Screening for breast and cervical cancer as a common cause for litigationBMJ 2000; 320 doi: https://doi.org/10.1136/bmj.320.7246.1352 (Published 20 May 2000) Cite this as: BMJ 2000;320:1352
A false negative result may be one of an irreducible minimum of errors
A delay in the diagnosis of cancer is now one of the commonest reasons for medical litigation. Increasingly women in whom breast or cervical cancers are diagnosed after a “normal” screening test are alleging negligence through a delay in diagnosis and are seeking compensation through the legal system. Medical staff involved in providing screening are highly concerned about this situation.1 How has it arisen?
The enthusiasm of the health service to promote screening has perhaps given women unrealistic expectations. Women may falsely believe that screening prevents cancer rather than detects it earlier. There is also a perception that cancers arising after a normal screening examination must have been “missed” and that the delays in diagnosis have prognostic significance.
Population screening is different from health care, which manages individuals with symptoms; most people who are screened are free from disease, and an acceptable balance between the sensitivity and specificity of the screening test must take this into account.2 The legal position here seems to differ from that of the health providers. In the case of a number of patients who were given false negative results on cervical screening in east Kent in England, the courts awarded the patients compensation. An appeal by the health authority was dismissed by the appeal court, which ruled that sensitivity in screening is paramount—“a false negative result could have very adverse consequences. A false positive would have nothing like this disadvantage to the patient … the patient could be caused anxiety, but this is a small price to pay for the protection against the adverse consequences.” 3 False positive screening tests, however, cause serious morbidity and anxiety to women several months after a screening recall despite reassurance.4 5 In screening for breast cancer, specificity is no more than 15% with about 5% of women recalled for further tests to diagnose 5–6 cancers per 1000 women screened. Reducing specificity further to improve sensitivity would be unlikely to increase appreciably the detection of early cancers or reduce mortality. It would also be to the detriment of the considerably larger number of women subjected to the anxiety of a recall for a false positive result.6 7
At best, screening mammography has a sensitivity for cancer of around 90%, and in the three year cycle of the NHS breast screening programme around 40% of breast cancers present symptomatically (interval cancers). A small proportion of these interval cancers (around 10%) are so called “false negative” screens in which, in retrospect, the previous mammogram shows abnormalities. An independent expert opinion may well conclude that in these individual cases a detectable abnormality was not identified and that this was negligent. There is, however, an irreducible number of false negative results in any screening programme; the reasons for this are complex.2 3 In cervical screening it is accepted that the proportion of false negative screening tests should be 5% or less. Negligent false negatives on one hand and an irreducible minimum of screening errors on the other are difficult to reconcile.
Are screening programmes doing all they can to ensure that false negative cancers are occurring at the minimum possible rate? The NHS breast screening policy is to have a single reading of mammograms and single view mammography after the first screen, but the evidence is that the best practice is two view mammography at every screen and double reading, with arbitration for discordant opinions. This results in considerably more small breast cancers being detected, with no increase in false positive results.8 Therefore, current NHS breast screening practice may be failing to achieve the irreducible minimum of false negative screens. The breast screening and cervical screening programmes are subject to rigorous quality assurance and quality control, involving external peer review of both the processes and outcomes, to ensure that the standards of care for each woman and the population are acceptable.
So what more can be done? People invited for cancer screening must be told about the risks, benefits, and limitations in a way that instils realistic expectations and ensures fully informed consent in those who participate. The quality assurance systems applied to screening programmes must be rigorous and involve individual peer review to ensure that performance is maintained above the prescribed minimum standards. The corollary of this is that screening must be given enough resources to ensure that standards can be achieved and false negative cases are indeed kept to the irreducible minimum. The legal system must take account of the difference between population screening and diagnoses in individuals based on symptoms and deal with alleged negligence in the context of population screening. Expert panels could be set up to offer advice to courts in these instances.