1 Department of Obstetrics and Gynaecology, University of Auckland, Auckland 3, New Zealand, 2 University Department of Obstetrics and Gynaecology, University of Cambridge, Cambridge CB2 2SW, 3 Nuffield Department of Obstetrics and Gynaecology, John Radcliffe Hospital, Oxford OX3 9DU, UK, 4 Department of Obstetrics and Gynaecology, Academisch Ziekenhuis Maastricht, Maastricht 6202AZ, The Netherlands and 5 Institute of Epidemiology & Health Services Research,Leeds LS2 9LN, UK
Abstract
The last two decades have seen a rapid explosion in research surrounding subfertility treatments. This ever-increasing volume of research has made it a difficult task for health professionals involved in the management of the subfertility to be able to assimilate the information easily. There is an urgent need for the findings from research to be synthesized into simple easy to read reviews that are both of a high quality and are based on the best evidence available. The Menstrual Disorders and Subfertility Group of the Cochrane Collaboration is attempting to address these issues by collecting a register of all the randomized controlled trials in the field of reproductive medicine and preparing systematic reviews on topics that will be of interest to healthcare workers and consumers. Readers are invited to participate in this process by identifying published and unpublished data and by helping in the process of preparing protocols and systematic reviews for inclusion in the Cochrane Library.
Key words: Cochrane collaboration/hormone replacement therapy/randomized controlled trials/subfertility/systematic reviews
Introduction
To offer patients effective treatments that do more good than harm is a fundamental and basic tenet of being a good doctor. This issue is particularly pertinent in the management of subfertility where outcomes with disastrous consequences are not uncommon. The dilemma for many specialists in reproductive medicine is that there is a huge volume of rapidly changing research and few resources for the translation of evidence from research into every day practice. How can clinicians know if the treatment they are offering is effective in improving fertility outcomes? Where do they go to look for evidence which is accurate, free of bias, and up-to-date? There are a number of journals that publish reproductive medicine articles and many other useful articles are published in general medical journals. For example, the database of the Cochrane Menstrual and Subfertility Group has almost 1800 randomized controlled trials relating to subfertility and these have come from >30 journals (S.Furness, personal communication). The majority of them were published within the last 10 years. A further example of the large amount of information is in the specialty of general medicine where physicians would need to read 19 articles a day, 365 days a year to keep up to date (Davidoff et al., 1995), yet self reporting by British medical consultants suggest that <1 h per week is available (Sackett, 1996
). Thus it had become increasingly difficult to sort, collate and translate research findings about effectiveness into appropriate patient care and it has become almost impossible for (and unreasonable to expect) individual providers, planners, or users of healthcare to have the facilities or time to remain informed or up-to-date with research evidence, even in limited areas such as subfertility.
In the past, the solution to this `information overload' has been to read summaries of primary research or overviews in journals or textbooks. Authors of such reviews usually attempt to collect all the reports on their chosen subjects, evaluating and summarizing the contents in a report. Unfortunately, often the quality of many reviews is lacking (Mulrow, 1987) or simply not clear. They are rarely comprehensive and do not contain a description of how the trials are identified and selected for inclusion. Failure to use either a systematic approach for data retrieval or appropriate statistical methods when these are indicated may further detract from the quality of many reviews. These areas may lead to conclusions being made that are at best biased and at worst often wrong (Altman et al., 1992
). This has meant that some highly effective forms of care have been delayed for many years while others are being used long after controlled research has shown them to be ineffective or even harmful. The most compelling illustration of this delay was the tardiness of obstetricians to introduce antenatal corticosteroids for women in preterm labour long after randomized controlled trials had demonstrated reduced mortality in the infants who were subsequently born. It was not until systematic reviews which incorporated meta-analysis were published that changes in practice began to occur (Crowley et al., 1990
). Although the need to apply scientific principles to the design and conduct of primary research is generally accepted, this is less appreciated in the preparation of a review (often described as secondary research). Reviews failing to observe these principles or based on studies with unreliable research strategies or poor quality, can reach invalid conclusions with serious consequences (Altman et al., 1992
). The failure of reviews to even consider the quality of the included trials is a further problem of the unsystematic review (Schultz et al., 1995).
