on the horizon

Screening for disease is a great idea in principle, opening the way to more effective and less costly treatment. In practice, the route to this happy outcome is no easy ride.

Take bowel cancer. A few families suffer from inherited disorders that predispose them to it - and doctors have devised effective protocols for checking and, where necessary, treating affected people.

It is also known that anyone with two first-degree relatives who have had bowel cancer, or with one such relative below the age of 45, has a five-fold increased risk of developing the disease. Are they, too, candidates for screening? Yes, according to various international guidelines that recommend regular colonoscopy for such people.

But what impact does this make on the detection of early bowel tumours among the population at large? Less than you might hope.

A group at Cambridge University has carried out its own screening project on more than 30,000 people living in East Anglia. Their figures confirm the generally accepted prevalence of the disease, but their interpretation is decidedly downbeat. In the latest Journal of Medical Screening (8: 69-72), Professor Kay-Tee Khaw and her colleagues point out that 'restricting screening to people with a family history will detect only 15 per cent of all prevalent colorectal cancers, as most (around 85 per cent) will not have a family history'. Schemes of this kind will benefit a few, but leave most of the problem untouched.

Testing for blood in stools is more suitable for whole populations - but here cost-effectiveness is open to debate.