Editor’s Choice

How do we know that new surgical techniques are safe and effective?

The introduction of novel drugs has for many years been strictly controlled, and data about their efficacy and safety have to be built up through a series of phase 1, phase 2, and phase 3 trials. A great deal of evidence has to be accumulated and presented to licensing bodies before approval for widespread use is given. In contrast, the introduction of surgical techniques has not been so closely regulated. By their nature, new surgical interventions are often developed in the heat of the moment when the clinical situation seems to demand a novel approach. An example in my own experience is that of the Hayman uterine compression suture for postpartum haemorrhage (Hayman et al., Obstet Gynecol 2002;99:502–6). Faced with a sudden emergency, and having been introduced to the concept of uterine compression sutures by the landmark paper of B-Lynch (Lynch et al., Br J Obstet Gynaecol 1997;104:372–5) but without the paper to hand and insufficient time to access it, I used my own modification. I later discovered that Professor Sabaratnam Arulkumaran had independently devised a similar modification, and so, we encouraged a junior colleague to write it up. Subsequently, a case series of the Hayman suture has been reported (Ghezzi et al., BJOG 2007;114:362–5). However, there has to my knowledge never been any prospective randomised trial either of the concept of uterine compression sutures or of the different variations employed.

So how should we approach the evaluation of surgical techniques? On page 492, Jacques Donnez et al. report a retrospective audit of 4505 hysterectomies performed by the same surgical team and compare their results using the vaginal route, laparotomy, and laparoscopy. Their complication rates were very low in all groups (about 1% or less), and they conclude that laparoscopic hysterectomy is not associated with any increase in major complication rates. In his commentary on page 473, Ray Garry expresses surprise at such low complication rates and suggests that retrospective studies are commonly flawed by confounding and bias. He advocates either prospective ‘real-time’ data collection (as opposed to retrospective chart review) as in the ‘VALUE’ study (Maresh et al., BJOG 2002;109:302–12, McPherson et al., BJOG 2004;111:688) or preferably a prospective randomised controlled trial such as the eVALuate study (Garry et al., BMJ 2004;328:129). Donnez, however, in his Mini Commentary on page 476 points out that some results are so obvious that randomised trials are hardly necessary (he quotes the use of parachutes in cases of ‘free fall’, but anyone who has seen the dramatic effect of compression sutures in a case of uterine atony would probably regard the efficacy of such sutures as equally self-evident). He argues that the learning curve can invalidate the results of small prospective randomised trials because the surgeons do not have the chance to become skilful (drugs do not have a learning curve). All three articles are provocative and deserve careful reading, with reflection on the importance of gathering proper evidence so that we can give reliable advice to our patients.

Clinical guidelines and BJOG

The aim of BJOG in its current and previous incarnations has been to ‘publish the highest quality medical research in women’s health, worldwide’. Clinical guidelines should of course be based on the highest quality evidence, but they do not of themselves constitute primary research, and therefore, as a matter of policy, we do not routinely publish them. Instead, we have recently introduced our international column, which informs readers on a monthly basis of the latest important clinical guidelines emerging worldwide. However, sometimes new guidelines have implications for clinical practice that need to be brought to the attention of our readers. Statins are now widely used to modify circulating lipid concentrations in people with high serum cholesterol levels and a strong family history and have been shown to reduce the risk of premature coronary heart disease. But are they safe in pregnancy? There are considerable concerns, which are highlighted in the commentary on page 478 by Thorogood et al. Moreover, Thorogood et al. highlight the fact that cholesterol concentrations should not routinely be measured during pregnancy because an increase is normal, and treatment is contraindicated at this time.

Ovarian cancer in pregnancy

Some years ago, we started to decline publication of case reports where the primary interest was rarity because they will have little relevance to the majority of our general readership. However, sometimes a case series will cast new light on a rare condition. In their paper on page 480, Julia Palmer et al. have ferreted out all recorded cases of ovarian cancer in pregnancy that they could find in the literature and have summarised their findings. As a result, they have put together the most authoritative account currently available. This will be helpful to anyone caring for a patient who is unlucky enough to have such a malignancy diagnosed while they are pregnant.