In 1979, Archie Cochrane (19091988), a British epidemiologist and clinical trialist, wrote: `It is surely a great criticism of our (medical) profession that we have not organized a critical summary, by specialty or sub-specialty, updated periodically of all relevant randomized controlled trials' (Cochrane, 1979). He suggested that for many, if not most forms of care, the randomized controlled trial was the form of research most likely to yield reliable estimates of the effects of care (Cochrane, 1972
). He was concerned that only those forms of care which had been shown to be effective without serious drawbacks should be encouraged. Forms of care which had been shown to do more harm than good should obviously be discarded. Forms of care whose effects are unclear should only be provided within the context of controlled trials.
Such criticisms of modern medical practice may seem unfair but assessing the management of reproductive medicine suggest that they may be justified. There has been a tendency in the reproductive medicine field, and in obstetrics and gynaecology in general, to introduce new treatments without full evaluation, often on the basis of physiological considerations. In fact, almost all treatments in reproductive medicine have been introduced without proper evaluation by a randomized controlled trial. Although in-vitro fertilization (IVF) in the case of bilaterally occluded tubes and intracytoplasmic sperm injection (ICSI) in the case of extreme oligozoospermia seem to make a randomized controlled trial redundant, spontaneous waiting list pregnancies have been reported for both, varying from 2 to 6% per year (Snick et al., 1997; Evers et al., 1998
). Another problem is that ineffective treatments are not abandoned for years even when adequate randomized controlled trials showing no effect are available. While using ineffective treatments may allow subfertility specialists to feel that at least they are doing `something', in many cases delays in commencing effective treatments occur and for a women in her late 30s this delay could be important. In other cases the treatments may incur a considerable cost to the patient, both financially and in personal health, through side effects or complications of a particular intervention. For example, more than two-thirds of one group of gynaecologists who were surveyed prescribed medical treatments for mild endometriosis in women who wished to conceive (C.M.Farquhar, unpublished data). This is in spite of a systematic review of randomized controlled trials published both in a prominent reproductive medicine journal (Hughes et al., 1993
) and in the Cochrane Library (since 1995) which has not demonstrated any benefit from this approach (Hughes et al., 1998
). Indeed, it is likely to expose these women to 6 months of contraceptive therapy and thus delay the commencement of effective treatments. Other examples, include the introduction (albeit briefly) of zygote intra-Fallopian tube transfer (ZIFT) (Tzafettas et al., 1994
), douching of the vagina to reduce cervical hostility, varicocoele surgery for oligozoospermia (Schatte et al., 1998
) and the widespread use of post-coital tests (Oei et al., 1998
). These are but a few examples of current practice where little evidence exists but effectiveness is assumed.
The Cochrane Collaboration
In response to Cochrane's challenge, The Cochrane Collaboration, an international network of individuals and institutions, has attempted to address this need. In 1992, the UK Cochrane Centre was funded as part of the Research and Development Programme developed to support the UK National Health Service (Chalmers et al., 1992). A year later this initiative was launched internationally as the Cochrane Collaboration. The aim of the Cochrane Collaboration is to prepare, maintain and disseminate systematic reviews of randomized trials of the effects of health care. Finding out what can be known from already existing evidence is not a minor task; it is estimated that as many as a million studies conducted over the last 4050 years may need to be considered. Since its inception in 1993, an international Cochrane Collaboration has evolved with further centres in Australia, Canada, Italy, Scandinavia, The Netherlands, France, South Africa, Spain, Germany, South America and the US.
The Cochrane Collaboration comprises an increasing number of Collaborative Review Groups; each represents an area of interest and is co-ordinated by an Editorial Group. Members of such groups include people with an interest in evidence-based medicine and a special interest or expertise in a particular topic. Members are both international and multi-disciplinary (usually with consumer input) who contribute in the preparation and maintenance of a systematic review of their choice or interest. A combination of knowledge about the topic and methodological rigour to the review process is required (Oxman, 1994). To find the right balance between these factors, and also for the purpose of quality control and reproducibility, a partnership of reviewers is required in most cases. Taking responsibility for a review is a long-term commitment since reviews need to be kept up-to-date and amended to take account of new evidence or knowledge, valid criticisms, or when mistakes are identified. Keeping a review up-to-date is not only necessary to identify `new' effects promptly, it also assists in the planning of new research by pointing out promising leads and avoiding research on questions that have already been answered.
