Intended for healthcare professionals

Views And Reviews No Holds Barred

Margaret McCartney: Are physician associates just “doctors on the cheap”?

BMJ 2017; 359 doi: (Published 20 November 2017) Cite this as: BMJ 2017;359:j5022
  1. Margaret McCartney, general practitioner
  1. Glasgow
  1. margaret{at}

Jeremy Hunt tweeted on 18 September, “PAs [physician associates] used to be looked down on as ‘drs on the cheap’ but now widely welcomed as reducing clinician stress & burnout.”1 Is he right?

In 2015 the NHS sought to bring 200 physician associates from the US into English primary care and hospital specialties, through the National Physician Associate Expansion Programme.2 The aim was to train new physician associates (PAs) in the UK to work with, and then replace, the US sourced associates. Around 600 PAs now work in the UK, to rise to 3200 in the next three years. The training is a two year postgraduate course, and Hunt has mandated 1000 new posts in primary care.3

The Royal College of Physicians (RCP) has set up a Faculty of Physician Associates. The college describes the posts as “not plugging or filling medical workforce gaps, but rather helping with the redistribution of the medical workload.”

The equivalent position in the US is “physician assistant,” but in the UK “physician associate” has been used in preference. This, says the RCP, is “to enable the profession to proceed towards statutory regulation and to distance PAs from another category of practitioner (still referred to as physician assistants) who work as technicians rather than clinicians, without a PA’s approved education and training.”4

This difference in terminology is confusing, as are the descriptions of what PAs can and can’t do. The RCP says that “PAs have the requisite knowledge and skill to prescribe” but can’t do so because they have no statutory regulation.

And this is the rub: despite plans to do so it is, as yet, an unregulated profession. It’s far easier to know when to prescribe than to know when not to. PAs can’t work unsupervised, and they can’t prescribe or request tests involving ionising radiation. The RCP thinks that PAs should have a mix of session types: “To ensure continued interest and long term job satisfaction a PA would ideally be involved in the entire scope of GP practice. PAs should be allowed to see any patient who presents, with their supervising GP assisting or intervening as required.”[4]

It is, as yet, an unregulated profession. It’s far easier to know when to prescribe than to know when not to

The RCP also says that PAs should be involved in service design, act as clinical placement leads for students, undertake minor operations, and take part in education and quality improvement projects. No wonder many junior doctors in secondary care—paid less for taking more responsibility and doing more unsocial hours—are concerned about this and the threat to their training opportunities. I’m sceptical that having more PAs join the health service will cut levels of burnout among junior doctors in secondary care.

The RCP adds that, in primary care, PAs’ inability to prescribe is overcome because “many PAs working in general practice have the ability to quickly interrupt their supervising GP for a signature.” I find that alarming. Taking responsibility for prescribing for a patient we’ve not seen or spoken to, whose notes we may not have reviewed in full, requires a lot of trust and means accepting a huge amount of risk. And being interrupted can easily become a safety issue: who else has forgotten to write down crucial information when our attention is pulled in several directions at once? General practice done right is difficult. Forgive me for my old fashioned view that GPs are best placed to see patients.

In primary care the evidence for PAs reducing workload is uncertain5; the longer time taken to see people, who are likely to have minor problems, negates any savings; and research has not considered the cost of GP interruptions and supervision or looked at physician stress or burnout.6 So, no: Jeremy Hunt isn’t right.



View Abstract