breaking. No one previously has had any standard to compare to, to understand whether the number of medical professionals they’ve got working in different wards within the hospital is right.”

What was the working party asked to do?

The 23 strong group of experts was asked to define benchmarks for safe staffing levels for different clinical situations (see visual summary). Dowdle says this meant analysing workload and working patterns to come up with a set of recommendations. “We started to examine what people do, how they should do it, how long it took and therefore how many people were needed.

“We tried to be sensible; we didn’t document best practice but what we thought was safe care,” says Dowdle.

What were the methods?

Focusing on four areas of hospital activity, the experts calculated how much time was needed to provide care from clinicians, whom they grouped into three tiers according to levels of knowledge, experience, and responsibility (see visual summary).

The working group says it is no longer appropriate to speak of work being done by specific grades of doctors, and the model reflects how the medical workforce has been evolving and “broadening out.” Dacre says, “What’s happened over the past few years, quite appropriately, is that other clinical staff have been able to pick up jobs that traditionally might have been done by doctors. One of those groups is physician associates. [The RCP hosts the Faculty of Physician Associates].

“We’re all clinicians and we work in teams and are codependent, and that’s because medicine is more sophisticated and complicated. So, how we reach those numbers can be quite flexible.”

Goddard says the tiered classification of medical professionals aims to make the analysis “future proof … So whatever the future looks like in how care is delivered in hospital this is the measure of what is safe and effective care.”

Can you give examples of recommended safe staffing levels?

Goddard explains this by describing a patient with pneumonia being admitted to hospital arriving first in the acute medical unit (AMU). He says it will take an average tier 1 clinician an hour and quarter of their time to do all the tests, take the blood samples and history, do an examination, request radiography, and be on hand to arrange any further investigations. “So that’s how we come up with the figure that says if you’ve got 10 patients in that setting you’re going to need 15 hours of tier 1 time.

“And in the report we’ve looked at an average AMU take of 45 patients and say you’ll need six to eight tier 1 doctors to do that—an estimated 13 posts. We’ve tried to say to a trust, this is how many doctors you are going to need to employ to provide this number of hours.

“You’ve got to remember that’s not necessarily the number of doctors you’re going to have on the ward or the unit at that particular time, but to cover that number of patients in that setting, given that some doctors will either be off shift or taking time off in lieu or for sick leave, annual leave, study leave, paternity leave, etc.

“One of the reasons we feel like we’re understaffed at the moment is because we haven’t taken those calculations into account.”

How many hospitals are likely to be meeting the new standards?

“We don’t know,” stresses Goddard. “It may be there are some places that have reached this standard easily and others may be a long way behind it. But until we actually start to measure against this standard it’s impossible to say.

“This is very much about setting a bar, so then we can measure against that bar and then start to talk about what is achievable [and] what’s not achievable.”

“One of the biggest arguments for not looking at this is that it’s too complicated and too difficult. We’ve got to stop that and try to see where we are.”

How long will it take to know how trusts are measuring up?

“I’d be disappointed if we didn’t have a reasonable idea after a year,” says Goddard. “There are 180 odd trusts in England, so we need to start in a selection of those.

“We hope people use this as a yardstick and over the next year or two begin to work with the Care Quality Commission and NHS Improvement and hospitals and healthcare systems to look at how many staff there are; then we can begin to think about how we help places where they are understaffed.”

How should trusts use the guidance?

Dowdle says, “We’ve produced a template which trusts can use if they want to assess where they are with their staffing.

“We suggest if your trust has concerns about your staffing, or if there are adverse events being reported, these can be used to trigger an audit. From that, they can use our indicative numbers to work out roughly what their workforce (levels) should be.

“The key thing about this document is it’s not the definitive product; it isn’t the end of the process—it’s the start. Our recommendations must be validated in action by people doing audits.”ow could regulators use the benchmarks?

Goddard says he believes the CQC would see it as a “positive” step that hospitals were looking at them. But could they consider taking action if trusts aren’t measuring up? “We’re nowhere near that point, because we don’t know where everybody is,” he says.

What might the cost implications be?

Goddard says that if trusts were found to be terribly understaffed this would affect costs. But he says alleviating staff shortages mitigates other costs. “Let’s not forget we spent £3bn last year on locums.

“The NHS already spends a lot of money trying to fill rota gaps and shortages and a lot of money trying to sort out problems when something’s gone wrong with patient care that could be prevented by having adequate staffing levels.”


  • Competing interests: I have read and understood BMJ policy on declaration of interests and have no relevant interests to declare.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract

Log in

Log in through your institution


* For online subscription