The screening muddle

Abstract

Screening detects disease, therefore screening must be good. That simple mantra is, sadly, far from true, as the latest errors in cervical-smear screening programmes become apparent. Poor performance at the Kent and Canterbury Hospitals NHS Trust in the UK is one of the worst examples in a series of setbacks that have bedevilled such programmes since their inception in the late 1980s. 91 000 cervical smears collected during 1990-95 and examined in Kent were rechecked: 333 women at high risk of progressing to cervical cancer (which would have led to immediate follow-up) had been told their smears were normal. Eight have died of cervical cancer and 30 have undergone hysterectomies. Unsurprisingly, many are seeking legal advice with a view to compensation. Detection of abnormal cells on the screening smear would have led to repeat testing, colposcopy, biopsy, and treatment, when appropriate. Therein lies the dilemma in any screening programme. The “correct answer” is that some proportion of the results will be incorrect, because no test is perfect. The concept of false-positive and false-negative test results is a difficult one for the public to understand, but simple non-jargon explanations must be put across. In the UK programme, between the invitation to the first screen and the possibility of colposcopy and treatment, a woman can receive up to ten letters and four leaflets. Those screened deserve the most accurate answer any test can provide, but they also need to realise that tests are not completely accurate. The UK information leaflets and letters about cervical-smear testing are a model of research-based wording and design, yet only allude to the concept of a false test result, with little or no explicit explanation of what a false result means. For instance, a Health Education Authority leaflet states that “cervical screening is not 100 per cent perfect” and a new poster says that the smear test “cannot be 100 per cent accurate”. Mass screening of the population is limited to potentially fatal diseases such as cancer (eg, cervical and breast in systematic screening programmes; colorectal and prostate in experimental or self-selection programmes), and receiving a false result has many consequences for the individual. A false negative means that the opportunity for investigation and treatment has been missed. A false positive means further and, commonly, invasive tests (and possible treatment) when neither is needed. In addition, there is immense psychological morbidity associated with a false-positive result in a cancer check, as the individual (and family) waits for further tests and results. The UK screens over 4 million cervical smears a year. In nearly a tenth of the smears, the sampling is inadequate. During 1996-97, 91% of adequate smears were reported as negative. Estimates of prevented cases of cervical cancer range from 1000 to 4000 a year. The advocates of cervical-smear screening programmes can only estimate the benefits of testing, because the introduction of the programme was never tested in randomised trials. One of the few who dares speak the unspeakable about screening is a Bristol-based public-health physician, herself involved in a smear-test screening programme. Angela Raffle came under fire for writing, in a research paper in The Lancet in June, 1995, that “much of our effort…is devoted to limiting the harm done to healthy women and to protecting our staff from litigation”. She concluded: “The screening programme is identifying one in ten young women as ‘at risk’ for a disease that is likely to affect one in many thousands”. She returned to the debate in a letter in our Jan 24 issue: “Screening has distorted public belief. In our desire for good population coverage we have said that screening is simple, effective, and inexpensive. In truth, it is complex, of limited effectiveness, and very expensive”. If screened people are not adequately informed about the possibility of a false-negative result, it is little wonder that, when they get such a result, their next port of call is a lawyer. By that stage, they will not be interested in the distinction between a less than perfect test and less than perfect provision of a testing service, as provided at the Kent hospital. JNC VI guidelinesCurt Furberg and Bruce Psaty1 express the belief that the just-released JNC2 report is closer to a consensus-based than an evidence-based practice guideline, and will disappoint clinicians who prefer evidence-based guidelines. As a co-author of the JNC VI, and a practicing family physician who adheres to an evidence-based approach to health care, I must disagree. Full-Text PDF