Relying on gut instinct to read the symptoms

Doctors must listen to a patient’s own views

MEDICAL MATTERS:ONE OF the great challenges in medicine, especially in general practice, is to unravel the mesh of symptoms expressed by people when they come to see a doctor. Each symptom is important to that individual's story, but patients rightly expect doctors to pick up on particular complaints that may presage serious illness.

Many symptoms, for example, could be due to cancer; in practice most of them are not. Family doctors don’t see enough people with individual cancers – typically a GP will see just seven new cases of cancer each year – while specialists see too few healthy patients, so that each group is immediately handicapped by an inherent bias when it comes to weighing up symptoms.

Family doctors traditionally use previous experience and “gut instinct” to form questions thought to have a predictive value in ruling out cancer, but really this is more the art of medicine rather than science.

Health systems have laid down ground rules under which people with “red flag” symptoms must be able to access specialist diagnostic services within a specified timeframe. However, there is little evidence that adopting this common sense approach actually results in a more rapid diagnosis of cancers.

So I was interested to read a paper in the current issue of the British Journal of General Practicethat set out to identify the predictive value for cancer of symptoms, signs and certain test results of patients routinely consulting a GP. Researchers from Keele University carried out a major review of the scientific literature and worked out which signs and symptoms carried a positive predictive value of 5 per cent or more.

They identified clinical features that pose a 5 per cent absolute risk of cancer, and that therefore merit further investigation. They are: bleeding from the rectum, a change in bowel habit, iron deficiency anaemia, blood in the urine, a breast lump, vaginal bleeding in a post-menopausal woman, and an abnormal prostate gland found on rectal examination in men.

But no sooner are these listed than important caveats emerge: age is a hugely important discriminator; iron deficiency anaemia in a 20-year-old woman is extremely unlikely to be due to bowel cancer, whereas the same finding in a 60-something male is likely to mean he has colorectal cancer.

Previous research at King’s College Hospital, London, found the risk of a cancer diagnosis in the three-year period after a patient presented with blood in the urine was 8 per cent in men and 3.7 per cent for women. Coughing up blood produced figures of 8 per cent and 4.5 per cent respectively and for patients who experienced blood loss from the rectum, the risk of cancer was 2.7 per cent in men and 2.1 per cent in women.

These figures show how few people with worrying symptoms actually have cancer. The bad news is how poor “red flag” symptoms are at predicting the risk of cancer and how many people must undergo invasive investigations to pick up the few with a malignancy.

It also suggests we are putting many people through unnecessary tests. If we investigate aggressively at too low a risk of cancer, are we causing avoidable worry and even harm? It's a growing dilemma, because of the push to ensure quality and consistency of care across the health system. As the Health Information and Quality Authority and others rightly address patient safety issues – as in a document published last week, National Standards for Better, Safer Healthcare– referral guidance and criteria will become the norm. But if this guidance is over-inclusive, it could trigger a raft of legal claims based on unrealistic practice and unjustified expectation.

All of which emphasises the need to continue to individualise a person’s care. Whatever a guideline might say, we must ensure a patient’s own views are listened to: some 60-year-old men will quite reasonably decline further investigation of a moderately raised prostate-specific antigen (PSA) level in the blood.

More quality research into the predictive value of red flag symptoms is clearly needed. In the meantime we can expect to see some tension between proponents of “watchful waiting” and those who advocate “timely testing”.

mhouston@irishtimes.com