Smear tests were not on trial but should have been

BMJ 1999; 318 doi: https://doi.org/10.1136/bmj.318.7189.1007 (Published 10 April 1999) Cite this as: BMJ 1999;318:1007
  1. C Mary Anderson, General practitioner.
  1. Heaton Moor Medical Centre, Stockport SK4 4NX

    EDITOR—In February three women were declared victims of medical negligence because their smear abnormalities were missed at the Kent and Canterbury Hospital.1

    This raises some serious questions. Did the judge understand how reliably the smear test performs under day to day conditions—that for every 90 000 slides reviewed several hundred will be reallocated a new category? Did he know that borderline cases and inadequate samples are not included in quality assurance tests in the United Kingdom because they cannot be agreed on? The rate of interobserver agreement even between experts is rarely better than 80% and is lower for milder abnormalities. Consent forms and disclaimers in the United States say that 5-15% of abnormal smears may be deemed normal. The guidelines of the College of American Pathologists for reviewing cervical smears in the context of litigation or potential litigation state that a false negative smear is not necessarily evidence of practice below standard care, that borderline smears represent a poorly defined diagnosis with poor interobserver and intraobserver reproducibility, and that smears should be reviewed without knowledge of clinical outcome and in an environment that simulates the normal screening practice as closely as possible.2

    Was this last condition met in this case? If not, how can the Bolam test have been applied? Furthermore, the three women argued that if their smear abnormalities had been picked up in time cancer would have been prevented. But this cannot be assumed: some women die of cervical cancer even when everything has been done properly in detection and treatment.

    In this case the screening programme itself seems to have been on trial and robustly defended. In his summing up, the judge urged women not to be afraid of having a smear test, adding that the national programme had been a great success in reducing the number of deaths from cervical cancer, which was declining at the rate of 7% each year. He did not say that only part of this reduction is due to screening: the rest is because cervical cancer is spontaneously declining in the United Kingdom. General practitioners have been urged to tell women that the smear test is not foolproof and that they should report abnormal bleeding to their doctor.3 After 30 years of screening, why has this message not got through and why was it not repeated here? A landmark opportunity for promoting honesty about the limitations of screening has been missed. The consequences for my patients, especially young ones, will be increased overdiagnosis and unnecessary treatments.


    • Competing interests Promoting greater honesty about the limitations of the smear test may conflict with my financial interest (target payments in general practice).


    View Abstract

    Sign in

    Log in through your institution

    Free trial

    Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
    Sign up for a free trial