|Home | About | Journals | Submit | Contact Us | Français|
I agree with William Hamilton that strategies to detect cancer should aim for evidence-based efficiency, and that more of the evidence should come from community-based research. His commentary is a thoughtful discussion of ways of improving cancer diagnosis in the context of current strategies of the British NHS.1 But the scaffolding of the argument puzzled me: I recognise none of the five misconceptions that he argues are impeding progress in improving cancer diagnostic services. And so, in the spirit of friendly debate, I offer my contrarian perceptions of Hamilton's misconceptions.
This doesn't mean much to me as a concept — the familiar pathway to cancer diagnosis that Hamilton describes looks like a reasonably rational team-based approach to achieving the answers the patient needs. When a patient comes to me with symptoms that might represent cancer, I use the resources at my disposal as a GP, the specialist those at theirs, and together we, hopefully, rule cancer in or out, preferably according to a shared guideline. Hamilton argues that the perception that secondary care is responsible for diagnosis leads to GPs using the wrong data when making referral decisions. This may well be the case, but, as he points out, the positive predictive values of most symptoms of cancer are lower in primary care than in referred populations. Guidelines based on predictive values derived in community populations will, if anything, increase the threshold for referral. Primary care research may show just how difficult it is for clinicians to decide which patients with diarrhoea to refer, but is unlikely to bolster the argument for expanding access to investigation for the low but not no-risk patients that Hamilton describes later in his discussion.
Who, other than media journalists, thinks this? It is certainly not our hospital colleagues. I regularly run joint medical student teaching sessions with consultant colleagues where we work through cases from first primary care presentation to final diagnosis and treatment. To a man and woman, these specialists acknowledge the difficulties facing patients and GPs in making rational decisions about referral in low risk populations. There is always room for better education, but I can't accept the argument that taking seriously the symptom of haemoptysis means that I will then miss lung cancer by ignoring patients with cough. It is possible for clinicians to hold more than one thing in their mind at one time, and as Hamilton has himself shown, the integrative judgement of the GP is often the most powerful factor in making correct decisions about cancer risk.
Where is the evidence? Screening has a role in cancer diagnosis, but GPs' involvement with cancer diagnosis continues to be mainly in patients with symptoms that worry them. I don't accept the argument that the existence of mammography and cervical screening programmes means I will act with any less diligence in evaluating a woman who comes to me with a breast lump or inter-menstrual bleeding.
For the reasons Hamilton gives, I agree that this statement is incorrect, but I don't know anyone who holds this belief. Surgeons, for example, were influential in liberalising the NICE guidance for investigation of dysphagia, so as not to exclude from investigation patients with curable oesophageal cancer who lacked red flag symptoms.
As Hamilton points out, this is not a widely held perception — almost everyone believes that early diagnosis is a good thing and behaves accordingly. Whether or not treatment improves prognosis, patients and doctors want speedy diagnosis of concerning symptoms to resolve uncertainty and plan treatment. It would be satisfying to have better evidence that we can reduce mortality by swift action, but I do not accept that patients or doctors slow down because they lack such proof.
Commissioned; not peer reviewed.