Systematic reviews
Systematic reviews and meta-analysis allow for the aggregation of data in a quantitative manner. Clinicians can therefore save time by searching for a relevant review and not individual studies. The reader is able to deal with the diversity of literature and hopefully arrive at a conclusion that can be applied in clinical practice (Oxman, 1994). Each systematic review represents the work of several individuals, usually over several months. Initially titles are registered with the collaborative review group, and both protocols and reviews are peer reviewed by the editorial team and by external referees (in the case of completed reviews this would be similar to the practice of any scientific publication) (Figure 1
). Cochrane reviews are based on the analysis of randomized controlled trials where available. Carefully controlled randomized trials are the most appropriate and reliable form of assessing most interventions under investigation; if such evidence is not available, this should be clearly stated. A systematic review undergoes the following stages: (i) statement of the objectives of the review, and outline of the eligibility criteria for trials; (ii) search for studies meeting eligibility criteria; (iii) assessment of the characteristics and methodological quality of identified studies; (iv) decision on inclusion or exclusion of identified studies; (v) data extraction; (vi) analysis of results, using meta-analysis if appropriate; (vii) sensitivity analyses; and (viii) preparation of a structured report to the review (Figure 2
).
|
|
It is the responsibility of each Collaborative Review Group to develop a specialized register of trials of interest to their scope. The Cochrane Collaboration provides organizational support in order to avoid duplication of effort. This is particularly relevant with regard to the development of an international register of randomized controlled trials. Until 1994, MEDLINE did not tag randomized controlled trials sufficiently accurately (both in terms of sensitivity and specificity) for electronic searches to be reliable as a means of identifying them. As a result of this, together with inadequate provision of information by authors, 50% (or more) of trials eligible for inclusion in reviews can be missed by relying solely on electronic searching (Dickersin et al., 1994). Trials from some specialities are particularly poorly located by electronic searches of MEDLINE. For example, in the subfertility area only 16% of all randomized controlled trials were found by this method (Vandekerckhove et al., 1993
). This failure by MEDLINE to identify randomized controlled trials has resulted in intense manual searching activity by various Collaborative Review Groups and Cochrane Centres. Randomized controlled trials identified by handsearching are tagged retrospectively by MEDLINE (though the Baltimore Cochrane Centre) as such in updated issues of the database. Identified randomized controlled trials are also available in the Cochrane Controlled Trial Register in the Cochrane Library.
Once systematic reviews have been prepared and submitted for editorial and peer review, the results are disseminated in electronic form through The Cochrane Database of Systematic Reviews and (if the authors wish to do so) by publication in medical journals or bulletins. This journal, Human Reproduction, and its sister review journal, Human Reproduction Update, have already published several systematic reviews in the past and it is hoped that this policy can be extended to Cochrane reviews (Hughes, 1997). The results of each review should be of interest to those both in clinical practice and in research, influencing decision making and the planning of future research strategies. Reviews are updated regularly and available on CD ROM and the internet (http://www.update-software.com/ccweb/Cochrane/cdsr.htm) (Figure 3
).
|
The Cochrane Library is the regularly updated electronic library presenting the work not only of the Cochrane Collaboration but also the National Health Service Centre for Reviews and Dissemination based in York, UK. The Cochrane Library consists of four databases: the Cochrane Database of Systematic Reviews, the Database of Abstracts of Reviews of Effectiveness, the Cochrane Controlled Trial Register, and the Cochrane Review Methodology Database (Figure 3). The Cochrane Database of Systematic Reviews contains the end products of the work of the various Cochrane Collaborative Review Groups. At present, it contains over 500 systematic reviews on an increasingly wide variety of subjects. Protocols of systematic reviews in progress or planned in future (over 600) are also listed. The Cochrane Database of Systematic Reviews aims to provide the information that is needed by health care workers and consumers to make clinical and purchasing decisions. The Database of Abstracts of Reviews of Effectiveness provides a bibliography and structured abstracts of reviews of effectiveness of health care performed by other groups outside the collaboration including the National Health Service Centre for Reviews and Dissemination in the UK. The Controlled Trial Register is the result of the searching for randomized controlled trials done by the various Cochrane Collaborative Review Groups. The Cochrane Review Methodology Database provides a bibliography on the science of research synthesis and on practical aspects of preparing systematic reviews.
