Reviews in Gynaecological Practice 2 (2002) 3
Editorial
Traditional or systematic reviews?
No one can read all the original research literature for even one clinical speciality, and certainly not if they are also actually treating patients. For example, in obstetrics and gynaecology alone, there are no less than 57 specialist journals indexed in Medline, as well as much relevant literature published in general journals. To evaluate the science properly, one would not only have to read all this but also have sufficient expertise to judge the quality, and to integrate the findings both with other studies, and with the larger scientific enterprise. Since this is obviously impossible we need reviews. These fall into two main types. Traditionally writers of clinical reviews were experts in the clinical area being studied. They were chosen for this expertise and wrote accordingly. They knew the major issues for their readers, and would interpret the original articles in the light of their experience. Doctors have kept up to date for hundreds of years by reading such reviews. However, traditional reviews had one major weakness. The reviewers were rarely expert in the science of reviewing, and were sometimes even inexpert in the scientific method itself. They tended to omit important studies or to combine the results of different studies in illogical ways. For example, they may have failed to give sufficiently greater weight to a large well-designed trial than to a small observational study. The solution is systematic reviews, where the reviewer describes the methods of the review just as an author of primary research should do. This permits the reader to judge the review quality. In many areas, typically if important questions can be clearly formulated and there have been well-designed studies, systematic reviews are much more reliable than traditional ones [1]. This is now almost universally recognised, and systematic reviews are now regularly published in many journals. An international organisation, the Cochrane Collaboration (http://www.cochrane.org) creates and regularly updates good quality ones, and the National Health Service keeps a database of quality-assessed ones
(The Database of Abstracts of Reviews of Effectiveness, DARE; http://nhscrd.york.ac.uk/darehp.htm). As a result, it has become much easier for clinicians to ensure that their practice is based on a proper evaluation of the best quality scientific evidence. The editors of Reviews in Gynaecological Practice support these developments and would be delighted to publish more systematic reviews in relevant areas. Nevertheless, there is still a place for clinicians to write traditional reviews. They can interpret the systematic reviews, give guidance when no systematic review has been performed, and indicate how individual patients may differ from those in the systematic reviews. In short, they can describe appropriate clinical practice in the light of the evidence. This is why Reviews in Gynaecological Practice will continue to publish traditional reviews as well as systematic ones. So long as traditional reviewers ensure that they include the evidence of relevant systematic reviews, our readers will get the best of both worlds. The above websites should make this relatively easy. References [1] Antman EM, Lau J, Kupelnick B, Mosteller F, Chalmers TC. A comparison of results of meta-analyses of randomised control trials and recommendations of clinical experts. JAMA 1992;268:240–8.
1471-7697/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 1 4 7 1 - 7 6 9 7 ( 0 2 ) 0 0 0 2 1 - 7
Jim Thornton, MD, FRCOG∗ Professor of Obstetrics and Gynaecology Nottingham City Hospital NHS Trust Nottingham, UK ∗ Tel.: +44-115-962-7914; fax: +44-115-962-7670 E-mail address:
[email protected] (J. Thornton) URL: http://www.iGreens.org.uk