Editor's Choice


  • Jim Thornton Editor-in-Chief

What is worthwhile?

Britain and the United States differ in their attitudes to screening for early neonatal group B streptococcal disease. Americans have long recognised it as a major preventable disease, especially in deprived urban areas, and have implemented screening. The US Centres for Disease Control advocates universal screening. British doctors have been more cautious, and if the paper by the Health Protection Agency Group B Streptococcus Working Group (pp. 1006–1011) is correct they may have been rather cavalier about the whole issue. In arguing that screening may not be worthwhile, doctors in the UK relied on prevalence studies that largely excluded deprived areas and probably underestimated the risk, and when they did detect at risk women British doctors rarely used the recommended antibiotic regimes. The present report shows that the disease is sufficiently common, at least in London, to justify the current RCOG recommendation of risk-factor based screening. Whether it justifies universal screening is debatable. I suspect most women would want screening if the issues were explained, but in Britain the final decision will, rightly or wrongly, depend on an economic cost benefit analysis.

Although there is no similar transatlantic divide over Rhesus prophylaxis with anti-D immunoglobulin, routine antenatal prophylaxis also remains controversial. It has been slow to catch on, partly because of limited supplies of naturally derived anti-D, and partly because of cost and safety concerns. The supply and safety issues largely disappeared with the development of synthetic recombinant anti-D, and the cost issue is addressed by an economic analysis on pp. 892–902 commissioned by the National Institute for Clinical Excellence from the Sheffield School of Health and Health Related Research (ScHARR). The bottom line is that the marginal cost of offering antenatal prophylaxis to primigravidae is about £12K per Quality Adjusted Life Year (QALY) gained. Extending it to all at risk women pushes the marginal cost up to about £50K per QALY. The authors leave it to readers and policy makers to make the final decisions about implementation, but such people may wonder how reliable such analyses really are.

There seems to be no major methodological flaw in the study, and the accuracy of the authors' quality of life estimates is probably as good as anyone could achieve. On a scale where full health is 1 and death zero, the authors valued minor developmental problems as 0.8 and major ones as 0.4, although even they baulked at valuing the parental grief associated with fetal death or handicap. This meant they had to ignore this in their calculations, although to their credit they discuss the problem.

The real problem is the difficulty of finding valid prices to put in the model. The whole point of economics is to allocate resources efficiently, i.e. where they will do the most good. If it cost £50K to gain a QALY from antenatal anti-D but only £30K to gain one from doing say more cervical screening - I made the latter figure up - then we should put money into cervical screening. The validity of all such calculations depends on using market prices, i.e. prices that reflect the balance of supply and demand for the resource. The problem is we know the market price for potatoes but not for health care. The NHS list price for anti-D is not a market price; bureaucrats set it. Neither does measuring the time staff spend on say amniocentesis, fetal transfusion or neonatal care and calculating a proportion of their government-regulated salaries give a market price.

Free market prices carry information, but they don't exist in the NHS. The USSR had no market prices either and eventually collapsed, partly for that reason; no-one knew whether they should be investing in women's fashions, or more steel production. As health care expenditure increasingly goes on treatments of relatively marginal benefit, it is becoming increasingly difficult for an NHS without market prices to allocate resources sensibly. Perhaps the private sector will provide the missing information. In the present study the authors turned to the independent Association of Radical Midwives to estimate the cost of administering each dose of anti-D.


Almost 50 years after its introduction the effectiveness of utero-sacral nerve ablation for the treatment of pelvic pain is at last being properly evaluated. We report one of the first randomised trials on pp. 950–959. The authors evaluated nerve ablation by the laparoscopic route (LUNA), and although recruitment was slow and the intended sample size not achieved, they should be congratulated. Surgical trials are difficult. Nevertheless we must not read too much into the results. In the subgroup of patients without endometriosis the rate of successful treatment, defined as an improvement by more than 50% in the visual analogue dysmenorrhoea pain score at 12 months, was higher in the LUNA group (P= 0.045). But 16 tests of statistical significance were done, four types of pain, recorded in patients with and without endometriosis, and assessed at two time periods. More trials are ongoing and it may be the fate of this one to end as part of a systematic review. There is nothing wrong with that.

Gene polymorphism not linked to pre-eclampsia

Readers should not be misled by our decision to publish a report of failure to find a significant link between a particular gene polymorphism (T594M) and pre-eclampsia (pp. 1012–1013). We do not normally publish negative gene linkage studies or reports of negative disease associations with gene polymorphisms, and we do not intend to alter our policy. Such studies may guide future researchers away from blind alleys but they are rarely interesting for clinicians. We published this one because the polymorphism was interesting in its own right. It had been shown to be associated with essential hypertension in some ethnic groups and people had argued that it might be involved in pre-eclampsia in places like Africa where the disease is particularly severe. It turns out that it is not.