The best known example is screening for breast cancer by X-ray mammography. This has proved highly effective in detecting small cancers, less than 1cm in diameter, which are mostly curable by surgery. Yet in countries such as the United States that have pioneered breast screening, the numbers of women found to have breast cancer have been rising steadily by around 1 per cent a year while the numbers dying have hardly changed. In the past 10 years the number of cancers detected has risen from 2.7 to 7.6 per 1,000 women screened as techniques have improved, and the proportion of very early cancers is now 30 per cent. How many of these would have gone on to become life-threatening, obvious tumours is unknown.
Research studies have shown that regular mammographic screening does save lives in women aged over 50, cutting mortality from breast cancer by around 15 per cent. No clear benefit has been shown in younger women, however - though in the United States women under 50 account for half the total number screened. The thought-provoking statistic cited earlier this year in a review in the New England Journal of Medicine is that a meticulous postmortem examination of the breasts of women aged 40-50 who had died from other causes found that no fewer than 39 per cent had microscopic evidence of breast cancer.
The problem beginning to emerge with screening has to do with the expectations of the doctors and their patients. Many years ago the criteria for a useful screening test were agreed. It should detect a common disease; treatment should have been shown to be most effective when given at the earliest possible stage; and the test should be safe and reliable, with few false positives (people alarmed unnecessarily) and few false negatives (people given false reassurance).
What no one had expected was the huge iceberg of unsuspected disease that sensitive tests would detect. For example, only one in 1,000 people aged 50-70 develops symptoms due to thyroid cancer, yet careful examination of the gland shows that around one-third of people have a small, symptomless cancer. The proportion of elderly men with microscopic cancers of the prostate is near 100 per cent.
Cancer is not the only disease with this iceberg effect. Ultrasound screening of men aged 60-75 shows that around 10 per cent have aneurysmal swellings of the lower aorta. Some of these small aneurysms would increase in size and may rupture, but most will probably not cause symptoms at all.
Once the condition is found it is difficult for doctor and patient to decide to watch and wait. If it is a cancer that has been detected, waiting may seem unacceptable, yet this may be the course of action most likely to prolong life. Answering dilemmas of this kind makes research an essential feature of health care.Reuse content