The Cochrane Library is work in progress and there are still many gaps; it will take many years before the Cochrane Library is fully fleshed out, and before the database is comprehensive. It could be said that the Cochrane Library will never be complete and the work of the Cochrane Collaboration will be ongoing. In many areas primary studies of higher quality need to be completed before unbiased assessment in a systematic review can be attempted. However, with each release of The Cochrane Library (four times annually), progress will be made both in extending the coverage of the material by the addition of new reviews, and in the functionality of the software used to present this material. In comparing the enormity of the task it has been stated that `the Cochrane Collaboration is an enterprise that rivals the Human Genome Project in its potential implications for modern medicine' (Naylor, 1995).
Systematic reviews in the management of subfertility
Systematic reviews in the management of subfertility are planned under the auspices of the Cochrane Menstrual Disorders and Subfertility. The Cochrane Subfertility Group registered with the Cochrane Collaboration in 1993 and was based in Leeds, UK. The Menstrual Disorders Group registered in 1995 and is based in Auckland, New Zealand. The two groups merged in 1997 and the administrative base is in Auckland, New Zealand. Members of the group include gynaecologists, reproductive medical specialists, endocrinologists, general practitioners, public health specialists, family planning specialists, physicians, scientists, pharmacists, academics, nurses and consumers. The topics relating to subfertility include all treatments of subfertility, endometriosis, unexplained infertility, infertility related to polycystic ovarian syndrome (PCOS), tubal infertility, assisted reproductive technology (ART), ectopic pregnancy and male infertility.
A number of systematic reviews of randomized controlled trials of treatments of subfertility are already published in the Cochrane Library and many more are in preparation (Table I). Currently there are reviews on endometriosis and subfertility, gonadotrophins in ART, ovarian drilling in PCOS, ovulation induction in PCOS, tubal flushing, ectopic pregnancy, oligozoospermia, pelvic surgery, and medical treatments of unexplained infertility. A number of reviews are in progress.
|
|
There are almost 2000 randomized controlled trials in subfertility. It is clearly not possible for the reproductive medicine specialist to know the content and quality of each of these trials, let alone implement the research findings within the practice of subfertility. (The Cochrane Menstrual Disorders and Subfertility Group is attempting to co-ordinate and focus this process in order for systematic reviews to be prepared and disseminated). There are a number of ways that readers can contribute to this process; searching a journal for randomized controlled trials, becoming a reviewer, or external referee, or making unpublished data available.
In particular, the Cochrane Collaboration wishes to be notified of any unpublished randomized controlled trials in which readers have been involved as investigators. These may have been trials for the purpose of registration of a new product which have never been formally published or may have shown negative or even adverse results. It is very important that all such trials are notified and reviewed to minimize the bias brought about by only including published trials with positive results. All pharmaceutical companies that currently market or have previously produced products for the management of subfertility are encouraged to notify the Cochrane Menstrual Disorders and Subfertility Group of the existence of such trial data. It is important from the pharmaceutical industry's point of view that their products and the adequacy of the data on them are fully represented in the Cochrane reviews as these evidence-based reviews may greatly influence prescribing and registration recommendations in the future.
Readers are invited to contact the editorial base: Dr Cindy Farquhar (c.farquhar{at}auckland.ac.nz), Co-ordinating Editor, or Ms Sarah Hetrick (s.hetrick{at}auckland.ac.nz), Review Group Co-ordinator, Cochrane Menstrual Disorders and Subfertility Group, Department of Obstetrics & Gynaecology, University of Auckland, School of Medicine, National Women's Hospital, Private Bag 92 189, Auckland, New Zealand, or to contact any of the authors for further information regarding the work of the review group.
Notes
6 To whom correspondence should be addressed
This opinion was previously published on Webtrack 61, April 16, 1999
References
Altman, E., Lau, J., Kupelnick, B. et al. (1992) A comparison of results of meta-analyses of randomized controlled trials and recommendations of clinical experts. Treatments for myocardial infarction. J. Am. Med. Assoc., 268, 240248.[Abstract]
Chalmers, I., Dickersin, K. and Chalmers, T.C. (1992) Getting to grips with Archie Cochrane's agenda. Br. Med. J., 305, 786788.[ISI][Medline]
Cochrane, A.L. (1972) Effectiveness and Efficiency, Random Reflections on Health Services. The Nuffield Provincial Hospitals Trust.
