Intended for healthcare professionals


Deaths from cervical cancer began falling before screening programmes were established

BMJ 1997; 315 doi: (Published 11 October 1997) Cite this as: BMJ 1997;315:953

This article has a correction. Please see:

  1. A E Raffle, Consultant in public health medicinea
  1. a Avon Health Authority, Bristol BS2 8EE

    Editor—The contrasting viewpoints in the letters from C Mary Anderson and Amanda Herbert illustrate the dilemma of cervical screening1 2: abnormal results are common, but the disease we are trying to prevent was rare before screening.3 Do the high rates of detection exist because we are averting a new epidemic of 7000 cases,2 or even 7000 deaths,4 each year in England? Or do the vast majority of lesions represent an early and reversible stage in the pathogenesis of cancer?

    Screening has been available in Britain since the mid-1960s, but it was not until call and recall programmes were widely established in the late 1980s that appreciable numbers of women aged over 35 were screened. The “epidemic” used to explain the high treatment rates must therefore be confined to women born since 1930. The 1 shows the number of deaths from cervical cancer since 1975 by year of birth in five year cohorts. These data do not give precise rates of deaths from cervical cancer. They do, however, show the actual number of deaths. A fall in the number of deaths from cervical cancer occurred in each successive cohort this century. This trend predated screening. (Data from 1950, available from the Office for National Statistics, show no real alteration to the 1.)


    Number of deaths from cancer of the cervix (ICD 180) in 1975, 1980, 1985, 1990, and 1995 in England and Wales in five year birth cohorts. Deaths among women aged over 80 are not included in the figure (Source: Office for National Statistics)

    Deaths in cohorts born after 1930 flatten off at an earlier age; this gives a strong indication of the effect of screening.5 An implausibly drastic change in cervical cancer has to be postulated if an additional 7000 deaths, or 7000 cases of life threatening disease, are being averted each year. Such an epidemic would have been manifest as a steep rise in death rates in health districts that were slower to organise screening programmes and in those groups of women who chose not to be screened. There is no evidence for this. It may be that deaths from cervical cancer in women born since 1930 are being halved by screening. If so, then the number of deaths in 1995 would have been 2000 in the absence of screening instead of 1339 (which would represent a doubling of the observed 661 deaths in women born since 1930). This saving of hundreds of lives comes at a high cost. Department of Health statistics show that in England and Wales each year about 800 000 women have abnormal smears; 166 000 of these abnormalities are severe enough to warrant referral for investigation and treatment. A tiny minority of these women are actually helped.

    The NHS cervical screening programme is among the best in the world, and 30 years' experience has shown much about the inherent complexities and limitations of early detection as an approach to preventing cancer. It is easy to make simplistic claims about screening. In the long term we will do more for the public by being honest.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    View Abstract