Photo: to sign from Reuben Strayer flickr photostream

Currently over 4 million people are on NHS waiting lists, hospital trusts’ deficits are at record levels, and the patient promise of treatment within 18 weeks of referral has been abandoned. For NHS England, the incentive to “do something” about GP referrals is clear. Clinical peer review, the analysis of clinicians’ behaviour by fellow clinicians, is a widely-adopted quality improvement and educational tool which NHS England now proposes to use to focus on GP referrals. It argues that reducing variation in GP referrals will improve the quality of referrals, benefit patients who suffer harm when they are referred unnecessarily, and reduce pressure on secondary care services. Panels of reviewers are being established and the management of some schemes has even been out-sourced to the private sector.

Models of retrospective peer review, traditionally built upon consent, supported learning, and agreed local priorities have been shown to be successful in influencing levels of GP referrals. GPs already engage in clinical peer review activities and participate in reflective practice to improve their performance, as required by the GMC Code, including work on monitoring their colleagues’ referrals.  The move to a compulsory, prospective, process predicated upon a set of centrally-determined aims, however, is the result of a  political choice to focus on GP performance, rather than looking at the wider population health problems which cause referrals, or more complex NHS system issues across secondary and primary care, or even the interests of the patient.

The processes involved in a decision to refer is a complex mix of GP confidence in their own  skills, training, and judgement; local service availability, patients’ wishes and circumstances. Multiple factors impact upon referral rates, but evidence shows that success in influencing referrals is crucially dependent upon the type of referral scheme implemented. The King’s Fund report that the most consistently successful approaches were  retrospective (after the patient had been referred and treated); supported by strong clinical engagement; and integrated clear referral criteria with evidence-based guidelines. Recognising that demand for secondary care cannot be managed in primary care alone, the best schemes looked at all sources of referral.  In contrast, less successful models tended to be centrally-led, prospective schemes, such as the NHS England initiative. Two major criticisms are that they interrupt and delay the patient journey, and focus only on general practice. The approach chosen by NHS England therefore runs against the tide of evidence about what works in referral management. The proposed scheme will have harmful effects upon the GP/patient relationship, and may not even be successful in its explicit aim of reducing referrals.   Irrespective of the motivation for referral management, there are also a wide range of possible consequences (both intended and unintended), and with some positive and some negative implications.

Research provides less than positive messages about the impact of centrally-led referral management schemes which intercept referrals. The GP/patient relationship is based on trust and confidence – centrally-led processes which intercede into that relationship and potentially over-ride GP opinion erode that relationship. Evolving patient attitudes toward their GP may deter them from consulting their GP, or encourage them to turn to alternative providers. Patients will rightly question their GP if and when their referral has been turned down.  This may be no bad thing – this could encourage GP engagement in the process – but, as Professor Martin Marshall, Vice Chair of the RCGP, said in a recent statement “our concern is that these schemes can undermine the important trust that exists between GP and patient.”

Positively, all referral management schemes, whether initiated locally or centrally mandated, are predicated on the idea that peer review opinion is more valid than that of the ordinary GP. All other things being equal, faced with the same patient, GPs with specialist training may make more accurate diagnostic decisions than GPs without specialist training.  But, of course, all things are never equal. GPs reviewing referrals cannot know the whole patient without access to their notes, knowledge of their background, needs, or personal circumstances. Considering referral through a purely, mechanistic, biomedical lens ignores the holistic psychodynamics of the doctor/patient relationship, described some 60 years ago by Michael Balint.

Somewhat positively, referral management could also potentially swap the bias of an individual GP for an alternative set of cultural and other organisational and team views held by reviewers. However, this raises a number of issues. How will reviewers monitor and assure their own best practice and demonstrate their own impartiality?  Are reviews undertaken “blind”?  The recent example of continuing care panels, where allegations of bullying and disregard of medical opinion have been made, offers a cautionary tale about what occurs when organisational agendas and professional opinions collide in a purportedly neutral, objective environment. Could future research into referral management test the hypothesis that “better” decisions are made by reviewers? Given the infinite variability implicit within the art of diagnosis, this is unlikely. Referral management as proposed by NHS England, therefore, appears to be little more than an irrelevant and misguided performance measure of general practice, suiting neither the interests of GPs or patients, aiming at arbitrary targets whilst missing the point.

Negatively, the harm caused to patients by unnecessary referral has been well-documented – referral management systems have the very real potential to risk patient safety, as referral review builds in delays for patients. If harm reduction is an aim of referral management, then a focus should be placed on GPs who under-refer, and whose patients miss out on appropriate, timely care. For GPs, paradoxically, there is evidence that pressure to review referral behaviour could be demotivating, reducing their willingness to tolerate uncertainty and manage problems in primary care, and increasing their levels of referrals to secondary care. Referral management systems established with the aim of improving quality, rather than performance managing general practice, would be setting acceptable levels of variation, risk, and quality within GP referrals. This could be a more positive approach, but these sorts of concerns are underplayed in the current NHS England proposals.

The identification of GP referrals as a policy priority is a political choice.  An alternative political reading is that the “problem” of GP referrals is a problem caused by capacity issues within the secondary sector; that referrals are directed to hospital because of the lack of community and primary care alternatives; and that GP decision-making would be improved if general practice was funded to increase consultation length.  There is a clear need for to be much more attentive to the wider repercussions of the focus on referrals: on GPs, patients, and the wider system, beyond any impact it may have on GP referral rates.

About the Author: Adrian Mercer Ph.d,  is a former primary care trust chief executive, NHS policy researcher, and occasional lecturer in health and social care policy. Based in Devon, he is currently researching and writing about NHS developments, including privatisation, democracy in the NHS, and digital politics. He is on twitter @adeindevon