An official website of the United States government
Here's how you know
Official websites use .gov
A
.gov website belongs to an official
government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you've safely
connected to the .gov website. Share sensitive
information only on official, secure websites.
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with,
the contents by NLM or the National Institutes of Health.
Learn more:
PMC Disclaimer
|
PMC Copyright Notice
Editor—The recent wave of litigation in relation to screening for breast and cervical cancer is consistent with the current “compensation culture” but to some extent reflects a failure of both the public and its legal representatives to appreciate the principles and shortcomings of screening programmes.1
There is an urgent need to educate women and give them accurate information so that informed decision making can occur and any adverse sequelae are better understood. Despite widespread publicity and vehement lobbying by patients' advocacy groups in the United States, surveys have revealed that women have little knowledge or understanding of the basic screening principles: over three quarters believed that the benefit of screening for breast cancer in women in the 40 to 49 year age group had been proved, and almost 40% thought screening was effective in women under 40 years of age, though no published data support such a benefit.2 Concepts of probability and numeracy may be difficult to convey during a clinical consultation even when the patient seems to understand the issues.3
It is perhaps not surprising that the occurrence of interval cancers after a normal screening investigation evokes doubt and confusion, leading women to seek legal and financial recompense. The heterogeneity of breast tumours, with variable mean sojourn times and a range of biological aggressiveness, makes it difficult to estimate the prognostic significance of any delay in diagnosis. However, if a national screening programme is offered to women, certain levels of expectation must be honoured and the screening process rigorously monitored for quality assurance and outcome measures. Data from the NHS breast screening programme indicate that the screening interval is too long: more than 80% of cancers occurring within the screened population in the third year after a screening mammogram are interval cancers.4 Thus almost as many cancers develop between the second and third years as would be expected in an unscreened population. Moreover, interval cancers tend to be more advanced than those detected by screens. False negative rates of up to 40% have been reported in screening programmes, with retrospective review of mammograms showing a radiological lesion in 20% of interval cancers and a suspicious abnormality in a further 20% (S Ciatto, Advanced workshop in breast cancer management, Hong Kong, 1998).
Two view mammography and double reading will help reduce false negative rates, but reducing the time between screens should also help to minimise true interval cancers. These changes will increase the staffing and financial costs of screening but will improve efficacy and in the longer term may avert potential claims of negligence.
References
1.Wilson RM. Screening for breast and cervical cancer as a common cause for litigation. BMJ. 2000;320:1352–1353. doi: 10.1136/bmj.320.7246.1352. . (20 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Woloshin S, Schwartz LM, Byram SJ, Sox HC, Fischoff B, Welch HG. Women's understanding of the mammographic screening debate. Ann Intern Med. 2000;160:1434–1440. doi: 10.1001/archinte.160.10.1434. [DOI] [PubMed] [Google Scholar]
3.Schwartz LM, Woloshin S, Black WC, Welch HG. The role of numeracy in understanding the benefit of screening mammography. Ann Intern Med. 1997;127:966–972. doi: 10.7326/0003-4819-127-11-199712010-00003. [DOI] [PubMed] [Google Scholar]
4.Woodman CBJ, Threlfall AG, Boggis CRM, Prior P. Is the three year breast screening interval too long? Occurrence of interval cancers in NHS breast screening programme's north western region. BMJ. 1995;310:224–226. doi: 10.1136/bmj.310.6974.224. [DOI] [PMC free article] [PubMed] [Google Scholar]
Editor—Wilson raises important issues in his editorial on screening for breast and cervical cancer, but they are obscured by the incorrect use of terms to describe the validity of tests.1-1
Specificity of a test measures its ability to correctly identify people without the disease being tested for. A test with 15% specificity means that only 15% of people without the disease tested negative and conversely that 85% without the disease tested positive. If 85% of women without breast cancer had positive mammograms, mammography would be useless. Wilson probably meant to say that mammography has, at best, only a 15% positive predictive value—that only 15% of those with a positive test will actually turn out to have breast cancer. However, the figures he quotes (5% of women recalled for a further test and 5-6 diagnoses of cancer per thousand women screened) give a positive predictive value for the initial screening mammogram of 10-12%.
The sensitivity of a test measures the proportion of people with the disease that is correctly identified by the test. Mammography is said to have 90% sensitivity—that is, of every 100 women with undiagnosed breast cancer at the time of the test, 90 will have a positive mammogram and 10 will have a negative one. Wilson goes on to say, however, that in the three year cycle of the NHS screening programme, 40% of all patients present with symptoms of tumours before the next screening. This means that the overall sensitivity of the programme (as opposed to the mammography) is, at best, 60%. The 40% of cases not detected by screening are all “false negatives” in terms of the performance of the programme—these patients had negative results on the test but did have or very soon went on to develop the disease.
Whether screening should be continued depends ultimately on its ability to reduce mortality or morbidity in the screened population at an acceptable total cost. Clear measures of the accuracy of the tests and of the risks and benefits that accrue are needed.
References
1-1.Wilson RM. Screening for breast and cervical cancer as a common cause for litigation. BMJ. 2000;320:1352–1353. doi: 10.1136/bmj.320.7246.1352. . (20 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
Editor—Wilson claims that “negligent false negatives on one hand and an irreducible minimum of screening errors on the other are difficult to reconcile.”2-1 I cannot agree with this statement, at least as far as screening for cervical cancer is concerned.
