Editor’s Choice

All of us will be familiar with the apparently inexhaustible supply of news stories about ‘breakthroughs’ that will not only improve health but ‘will reduce the pressure on health care provision’. Recently, we have been bombarded with messages that if we could stop smoking, slim down, and drink less alcohol, a huge burden would be lifted off the health services. But journalists seem blind to the fallacy of their argument. We would certainly live longer, healthier lives, but we all have to die some time, and the bulk of our healthcare consumption is in the last few months of our lives, whether that is when we are 18 or 80 years. Indeed, the degenerative diseases of old age are often the most expensive way to die. Decades ago, a popular view among cynics was that governments were reluctant to do anything to curb tobacco consumption not only because it raised billions in tax revenue but also because dying of lung cancer was quick and cheap. So, it makes sense that the emphasis should instead be on maximising quality of life and minimising disability, while we are fortunate enough still to be alive. A widely used measure is the ‘QUALY’ or ‘quality-adjusted life-year’. This is the arithmetic product of life expectancy and a measure of the quality of the remaining life-years. A year of perfect health is worth 1, while a year of less than perfect health life expectancy is worth less than 1. Death is considered to be equivalent to 0; however, some health states may be considered worse than death and have negative scores (www.evidence-based-medicine.co.uk/ebmfiles/WhatisaQALY.pdf). Cost-effectiveness analysis compares the costs and health effects of an intervention to assess whether it is worth doing from the economic perspective (www.evidence-based-medicine.co.uk/ebmfiles/Whatiscosteffect.pdf). To be truly accurate, it needs to include the lifelong consequences of the intervention, and large university departments have now developed to address this complex task. On page 749, Hadwin et al. address one such issue in relation to colposcopy referral guidelines. Until 2004, in the UK, the national recommendation for the management of women with a mildly abnormal cervical smear result (mild dyskaryosis) was to rescreen at the local primary care centre 6 months after the initial test. However, a minority of these women subsequently prove to have high-grade cervical intraepithelial neoplasia (CIN3). Earlier referral for colposcopy allows improved detection of preinvasive disease and therefore presumably a further reduction in invasive carcinoma incidence. One can imagine the journalistic headlines: ‘testing breakthrough reduces women’s cancer deaths’. But what about the false positives? The potential for unnecessary surgical excision is substantial, carrying with it the risk of, for example, preterm labour in any future pregnancy (as highlighted by Bruinsma et al., and Paraskevaidis et al., in BJOG January 2007). Indeed, the authors warn against a policy of early treatment for mild dyskaryosis; it reduces workload but exposes many women to the risks of treatment who cannot benefit because their abnormality would have resolved spontaneously anyway. Moreover, money spent on increasing the number of colposcopies cannot be spent on other interventions that might produce greater overall benefit. Hadwin et al. describe how they built their cost-effectiveness model, which included assessing the effects of the natural history of the condition with and without treatment. Some of the techniques used will not be familiar to most clinicians. For example, they used a state-transition (Markov) approach to simulate disease outcomes over a period of 3 years (which they call a ‘time horizon’—all specialists have their own jargon). The purpose of this mathematical model is to simulate not just a simple divergent model in which things either happen or they do not but a more realistic situation in which events may recur at intervals, as abnormal smears have an unfortunate tendency to do (www.treeage.com/learnMore/MarkovModels.html). They also used two subcomponents (or ‘matrices’) in their model: one in which the smear result changed when rescreening took place and one to simulate disease progression. They use their model to determine the ‘cost per life year gained’ (LYG) and also the total QUALYs gained. Their first conclusion is that colposcopy for mild dyskaryosis would increase the number of colposcopies by one-quarter; in clinics, which operate low-intensity screening, the capacity increase needed would be over 50%. Such an increase could clearly not be implemented overnight, and so there might have to be balancing changes in policy, for example not offering colposcopy following treatment. What of the benefit to the patient? The cost per LYG as a result of increased referrals ranges from £10,000 to £35,000 according to the screening interval, while the cost per QUALY would be approximately £13,500. For comparison, the National Institute for Health and Clinical Excellence in the UK will currently consider for introduction procedures costing less than £30,000 per QUALY gained. On this basis, colposcopy for mild dyskaryosis can be considered cost-effective. Will journalists hail this as a new ‘breakthrough’? Will resource constraints allow it to happen? Time will tell. So, much of healthcare implementation depends on emotion and pressure groups rather than on science; however, cost-effectiveness analysis does provide a rational approach to such decisions. It is likely, however, that they will always be controversial. In February, we published a paper by Khalaf et al. that described their technique for reducing the number of blastocysts transferred and thus reducing the multiple pregnancy rate, while maintaining success rate. In the substantial publicity that followed, successful policies in Sweden and Belgium to reduce the multiple pregnancy rate following IVF were highlighted. In this month’s journal on page 758, we publish a paper from Sheffield (UK) that evaluates the cost-effectiveness of different embryo transfer strategies in England. A major part of the problem in the UK is that 91% of the state