Evidence to be given to the public must be presented accurately and fairly

Editor-Promoting informed choice is commendable provided the evidence is presented accurately and fairly. The case study by S Oliver and colleagues exemplifies how evidence may be biased.1 The authors studied reactions to leaflets prepared by the Midwives' Information and Resource Centre and the NHS Centre for Reviews and Dissemination, organisations of which four of the authors are employees. We are not told whether these same four investigators were also authors of the leaflets, but the question of conflict of interest must arise.

The leaflet “Informed choice for women” states: “one in 200 babies who were aborted as a result of what a scan showed were in fact normal or had only minor things wrong,” and Oliver et al quote this figure four times. As an author of the paper from which this statistic is derived,2 I suggest that the evidence should be interpreted in its true context, otherwise misinformation will be the result.

We studied 2261 pregnancies with a fetal anomaly diagnosed by ultrasonography in Yorkshire during 1989-91. Altogether 369 of these pregnancies were terminated, and in two the anomaly proved less severe than had been predicted on ultrasonography. As Paul Chamberlain and P A Boyd say in their commentary on Oliver and colleagues' paper,1 one of these fetuses had a gastroschisis (which still carries a considerable perinatal mortality); the other fetus had enlarged hyperreflective kidneys, and, despite normal histological findings, the outlook for renal function in such a fetus remains open to speculation.3 This 99.5% specificity for the diagnosis of fetal anomalies warranting termination was achieved from a region-wide hospital mix with ultrasonography equipment of varying sophistication and by ultrasonographers with differing levels of experience.

The diagnosis of fetal anomaly has continued to advance since 1991, with developments in ultrasonographic technology and with improved training programmes for sonographers, radiologists, and obstetricians. The specificity of ultrasonographic diagnosis of fetal anomalies is confidently thought to have increased beyond 99.5% already.

It is a pitfall of evidence based medicine that evidence may be misinterpreted or may be presented in a biased manner. There are few diagnostic tests in medicine that perform to such high accuracy, and it would have been appropriate to inform the public in language that conveys this good news rather than in an alarmist manner. I share the concerns of the ultrasound professionals quoted by Oliver and colleagues that these leaflets may do more harm than good, and I suggest that the relevant health professionals are consulted when such leaflets are next prepared.