Several studies have shown that two types of smear give rise to false negative results.2-2,2-3 One type contains a small number of abnormal cells (<200 per slide) that are hypochromatic, small in size (diameter <15 μm), and dispersed sparsely across the slide. This type of smear accounts for the “irreducible minimum” of screening errors, and a screener who misreads such a smear does not deserve to be considered negligent. The other type contains numerous hyperchromatic tumour cells (>2000 per slide), which under optimal screening conditions should not be missed. I am frequently asked to review the previous smears of women presenting with invasive cancer who have had a long history of negative smears. Some of these women had five or more “negative” smears in the previous 10 or 12 years. Review of the smears almost always shows that at least one smear contains many obvious tumour cells. Although it is uncomfortable to admit it, there is no satisfactory explanation other than negligence for why these smears were misread.
Women are correct in their perception that cancers arising after a negative smear might have been missed and that delay in diagnosis has prognostic significance. Delay also affects treatment, as preinvasive lesions of the cervix can be treated by laser ablation whereas treatment of invasive cancer usually involves hysterectomy.
Like my colleagues I am deeply concerned at the increasing demand for compensation, but women are justified in assuming that if they present themselves for screening they are entitled to the best possible standard of care. I endorse Wilson's view that screening programmes for both breast and cervical cancer must be given enough resources to ensure that women are offered a service of the highest quality—and so that compensation becomes a thing of the past.
References
2-1.Wilson RM. Screening for breast and cervical cancer as a common cause for litigation. BMJ. 2000;320:1352–1353. doi: 10.1136/bmj.320.7246.1352. . (20 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
2-2.Smith PA, Turnbull LS. Small cell and pale dyskariosis. Cytopathology. 1997;8:3–8. [PubMed] [Google Scholar]
2-3.Baker RW, O'Sullivan JP, Hanley J, Coleman DV. The characteristics of false negative cervical smears—implications for the UK cervical cancer screening programme. J Clin Pathol. 1999;52:358–362. doi: 10.1136/jcp.52.5.358. [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2000 Sep 23;321(7263):760.
“Blind” rereading of test results provides objectivity
Editor—There are two points to be made about Wilson's editorial.3-1 Firstly, testing in asymptomatic patients differs fundamentally from testing in patients with symptoms. There are no perfect tests, and false negative results of mammography in a patient with symptoms may be “corrected” by clinical or cytological findings, or both, which will not be the case for asymptomatic patients in screening programmes. The clinician interprets the test result in the light of this other information, using bayesian techniques. In asymptomatic patients this is impossible.
Secondly, many if not most false negative smears can be detected on re-examination, but what does this mean from the legal point of view? Does this imply negligence? The Dutch Society for Clinical Pathology recently approved a procedure whereby in a litigation a disputed microscopic slide is not read by a single expert witness but is either reintroduced into normal daily practice in several laboratories where the pathologists are unaware of the outcome or is read by a panel of pathologists who are given a set of similar slides without any knowledge of the outcome or which is the slide under question. This approach circumvents biased reading of slides and gives results that are closer to those of everyday practice.
References
3-1.Wilson RM. Screening for breast and cervical cancer as a common cause for litigation. BMJ. 2000;320:1352–1353. doi: 10.1136/bmj.320.7246.1352. . (20 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
Editor—Benson et al raise the issue of frequency of screening for breast cancer and say that the three year interval of the United Kingdom NHS programme is too long. This seems correct intuitively, but I caution against assumptions made on the basis of data from the early years of this programme. Recently published data show that the breast screening programme as a whole has only recently achieved the cancer detection rate required to achieve the target mortality benefit.4-1 The results of the multicentre, randomised breast screening frequency trial of the United Kingdom Co-ordinating Committee for Cancer Research show that screening every year, rather than every three years, would provide only a marginal benefit in predicted mortality.4-2 However, I am sure that all of us involved in breast screening welcome the recent announcement that breast screening will include two view mammography at every screen and invitation up to the age of 70, when resources allow.4-3
I am grateful to Rooney et al for their tutorial on statistical nomenclature; they are quite correct in pointing out that I should have used positive predictive value rather than specificity in describing recall for assessment. Whichever measure is used, 85-90% of women recalled because of a mammographic abnormality do not have breast cancer. All cancers occurring after a normal screen are indeed false negatives, but in screening parlance the term “false negative” is used, rightly or wrongly, specifically to describe cancers that are detectable on the previous screening test.4-4 The sentiments of my editorial remain the same regardless of nomenclature: we all await mortality data that tell us whether population breast screening is worth while or not.
Coleman correctly reiterates the point that errors range from the obscure to the obvious. However, I take issue with the opinion that all obvious errors are due to negligence. Experts review cases with the benefit of hindsight, as Giard points out, and not in the conditions of the original test.4-5 Errors that seem obvious in retrospect occur even in the best quality screening programmes staffed by highly trained health professionals. To err is human . . .
References
4-1.National Health Service Breast Screening Programme. NHS breast screening programme review 1999. Sheffield: NHSBSP; 1999. [Google Scholar]
4-2.Blamey RW. The UKCCCR breast screening frequency trial. Eur J Surg Oncol. 1999;25:659–668. [Google Scholar]
4-3.The NHS plan: a plan for investment, a plan for reform. London: Stationery Office; 2000. [Google Scholar]
4-4.National Health Service Breast Screening Programme. Quality assurance guidelines for radiologists. Sheffield: NHSBSP; 1997. [Google Scholar]
4-5.Harvey JA, Fajardo LL, Innis CA. Previous mammograms in patients with impalpable breast carcinoma: retrospective vs blinded interpretation. Am J Roentgenol. 1993;161:1167–1172. doi: 10.2214/ajr.161.6.8249720. [DOI] [PubMed] [Google Scholar]
