Ways of Knowing

‘Ovarian cancer isn’t a silent cancer — it is a cancer that whispers’. This comment encapsulates both the challenges and the opportunities for making accurate diagnoses of conditions which are infrequently seen in general practice, or often present with non-specific, unremarkable symptoms, or both. Vague abdominal pelvic or systemic symptoms are difficult to evaluate, but if they persist or recur, they may be early warnings of something more serious — they whisper.

Diagnosis is the theme of this issue of the BJGP. The study by Charles Helsper and colleagues from the Netherlands addresses the problem of identifying women with a history of ovarian cancer in order to carry out genetic testing on them and their relatives to determine their future cancer risk related, for example, to carrying BRCA1/2 mutations. They appear to have established the acceptability and likely effectiveness of two strategies, one with and the other without information communications technology support.

The problems of making an accurate diagnosis in polymyalgia rheumatica (PMR), pneumonia, and liver disease are described in three further research papers. Toby Helliwell and colleagues from Keele University discuss the diagnosis of PMR with GPs, and find variation and inconsistency in the use of investigations and guidelines, and concerns about corticosteroid therapy which, they point out, are unsurprising because of the lack of research on PMR conducted in primary care settings.

Mark Ebell and colleagues, reporting from the universities of Georgia, US and Lausanne, Switzerland use case vignettes of patients with cough and a range of other symptoms to study clinicians’ estimates of the likelihood of community-acquired pneumonia (CAP). They found that the likelihood of CAP was generally overestimated, and that treatment thresholds were variable across their sample, suggesting that clinical decision rules may in future have a role to pay in providing guidance on testing, diagnosis, and treatment.

Liver disease has, it seems, been something of a closed book for many GPs, and the interpretation of liver function tests an arcane pursuit. Helen Jarvis and colleagues at Northumbria and Newcastle universities found that their sample of GPs did, indeed, have some gaps in knowledge and lacked confidence about diagnosis and management of liver disease, particularly non-alcoholic fatty liver disease. This is an important shortcoming given the rising tide of chronic liver conditions, and needs to be addressed through a range of educational initiatives to increase awareness and confidence.

Given the current intense professional and media focus on the use of new technologies in medicine, it is reasonable to ask whether technology such as artificial intelligence might have a useful role in these diagnostic scenarios, and the answer is almost certainly yes for all four. Electronic health records linked across the primary–secondary care interface, including family tree/genogram data in the patient record, should be extremely useful in identifying patients at increased risk by virtue of clinical history, family history, and persona demographic data. Computer-generated prompts of diagnostic probabilities and nudges toward appropriate investigations, displayed in real time as the computer ‘listens in’ to the consultation, must have the potential to sharpen the precision of diagnosis of PMR and CAP, and also mitigate against inappropriate prescribing. The interpretation of abnormal liver function tests in the context of the history and examination, and the generation of diagnostic probabilities and prompts for further action must be child’s play for any interested robot. I’m not sure what we are waiting for.

Finally, this year’s RCGP Research Paper of the Year award is reported in Life & Times. Two BJGP papers were category winners. One was a very timely study on the barriers and facilitators that face GPs seeking treatment for distress.¹ The other was the widely-reported and practice-changing study highlighting the significance of thrombocytosis as an important predictor of malignancy.² Great stuff!

• © British Journal of General Practice 2018