Cochrane, A.L. (1979) 19311971: a critical review, with particular reference to the medical profession. In Teeling Smith, G. (ed.), Medicines for the Year 2000. Office of Health Economics, London, UK, pp. 111.
Collins, J.A. (1997) Clinical research evidence and clinical practice. Hum. Reprod., 12, 18471850.[Abstract]
Crowley, P., Chalmers. I. and Keirse, M.J. (1990) The effects of corticosteroid administration before preterm delivery: an overview of the evidence from controlled trials [see comments]. Br. J. Obstet Gynaecol., 97, 13.[ISI]
Davidoff, F., Haynes, B., Sackett, D. et al. (1995) Evidence Based Medicine: a new journal to help doctors identify the information they need. Br. Med. J., 310, 10851086.
Dickersin, K., Scherer, R. and Lefebvre, C. (1994) Identifying relevant studies for systematic reviews. Br. Med. J., 309, 12861291.
Evers, J.L., deHaas, H.W., Land, J.A. et al. (1998) Treatment independent pregnancy rate in patients with severe reproductive disorders. Hum. Reprod., 13, 12061209.[Abstract]
Hughes, E.G. (1997) The effectiveness of ovulation induction and intrauterine insemination for the treatment of persistent infertility: a meta-analysis. Hum. Reprod., 12, 18651872.[Abstract]
Hughes, E.G., Fedorkow, D.M. and Collins, J.A. (1993) A quantitative overview of controlled trials in endometriosis-associated infertility. Fertil. Steril., 59, 963970.[ISI][Medline]
Hughes, E.G., Fedorkow, D.M., Collins, J.A. et al. (1998) Ovulation suppression versus placebo in the treatment of endometriosis (Cochrane Review). In The Cochrane Library. Issue 4. Update Software, Oxford, UK.
Mulrow, C.D. (1987) The medical review article: state of the science. An. Intern. Med., 106, 485488.[ISI][Medline]
Naylor, C.D. (1995) Gray zones of clinical practice: some limits to evidence based medicine. Lancet, 345, 840842.[ISI][Medline]
Oei, S.G., Helmerhorst, F.M., Bloemenkamp, K.W. et al. (1998) Effectiveness of the postcoital test: randomised controlled trial. Br. Med. J., 317, 502505.
Oxman, A.D. (1994) Checklist for review articles. Br. Med. J., 309, 648651.
Sackett, D.L. (1996) Surveys of self reported reading times of consultants in Oxford, Birmingham, Milton Keynes, Bristol, Leicester and Glasgow. In Rosenberg W.M.C., Richardson, W.S., Haynes, R.B. and Sackett, D.L. (eds), Evidence Based Medicine. Churchill Livingstone, London, UK.
Schatte, E.C., Hirshberg, S.J., Fallick, M.L. et al. (1998) Varicocelectomy improves sperm strict morphology and motility. J. Urol., 160, 13381340.[ISI][Medline]
Schulz, K.F., Chalmers, I., Hayes, R.J et al. (1995) Empirical evidence of bias: diversions of methodological quality associated with estimates of treatment effects in controlled trials. J. Am. Med. Assoc., 273, 408412.[Abstract]
Snick, H.K., Snick, T.S., Evers, J.L. et al. (1997) The spontaneous pregnancy rate on untreated subfertile couples: the Walcheren primary care study. Hum. Reprod., 12, 15821588.[Abstract]
Tzafettas, J., Loufopoulos, A., Stephanatos, A. et al. (1994) Tubal catheterization for intrafallopian insemination and transvaginal gamete (GIFT) or zygote intrafallopian transfer (ZIFT): our experience in a total of 1128 treatment cycles. J. Assist. Reprod. Genet., 11, 283288.[ISI][Medline]
Vandekerckhove, P., O'Donovan, P.A., Lilford, R.J. et al. (1993) Infertility treatment: from cookery to science. The epidemiology of randomized controlled trials. Br. J. Obstet Gynaecol., 100, 10051036.[ISI][Medline]