The truth about; smear tests

Would you like peace of mind for the price of a pair of shoes or a new hairdo? This is the promise offered by a new cervical smear test being marketed in Britain, with the claim that it is more accurate than that used by the National Health Service. The test, called Papnet, was developed in the US and can be obtained privately, and through some GPs, at a cost of about pounds 60-pounds 70. It is already being bought by about 1,500 British women a month - a sign, perhaps, that reports of faulty smear results are undermining many women's confidence in the NHS. (One of the largest such cases was at Kent and Canterbury Hospital, where 1,800 women were wrongly given the all-clear; the hospital is being sued by 15 women who had to have hysterectomies and by the families of three who died of cervical cancer.)

The Papnet test, it is claimed, has a 30 per cent greater sensitivity to abnormal cervical cells than traditional methods of detection using a microscope. Some health service specialists, however, say it has yet to be properly evaluated and have warned that it will undermine the NHS screening programme.

So who is right?

Undoubtedly the computerised testing of smears, which uses a "neural network" technology to detect possible abnormalities, has some potential benefits. For a cytologist, it cuts out the utter dreariness of having to scan some 300,000 cells on each slide, a task that has been likened to searching for a needle in a haystack and which needs phenomenal concentration levels. If this part of the job disappears, it is likely that some degree of human error can be eliminated, too. Another advantage is that Papnet is not totally automated: having scanned all the cells on each slide, it selects 128 with possible abnormalities for examination under the microscope. So while eliminating the drudgery that can lead to error, it keeps the human skill necessary for a final diagnosis.

Studies of Papnet in the US and Australia suggest that it can detect higher numbers of abnormal smears than manual screening, thereby reducing the number of "false negatives" (which in the NHS are currently running at 30 per cent). It is certainly being taken seriously by researchers here: a major trial comparing it with manual screening, involving smears from some 20,000 women, is now being co-ordinated by specialists at the Imperial College School of Medicine and St Mary's Hospital, London. Results should be completed by May 1988.

None of this, however, means that Papnet is infallible. While some faulty smear results are due to laboratory errors, others (about half) are caused by inadequate smears being taken by GPs and practice nurses (in one notorious case, a tongue depressor was used instead of a spatula; in another, the GP used a finger). Unless the smear is taken correctly, abnormalities will be missed, no matter what system of screening is used.

In the NHS, meanwhile, there have been some major improvements in standards of accuracy and methods of recall. Despite all the horror stories about blunders, the programme is thought to prevent some 2,000 cases of cervical cancer each year, with rates falling from 16 per 100,000 smears to 11.2 in the last decade. A 30 per cent false negative rate sounds terrifying, but since only 5-10 per cent of smear results are expected to be positive, what this actually means is that out of 5 million smears taken each year,well over 90 per cent of women get the correct result.

Any test that improves detection rates is good news, but cervical smear testing, whether computerised or not, is never going to be foolproof. It is a bit like wearing a seat belt: in an accident it may save your life, but then again, it may not. As one eminent NHS specialist puts it, the Papnet test may provide a stronger seat belt - one which goes over the shoulder rather than just across the lap. Whether the NHS will ever be able to afford it is uncertain.

Another test is currently being developed for human papilloma virus (HPV), which triggers the cell changes that cause cervical cancer. This test can detect infection of cervical cells with HPV and will be able to help doctors pinpoint which women with mild abnormalities need treatment. Some 250,000 women are diagnosed with mild cell changes each year, but until recently doctors had no way of knowing which of these were harmless: hence all the anxiety about "false positives". The new test will show who is at risk of developing cancer.

With or without computerised testing, the best chance of avoiding cervical cancer is to have a smear taken regularly. Government guidelines specify one every five years, but most specialists would recommend one every three. Women whose GPs operate a five-year interval can obtain a test privately, for a cost of about pounds 40-pounds 50.