Contribution of the Cochrane Library to the evidence-based journals
- Cindy Walker-Dilks, MLS,
- for the Evidence-Based Journals Group
Systematic reviews, particularly those published in the Cochrane Library, now form a substantial part of the content of ACP Journal Club, Evidence-Based Medicine, Evidence-Based Nursing, and Evidence-Based Mental Health. This contribution has grown steadily since the first publication of the Cochrane Library. What makes the systematic reviews published in the Cochrane Library so amenable to the evidence-based journals?
A BRIEF HISTORY
Up to the end of 2002, 4 evidence-based abstract journals were produced by the Health Information Research Unit (HIRU) at McMaster University in Hamilton, Ontario, Canada. The first was ACP J Club, begun in 1991. Published bimonthly by the American College of Physicians (ACP), each issue contains 25 research abstracts with clinical commentaries on internal medicine topics. Evidence-Based Medicine, published bimonthly by the BMJ Publishing Group starting in 1995, has a primary care focus including relevant obstetrics and gynaecology, paediatrics, surgery, psychiatry, and general and family practice. Evidence-Based Nursing and Evidence-Based Mental Health, published quarterly by the BMJ, began in 1998.
To fuel the content of these evidence-based publications, a team of 6 research associates reads all issues of about 115 journals, plus the Cochrane Database of Systematic Reviews. All original studies and systematic reviews meeting methodological criteria (see Purpose and Procedure) are flagged and forwarded to the editors of the evidence-based journals to which the content pertains (eg, internal medicine to ACP J Club, nursing interventions to Evidence-Based Nursing). About 3000 original studies and reviews meet criteria per year. From these, the most clinically relevant are chosen by “vote,” and the 4 evidence-based journals abstract up to 414 articles per year across the 4 journals (some articles may be abstracted in >1 journal).
The number of abstracts of reviews has increased substantially since ACP J Club first began publication in 1991. In 1991 the proportion of systematic reviews abstracted was 22%. By 1996 the proportion was 26%, and by 2002 it was 38%.
Systematic reviews meet criteria for inclusion in the evidence-based journals if they contain all of the following:
an identifiable description of methods and sources used for searching for articles
a description of the clinical topic and inclusion/exclusion criteria for selecting articles
⩾1 study that meets criteria for original studies
With respect to the last point, this would be met by a systematic review of a therapeutic intervention, for example, if it contained ⩾1 randomised controlled trial with ⩾80% patient follow up and examined clinically important outcomes.
The Cochrane Library has been published quarterly since 1996. For each issue of the Cochrane Library, >100 new and updated reviews meet inclusion criteria. Virtually all of the reviews pass criteria unless no randomised controlled trials on the content topic are found by the review team. Thus, Cochrane reviews have the potential to represent a substantial amount of evidence-based journal content.
The journal titles that yield the largest proportion of the articles that are abstracted in the 4 evidence-based journals are the New England Journal of Medicine, JAMA, BMJ, Lancet, and the Cochrane Library. The figure shows the relative contributions of the 5 journals, which have changed over the years. Cochrane captured just over 1% of the abstract content in 1996; 7% in 1998, close to the other 4 journals; second highest in 1999 with 11%; and then became the top yielder in 2002 with almost 16% of the abstract content.
Cochrane reviews have steadily increased in abstract share in the evidence-based journals since their first appearance in 1996. The Cochrane Library provides a regularly published source of large numbers of high quality systematic reviews on clinically relevant topics. Both new and updated reviews are of interest, and almost every review meets criteria for abstracting in the evidence-based journals.