It is rare for a study to generate excitable headlines when its main conclusion is that more research is needed, but that is what happened last week when the British Medical Journal published a “rapid systematic review” of how physician associates (PAs) fare in the NHS. The authors looked at the academic literature and, after employing some strenuous exclusion criteria, concluded that it is “sparse and of variable quality”. Only six studies looking at the clinical performance of PAs met their exacting standards.
This was greeted by the Telegraph with the headline “No evidence it’s safe for physician associates to work in NHS, says BMJ review”. The Independent went with the slightly more reserved headline: “Almost no evidence that employing physician associates in NHS is safe, say Oxford researchers”. The messaging on social media was, as ever, more hysterical. The incorrigible bandwagon-jumper Carol Vordermann tweeted: “Physician associates — not safe to work in NHS says Oxford Uni study” while the inexplicably popular left-wing X account “Blade of the Sun” tweeted: “A new review by Oxford University says that Physician Associates (PA’s) are neither needed nor safe. It’s time to get rid of this scam.”
If this seems an over-reaction, it was one encouraged by the study’s authors, both of whom happen to be members of Independent Sage: Trisha Greenhalgh and Martin McKee. Some readers may fondly recall Professor Greenhalgh encouraging the public to wear panty liners over their faces to protect themselves from Covid-19 while some vapers will know McKee, who is at London School of Hygiene & Tropical Medicine, as one of the Britain’s most passionate and least informed opponents of e-cigarettes. Greenhalgh works at Oxford University’s Nuffield Department of Primary Care Health Sciences at that, apparently, is enough for their paper to be called an “Oxford University study” and for them both to be described as “Oxford researchers”.
In the abstract of their study (but not in the study itself) they claim that “No evidence was found that physician associates add value in primary care”. In their conclusion, they say: “Conflating absence of evidence of safety incidents in a small number of research studies with absence of safety concerns when physician associates directly substitute for doctors is an error of logic that is likely to cost lives.” Greenhalgh repeated this on the Today programme, saying that “to conclude that they’re safe from the basis of the evidence we’ve got is an error of reasoning that is likely to cost lives.” The BBC then wrote the story up under the headline “NHS associates role could ‘cost lives’, study says”.
Physician associates … have been the target of an extraordinarily vicious campaign of abuse
It is sometimes true that the absence of evidence is not evidence of absence, but it can equally be fatuous mid-wittery. Researchers who test a hypothesis and fail to confirm it are not dealing with the absence of evidence, rather they have found evidence that the hypothesis is wrong. So it is with physician associates who, as I described last year, have been the target of an extraordinarily vicious campaign of abuse from certain medics who are trying to protect their wallets. Greenlaugh and McKee’s study gives them ammunition, but its content does not match its conclusion.
One of the few studies that meet Greenlaugh and McKee’s inclusion criteria was published in BMJ Open in 2019. It found that: “PAs were found to be acceptable, appropriate and safe members of the medical/surgical teams by the majority of the doctors, managers and nurses.” Awkwardly for the locum doctors who are waging war on PAs, it also found that: “Consultants and managers considered locum doctors that were new to their service as less efficient, less safe and costlier than PAs.” Moreover, it noted that a systematic review had already concluded that PAs are safe. That review, published in 2014, said that “physician assistants (PAs) are cost-effective substitutes for physicians, with similar outcomes in primary care and surgery.” Greenlaugh and McKee ignore that review because it was published before 2015 (!) and they implicitly dismiss the BMJ Open study because identifying safety concerns was not an explicit part of the study’s design.
A study comparing PAs to junior doctors, published in BMJ Open in 2020, concluded that: “Physician associates in emergency departments in England treated patients with a range of conditions safely, and at a similar level to foundation year two doctors-in-training, providing clinical operational efficiencies.” Greenlaugh and McKee cite this study, but put it to one side because its “primary outcome measure (reattendance at the same healthcare facility with the same complaint within seven days) was not designed to detect or examine safety critical decisions”.
In general practice, a study published in PLoS One in 2016 used experienced doctors to assess consultations by PAs and GPs and found that “experienced GPs could not differentiate PAs from GPs, and importantly they considered that PAs provide safe consultations”. Greenlaugh and McKee also cite a study published in the British Journal of General Practice comparing physician associates to GPs which found no significant difference in re-consultation rates (a proxy for outcomes) or patient satisfaction. It concluded that PA consultations “result in similar outcomes and processes for similar consultations by GPs at a lower consultation cost”. Greenlaugh and McKee downplay these findings because the PAs tended to deal with less complex health needs (which is the whole point of them) and because the research was conducted in 2011-12 (“before the current workload crisis in general practice”).
One can argue that research of this kind is not perfect and could be improved, but claiming that there is no evidence at all just isn’t true, is it? And that is without looking at all the international evidence on physician associates elsewhere in the world which Greenlaugh and McKee appear to have ignored completely. Even a brief reading of the literature shows that the claim that there is “no evidence” that physician associates “add value in primary care” is a lie. Their value has been confirmed by patients, managers and independent assessors. Greenlaugh and McKee assert in the abstract (but, again, not in the study itself) that PAs “seemed to struggle in primary care”, but insofar as the evidence they cite supports this claim, it is considerably weaker than the evidence that PAs are “safe”.
If there is an “error of logic” here, it is the demand that those who support physician associates prove a negative. Where is the evidence that employing more PAs “is likely to cost lives”? There is none to be found in this study but Greenlaugh and McKee say it anyway because the absence of evidence is no problem if you’re on their side of the debate.