funding agencies that support IVF will only fund one cycle, thus putting enormous pressure on clinics to replace two embryos and give the parents a chance of two children rather than one, despite the greatly increased risk of perinatal death and cerebral palsy associated with twin pregnancies. Dixon et al. therefore decided to evaluate the cost-effectiveness of different embryo transfer policies, considering not only the cost-effectiveness of the transfer itself (in terms of pregnancies achieved) but also the costs of the resultant adverse outcomes. Their units of comparison were between the transfer of two fresh embryos at one time, single embryo transfer, and the transfer of one fresh embryo followed by the transfer of the other (frozen) embryo in a subsequent cycle. This makes sense because a frozen cycle costs only about 30% as much as stimulated cycle, and the risks to the woman are also much lower. They use another fancy statistical technique known as the Dirichlet distribution. This looks at continuous multivariate probability distributions. It is not easy to explain in words, but there are some helpful pictures on Wikipedia (http://en.wikipedia.org/wiki/Dirichlet_distribution). Such a technique is necessary because the outcome of a twin pregnancy is obviously multivariate! The main adverse outcomes assessed were preterm birth, neonatal intensive care admission days, and cerebral palsy. Their analysis shows that when the adverse outcomes are factored in, double embryo transfer is the least cost-effective. However, it should be remembered that cost-effectiveness is not the only consideration. Previous papers published in BJOG (Scotland et al., BJOG 2007;114:977–83) have shown that for many parents, achieving a pregnancy is more important than the likelihood of an adverse outcome for the child. Hopefully, the days are long gone when low Apgar scores were routinely interpreted as indicating ‘birth asphyxia’, with its associated hypoxia and acidosis. We now know that babies can be depressed at birth for a variety of reasons, including infection, trauma, and meconium aspiration, as well as acidosis. But surely at least measuring the pH or the lactic acid level in the umbilical cord vessels is a straightforward indicator? On page 697, Wiberg et al. report that there are significant changes in these variables if the cord is not clamped immediately after birth. They hypothesise that this is due to opening up of the vascular beds in organs of the fetus that have been shut off to reduce oxygen demand during the stress of labour (‘hypoperfused nonpriority organs’) as soon as the baby starts breathing. The mean drop in arterial cord blood pH was only 0.03 over the first 90 seconds, but in some cases, the fall was substantially greater than this. The same group have also studied the variation of lactate levels in arterial and venous umbilical cord blood at birth with gestational age in almost 18 000 newborns (Wiberg et al., on page 704). Their results confirm previous studies on pH that acidosis at birth even in vigorous newborns increases steadily from 34 weeks of gestation onwards. This presumably reflects the gradually failing placental function that may be an important trigger for the onset of labour and is likely linked to the rise in stillbirth rate towards and after term. So, it seems that everything in life is relative. As I have said before in my editor’s choice, the problem with case reports is that they often deal with rarities and are therefore irrelevant to the majority of our readers who will never see such a case during their clinical career. Some can illustrate a new approach to a common problem, or a new complication associated with a recently introduced technique, and these we think can be worth publishing. Another category is when rarities are gathered together in a substantial case series, thus generating the opportunity of defining the characteristics of the group. This is illustrated in the case series of 29 women with Swyer syndrome that we publish on page 737. They have an XY karyotype but normal female external genitalia—and unlike women with testicular feminisation, they also have a uterus. So, if you ever come across a case, this is the paper to read to give you background information. For many years, there have been calls to reduce the caesarean section rate in developed countries, and yet year on year, the rate continues to increase—while the maternal mortality continues to fall. Those who argue for a reduction have to persuade women who would otherwise have a caesarean (and their obstetricians) to pursue more doggedly a vaginal birth. Reports of the adverse effects of caesarean section are grist to their mill, and certainly, the increased rate of percent of praevia and accreta in subsequent pregnancies (together with the increased risk of caesarean hysterectomy) appears indisputable. A paper by Gordon Smith et al. in the Lancet in 2003 (Lancet 2003; 362:1779–84), which reported that the risk of ‘unexplained’ stillbirth after 39 weeks of gestation was more than doubled following a previous caesarean section compared with a previous normal birth, attracted a lot of attention and has to date been cited 64 times (which is a lot!). On page 726, Wood et al. report a study of more than 158 000 second births in Alberta Canada and found no significant difference in the rate of unexplained antepartum stillbirth by mode of first delivery after adjusting for maternal age and blood pressure. Will their findings be equally highly cited? We still receive many submissions to BJOG that are not correctly formatted for the type of article, be they randomised controlled trials, observational studies, or meta-analyses. A new website provides invaluable guidance, with a plethora of references to papers describing how to write up your research correctly. It is a new initiative that ‘seeks to improve the quality of scientific publications by promoting transparent and accurate reporting of health research’ and is called ‘The Equator Network’. It can be found at www.equator-network.org/. I commend it to you wholeheartedly. Enjoy!

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.