Barriers to preventive health care—and participation in clinical trials

When people feel ill, the majority will seek help. Many of us would like this help to be provided universally and based on treatments that have a strong evidence base. However, there has always been (and may always be) a strong uptake of ‘complementary therapies’, which usually have little evidence base. It is a truism to state that prevention is better than cure, but the paradox is that while many people will pay for therapies with little or no evidence base, equally large numbers fail to take advantage of preventive measures, which have been shown to be effective. Examples abound, for example, the use of condoms to prevent HIV transmission. On page 501, Ngan et al. report on the barriers to implementation of human papillomavirus (HPV) vaccination in China. They found that awareness of HPV infection was low, and its link with the development of cervical cancer was poorly understood. Moreover, there was a generally negative attitude towards vaccination against sexually transmitted diseases. About one in four participants in their study did not agree with vaccinating sexually naive females, thus indicating that they had no proper concept of preventive medicine as opposed to medicine to treat an established condition. While most of us will be keen to know about a condition from which we suffer, conditions that we might get (but often do not) pose only a theoretical threat rather than an immediately obvious one. Accordingly, ways have to be found to motivate the population to take out a form of ‘health insurance’ for conditions that we hope to prevent. This is easiest while the condition we are preventing remains common, but the paradox is that it becomes ever more difficult as the prevention succeeds!

A similar difficulty occurs in the practice of antenatal care. The more successful we are at preventing the most disastrous outcomes of pregnancy, the safer pregnancy appears to be, and this can reduce the motivation for women to seek pregnancy surveillance. Moreover, the paradox is further complicated by the fact that it is often the women most at risk of complications who are least likely to attend for appropriate antenatal care. On page 518, Soo Downe et al. have studied the barriers to participation in antenatal care by a technique called meta-ethnography. This is an approach to meta-analysis but based on qualitative rather than quantitative literature. They discovered a number of key issues, including denial of pregnancy (especially when these were unplanned, unexpected, and undesired), chaotic lifestyles (when any form of personal organisation becomes difficult), and a perception that antenatal care is not effective. The travel costs of attendance at clinics were also found to be a significant barrier, as was the time involved, especially for women with pre-existing major domestic responsibilities. Cultural disparity between the caregivers and the care receivers can also be a barrier, even unavoidable factors such as the midwife not yet having had her own children.

Some may ask why we should worry about people who choose not to take advantage of care that is offered. Do we have an ethical duty to try and persuade people to take advantage of evidence-based care or does that infringe their autonomy? A similar question also arises in the context of prospective randomised controlled trials. When the value of alternative treatments is in equipoise, but they have different advantages/disadvantages and perhaps different costs, there is a need to establish which is the best, from the point of view of both the society and the individual. Do we have a right (or indeed a duty) to persuade subjects to participate in properly formulated trials or should we simply offer participation and wait for those with a sense of public duty to come forward? This question is addressed in a Mini Commentary on page 542 by the ethicist Raanan Gillon. Professor Gillon considers that ‘it is entirely ethically justified to encourage potential recruits to sign up for clinical trials, provided of course that such encouragement is not threatening, misleading, or coercive’. On page 537, Hendrix et al. report on an attempted randomised controlled trial of home versus hospital birth in the Netherlands. An initial pilot study suggested that 50% of women would agree to participate. However, once the trial proper got under way, despite 115 women agreeing to participate in a cohort study, only 1 agreed to be randomised with respect to planned place of delivery. The most common reason they gave was that since there was doubt about the relative safety of home versus hospital births, they wished to make their own decision based on personal criteria, and ‘researchers should not be meddling in this decision’. Unfortunately, equipoise about the relative safety of place of delivery is largely due to lack of adequate information rather than a confirmed equivalence between the two alternatives (as explained in the paper by Mori et al. in last April’s BJOG—watch this space for further developments in this discussion). Professor Gillon makes the point that the approach of the counselling midwives to recruitment is also deserving of study; whether the recruiter is in favour of one of the alternatives offered is likely to make a big difference to the way they put the issues across. He considers it entirely ethical to use only recruiters who genuinely favour the trial going ahead. The importance of consistency and appropriateness in how trialists put information across to potential participants equally applies to the use of questionnaires in qualitative research, as explained in a linked Mini Commentary by Professor Tina Lavender on page 541. We receive many papers based on questionnaires, and only too often they have not been properly piloted or assessed for content validity. Not sure what ‘content validity’ means? Read Professor Lavender’s Mini Commentary and find out (further information can be obtained from Kelley et al., Int J Qual Health Care 2003;15:261–66). Professor Lavender quotes Oppenheim’s comment that ‘questionnaires do not emerge fully fledged; they have to be created or adapted, fashioned, and developed to maturity after many abortive test flights’. As with any flight, piloting is essential.