After Achilles

Almost 45 years ago John Howie, one of the heroes of academic general practice and Richard Scott’s successor to the world’s first chair in general practice in Edinburgh, published a article in the Journal of the Royal College of General Practitioners entitled Diagnosis — the Achilles heel? ¹ Howie argued that the decision by a GP to prescribe for a set of symptoms frequently preceded the allocation of a diagnostic label to those symptoms. A patient with a cough and sputum might get an antibiotic, and someone with a vague feeling of unease without any obvious cause might receive a benzodiazepine. The diagnosis followed the prescription. Although in many regards these observations reflected the more general state of medicine in the early 1970s, general practice has still not really shaken off its struggle with accurate diagnosis. Notwithstanding the need to ‘marginalise danger’, in other words to identify those patients likely to have more serious disease mandating more intensive assessment and investigation, diagnostic decision making in general practice has floundered among unhelpful phrases such as ‘tolerating uncertainty’, ‘using time as a diagnostic tool’ and ‘letting the diagnosis emerge’, which have sadly passed into our lexicon. At worst, this approach to diagnosis is sloppy and idle, and seems to lie at the other end of the spectrum from the need to make early, accurate diagnoses in patients presenting with even the most vague symptom complexes.

It is time for a real paradigm shift in the approach to diagnosis in general practice. It is time to emerge from the shadows of guesswork, reluctance to investigate and willingness to take chances, and to use evidence, risk assessment tools combined with clinical judgement, and technology to cone down on an accurate diagnosis and, where the evidence is inconclusive, to seek better evidence for diagnostic decisions. Our understanding of the natural history of minor and major illness in general practice, of the significance of so-called ‘alarm symptoms’, of the predictive value of individual and multiple symptoms for specific diagnoses, and the ways in which computerised decision support has the potential to improve diagnostic decision-making, have moved on immeasurably over the past 40 years, and it is high time that they formed part of every diagnostic decision in every consultation.This is particularly important in the field of early cancer diagnosis, where primary care research really has led the way in identifying symptoms and symptom complexes requiring early investigation or intervention, and also highly relevant in the diagnosis of serious, non-malignant disease, including infective, vascular, and inflammatory disorders.

Prompt recognition of sepsis, particularly in children, is a current matter of concern, and Claire Gilham’s timely editorial highlights the scale of the problem and the place of primary care in dealing with it. In counterpoint, Treadwell and McCartney highlight the dangers involved in overdiagnosis and overtreatment. The article by Looijmans-van den Akke and colleagues, from the Netherlands, provides a good clinical example (in this case, asthma) where overdiagnosis, often without using available diagnostic facilities, may be a problem. Hamilton and colleagues article on the symptoms of adult chronic and acute leukaemia before diagnosis emphasises the value of large primary care database analysis in identifying key diagnostic features, while a qualitative study by Horwood and colleagues teases out some of the difficulties faced by primary care clinicians in making diagnoses and prescribing decisions in children with respiratory tract symptoms. Renzi and colleagues look at the unintended consequences of giving an ‘all-clear’ diagnosis in patients with potential cancer symptoms — essential reading and of great importance in understanding the need for ‘safety netting’ — while the difficulties of choosing the right test at the right time are highlighted in Watson and colleagues’ study on the use of inflammatory marker testing in primary care. Finally, Lyratzopoulos and colleagues describe an important study on patient-reported consultations before an eventual diagnosis of a rare cancer; emphasising both the difficulties of making an early diagnosis of a rare disease, and also the importance of pursuing symptoms in order to nail down a plausible explanation.