Helen Salisbury: Risk and responsibility when working with physician associates
BMJ 2024; 385 doi: https://doi.org/10.1136/bmj.q1270 (Published 11 June 2024) Cite this as: BMJ 2024;385:q1270
One of the issues troubling doctors in general, and GPs in particular, is the question of whether they’re responsible—and indeed medicolegally liable—if a physician associate (PA) under their supervision makes a mistake. Other members of our multidisciplinary teams have their own professional registration, but PAs currently do not. There’s been much discussion of one case where a doctor in an emergency department was sanctioned by the Medical Practitioners Tribunal Service for failing to go back to take a collateral history and perform his own examination of a patient already seen by a PA.1
Last week the General Medical Council’s chief executive, Charlie Massey, wrote to NHS leaders to offer reassurance that “doctors are not accountable for the decisions and actions of PAs and AAs [anaesthesia associates] who they supervise.”2 Although this will come as a relief, there was a rider: “doctors will not be held accountable provided they have delegated responsibility to them in line with the standards and guidance in Good Medical Practice,3 Leadership and Management for All Doctors,4 and Delegation and Referral.5”
This last document, Delegation and Referral, advises that you “must be confident that the colleague you delegate to has the necessary knowledge, skills, and training to carry out the task, or that they will be adequately supervised to ensure safe care.”5 So, how do we gain that confidence? When I’m supervising specialist registrars I know that they have their own professional registration, but more importantly I know what knowledge they should have by now—what experience they’ve had and what exams they passed to get to their current position. Even with this secure knowledge, they’re closely observed when they first join us and debriefed after every surgery until they finish their training.
Assessing competence
I don’t know all the details of the training of other members of the multidisciplinary team, but I’m not delegating to them. The nurses are doing nursing, the physio is doing physiotherapy, and in both cases they’re fulfilling their own professional role.
However, if PAs are employed, as stipulated, under the Additional Roles Reimbursement Scheme, they’ll be seeing undifferentiated, undiagnosed patients. As it takes an average of 10 years for a doctor to acquire the skills to do this, my baseline assumption will have to be that a PA, after a two year course, doesn’t have the necessary knowledge and skills to do what NHS England says they should be doing. The current qualifying exam—which has a 100% pass rate in some institutions6—doesn’t give us an assurance of competence, so we’ll have to make that assessment ourselves. Unlike registrars, physios, and nurses, competence isn’t guaranteed by the qualification gained nor the role assigned and will vary hugely between individuals.
The rules set out by the GMC might be just about workable if everyone operated in stable teams and there was time for each GP or consultant supervising a PA to fully assess their knowledge and observe their skills. The reality of many workplaces is that different doctors supervise PAs each day, and the existence of locum PA agencies makes a nonsense of the idea that all doctors could know the abilities of the people they’re delegating to in this role.
The only option, if we’re to keep within GMC regulations, is to supervise to the hilt, debrief thoroughly after every patient encounter, and repeat important physical examinations. Otherwise, despite Massey’s attempt to reassure us, we may well be at risk of the medicolegal consequences of unsafe delegation or inadequate supervision. This will take a lot of GP time, but the alternative is that our patients may be at risk of missed diagnoses, and their safety should always be our first concern.
Footnotes
Competing interests: See www.bmj.com/about-bmj/freelance-contributors
Provenance and peer review: Commissioned; not externally peer reviewed.