Guidelines and standards
Formal guidelines play an increasing role in clinical care, with a belief that they can result in improved quality of care and ensure that similar quality is delivered across the healthcare system. We are used to matching our clinical performance against clinical standards often set by guidelines to the extent that there is a financial incentive within the Clinical Negligence Scheme for Trusts to both have particular guidelines and perform regular audit of compliance with them. There is less information on how good clinical guidelines are in the first place, and they can derive from a variety of sources and methods. It is common sense to infer that guidelines that are produced badly or that are impractical are likely to impair the quality of care rather than improve it.
Tools exist by which guidelines can be assessed, and Appleyard et al. (p. 749) have examined guidelines on the management of pelvic pain due to endometriosis using two of these tools. Cluzeau’s tool examines three domains—rigour of development, which includes examination of whether the group producing the guidance had all the relevant players and assessment of how evidence was sought, content and context, which includes looking at the stated objective and clinical context of the guidance, and applicability, which includes evaluation of dissemination strategies and advice on implementation and audit. The AGREE instrument looks at six domains—scope and purpose, stakeholder involvement, rigour and development, clarity and presentation, application and editorial independence.
Using a very wide search strategy, eight guidelines were identified that fulfilled the prespecified criteria of the authors. The overall methodological quality of these guidelines was not reassuring, with only three of the eight meeting more than half of the standards for rigour of development. Context and content of guidelines appeared to score highest and applicability standards the poorest. Poor scoring in this domain means to some extent that however good the guidance is, it is unlikely to be well implemented, somewhat defeating the purpose of the exercise. The Royal College of Obstetricians and Gynaecologists (RCOG) guidelines at least made the cut, but examination of the domains does not make comfortable reading with scores of 50 and 40% for rigour and 10 and 20% for applicability.
Most investigations of guidelines have thrown out similar messages, and given the plethora of guidelines and their perceived importance, it is disappointing that standards are not what they should be. The RCOG’s strategy for the production of ‘green top’ guidelines is out for review, in part because it was appreciated that the previous methodology had become obsolete and was failing to meet modern quality criteria. Professional bodies in particular need to ensure highest standards in the production of guidelines, and this article is a reminder that guideline production needs as much scientific rigour as primary research.
Van der Steeg et al. (p. 825) address a different aspect of guidance—what happens to clinical decision making where there is either poor or poorly disseminated information that can guide care? They produced 16 fictional vignettes of couples undergoing fertility investigations and asked 35 specialists to predict pregnancy likelihood for spontaneous conception, following intrauterine insemination and following in vitro fertilisation. They then asked for the management plans. The vignettes differed in seven factors, five of which were known to influence the likelihood of pregnancy. After a period of 3 months, the process was repeated, but this time, predictive information from existing models was provided.
There was poor concordance between clinicians in predicting pregnancy success, with resultant wide variation in the treatment that would have been offered. Unsurprisingly, treatment followed individual predictions. Providing predictive information improved concordance and led to an increase in conservative management (awaiting spontaneous conception for a period of time) from 39 to 51%.
The study suggests that without adequate or easily available information, clinical practice will vary widely, with both inappropriate intervention and nonintervention possible. Providing guidance, in this study as prediction models, at least results in more consistency of care, and hopefully more appropriate care.
Blair Bell Research Society abstracts
We publish the Blair Bell abstracts every year. I am never sure how widely they are read. These abstracts showcase current research in our specialty and generally represent the cutting edge across all the subspecialties. Many of these will see the light of day as full papers, and although we would love to see all of these in BJOG, many will be published in other specialist basic science journals. So this may be your only opportunity to find out what is going on and what might be around the corner, whether you are interested in new treatments for vulval intraepithelial neoplasia, normal ranges for coagulation parameters, transferring genes, the pathogenesis of pre-eclampsia or the how the uterus works. Don’t miss it!
Minimally invasive gynaecology?
Two articles look at the use of laparoscopy in this issue—one seeking to avoid its use and one using it in a clinically radical way.
Clomiphene-resistant polycystic ovary syndrome remains a problem in reproductive medicine services, and a number of techniques have been devised to establish ovulation, including wedge resection and more recently laparoscopic drilling techniques. Ultrasound-guided ovarian follicle aspiration is a common technique during assisted conception. A Chinese group (Zhu, p. 810) have combined these approaches to carry out transvaginal-ultrasound-guided laser coagulation within the ovary. Under fairly heavy sedation, the ovary is punctured in its largest plane through the vaginal fornices, guided by ultrasound. A fibro-optic laser cable, where the tip has been unsheathed, is then passed down the needle and low intensity coagulation carried out. Several areas of each ovary, where possible, are subjected to this. Twenty-three women agreed for treatment and it was possible to access both ovaries in 21 women. Follow up demonstrated a return in ovulatory hormonal parameters in 19 women and 8 pregnancies have occurred, albeit with additional therapy. There were no apparent surgical complications. The technique clearly has a long way to go before it could be considered as part of practice and will need careful evaluation within trials comparing other more widespread methods, but watch this space.
The second article by Swanton (p. 792) on first sight confirms all my prejudices about gynaecology. Chronic pelvic pain is as common a problem as chronic back pain; yet, often no clear cause is found and various surgical interventions may not alleviate the pain. The Oxford group have taken the basic clinical question of ‘does it hurt when I press here?’ to a new level. Using laparoscopy under local anaesthesia, they ask the woman if the pain that brought her here is reproduced when the pelvic organs are probed during laparoscopy. The technique is called conscious pain mapping.
Of the 43 women consenting, conscious pain mapping was possible in 39, and in 35 of these women, probing reproduced the pain. Of interest is that in the 18 women with demonstrable pelvic pathology, 6 had no reproducible pain on probing of that pathology and 5 of them had another site for their pain. Where the pelvis appeared normal, women had exquisite tenderness on probing of vaginal vault, uterus or ovary.
Treatment was guided by the results of conscious pain mapping, though many came ultimately to surgery. The authors are careful not to make too much of the technique, calculating that only 27% of the women avoided inappropriate treatment as a direct result of its use. However, it seems a technique worth further exploration and comparison.
The article by Brown et al. (p. 803) seeks to establish cost-effectiveness by comparing two treatments for heavy menstrual bleeding. Although we all subscribe to the concept of cost-effectiveness, few will have knowledge of the science behind it and its limitations. Brown et al. have included a cost-effectiveness analysis within a randomised clinical trial of thermal ablation versus the levonorgestrel intrauterine system for heavy menstrual loss. By embedding within a RCT, many of the biases that can arise in carrying out cost-effectiveness work are minimised. Costs are derived largely from primary sources, another advantage for this type of study, and good follow-up data on treatment failure and the costs to women have been obtained. The assumptions are transparent, another important feature of good cost-effectiveness studies.
The methodology is clearly described, and anyone wanting to understand the ‘how’ of cost-effectiveness should take the time to go through this article, as it will give them a better understanding of the principles and assumptions of the technique.
Seventy-nine women were randomised and so the data are not derived from large numbers. The levonorgestrel system appears to be more cost-effective, and this holds at different levels of sensitivity, i.e. altering costs or outcomes by up to 25%. There are clearly limitations to the data and the authors acknowledge these, both from a methodological perspective and from a clinical viewpoint. However, it would seem that if there were an equal choice for an individual woman needing treatment for heavy menstrual loss, then the hormonal device is more cost-effective.