Intended for healthcare professionals


How will expansion of physician associates affect patient safety?

BMJ 2024; 386 doi: (Published 05 July 2024) Cite this as: BMJ 2024;386:q1377
  1. Rachel Greenley, research fellow,
  2. Martin McKee, professor of European Public Health
  1. London School of Hygiene & Tropical Medicine

Experience with nursing assistants calls for extreme caution, write Rachel Greenley and Martin McKee

Too often what seems like a good idea turns out not to be. The NHS is facing a staffing crisis. It has over 120 000 vacancies and almost three quarters of employees think that their organisation has insufficient staff.1 There are concerns that the NHS is entering a downward spiral as departures increase pressure on those left to unsustainable levels.

Those in charge are taking drastic and highly controversial action.2 Medical school places are expanding, although where the necessary infrastructure, trainers, or clinical placements will come from is unclear.3 New occupations, like physician associates (PAs), are being created. Yet this development has proven highly controversial, for many reasons.4 One is concern about patient safety, with increasing examples of misdiagnosis of patients by physician associates. However, at this point, these are individual cases, so it is unclear whether there is a larger problem. Intuitively, providing additional staff to support teams that are under great pressure must surely be a good idea.

There is, however, an analogous development from which we can learn. The UK, like many other countries, has developed occupations intended to alleviate the work of nurses. Variously called nursing assistants or associates, there are now over 10 000 registered in the UK.5 Unfortunately, the evidence on patient safety is far from reassuring.

A study of the effect of differences in skill mix in six European countries found that each 10% reduction in registered nurses as a proportion of total nursing care personnel was associated with an 11% increase in the odds of patient deaths.6 It did not matter whether the change in proportion of registered nurses was caused by reducing their numbers or adding nursing assistants or other types of non-registered nurses. Moreover, a higher share of registered nurses in the mix was associated with greater patient satisfaction and lower nurse burnout and job dissatisfaction.

Another study, this time of changes in nurse staffing over time in English hospitals, found that additional shifts by registered nurses already on the hospital payroll was associated with a significant decline in in-hospital mortality, with the effect twice as high for the senior nurses than junior nurses.7 However, increased use of agency nurses and nursing assistants achieved no reduction.

A German study found that three dimensions of patient-perceived quality of nursing care significantly declined as nurse staffing decreased and the proportion of assistant nurses in a hospital unit increased.8 These effects were more pronounced for patients whose clinical condition was less complex or were admitted to smaller hospitals.

These findings from Europe are supported by a study from the United States where a 10 percentage-point reduction in registered nurses was associated with a 7% higher odds of in-hospital death, 1% higher odds of readmission, and lower patient satisfaction.9 Another study, from Australia, found that increasing the proportion of time that patients spent on wards where nursing assistants were employed increased the odds of urinary tract infections and pneumonia among patients.10

Meanwhile, a systematic review that included studies from the United States, China, and Australia—as well as Belgium and the United Kingdom, concluded that policies that lead to a reduction in the proportion of registered nurses in nursing teams could produce worse outcomes at increased costs and there is no evidence that such approaches are cost-effective.11

These findings seem counterintuitive. As the authors of the Australian study noted: “Logically, maintaining the skill mix combined with enhanced resources should improve patient outcomes. We hypothesised that the addition of [nursing assistants] to existing ward staffing would decrease adverse patient outcomes due to having another person on the ward to help the regulated nurses.” Yet it did not. They speculated that “delegating basic patient care tasks to [nursing assistants] reduces the opportunity for ongoing monitoring, assessment and evaluation by [registered nurses], and that important cues are not recognised by the [nursing assistants].” It is at least plausible that this will also be the case when substituting doctors with PAs who have much less training. It is argued that their two-year training equates to the clinical part of a medical degree. Yet it is ludicrous to suggest that what comes before that, an undergraduate science degree, which could be in astrophysics or inorganic chemistry, equates to pre-clinical training in anatomy, physiology, and pathology. However, in a world where the chair of NHS England, an accountant with a degree in history, argues that medical training can be “slashed” without citing any evidence, anything is possible.12

No one can be in any doubt that the NHS needs more staff. Yet, given evidence that some of these new occupations, like anaesthetic associates, are not cost effective because of the greater supervisory demands,4 and the consistent evidence that adding nursing associates to the mix threatens patient safety, NHS England and the General Medical Council are taking a huge risk by conducting a large, and unevaluated experiment of patients.


  • Competing interests. MM is Past President of the British Medical Association and was one of the signatories to a letter calling for an extraordinary general meeting of the Royal College of Physicians because of concern about the expansion of physician associates. This paper draws on a review conducted within the project Magnet4Europe: Improving Mental Health and Wellbeing in the Health Care Workplace funded by the European Union’s Horizon 2020 research and innovation programme (Grant Agreement Number 848031) and the National Institute of Nursing Research, National Institutes of Health (R01NR014855 and T32NR007104). The investigation presented here is the responsibility of the authors only. The EU Commission takes no responsibility for any use made of the information set out.

  • Provenance and peer review: not commissioned, not peer reviewed