The Cochrane Library 1998 Issue 4

British Medical Journal, March 27, 1999 by Albert G Mulley

The Cochrane Library 1998 Issue 4

Update Software, 141 [pounds sterling] for individual subscription, updated quarterly ISBN 0 72791089 2

Rating: ****

Illness and its treatment are often transforming experiences. It follows that decisions about treatment can be fateful for patients. Such decisions ought to be informed by the best evidence regarding the effectiveness of available interventions. However, timely access to the best evidence is elusive. It requires an organisational and analytical enterprise so prodigious that, until now, it has been beyond the grasp of individual clinicians and the profession as a whole. The Cochrane Library represents a formidable attempt to provide a response worthy of the vision of effective and efficient health care articulated decades ago by the project's namesake, the late A L Cochrane. He would be proud of the library that carries his name.

The Cochrane Library consists of (1) the Cochrane Database of Systematic Reviews, currently containing 438 protocols for reviews in progress and 481 completed reviews, including many that address problems commonly seen in general practice; (2) the Database of Abstracts of Reviews of Effectiveness, containing nearly 2000 abstracts of evidence based reviews, some with commentaries about their quality; (3) the Cochrane Controlled Trials Register, with more than 200 000 trials listed; and (4) the Cochrane Review Methodology Database, with 836 citations and abstracts that address methods for unbiased collection and interpretation of evidence. For most users, the completed systematic reviews will be most valuable.

Because medicine's knowledge base is constantly expanding, any useful compendium must be cumulative. This is the fourth issue of 1998, and revised issues are published quarterly. But the library is a work in progress in other ways. Updating is uneven. The coverage of different conditions and of various treatments for the same condition varies. The level of derail is more likely to reflect the interests of those committed to the enterprise than the burden of illness conferred by a condition or other measures of priority. For example, about a third of the completed reviews address evidence regarding care related to pregnancy and childbirth.

The formats used to present and catalogue reviews reflect more the mindset of a researcher or a librarian than that of a user--doctor, patient, or policy maker--trying to make an evidence based decision. But the search and display functions are well designed and easy to learn. The library will not always inform decisions, however, even if relevant reviews have been completed. It is, after all, a compendium of available evidence, and much of what needs to be known to make a good decision has not been well studied. Moreover, good clinical decisions will often depend on preferences for possible outcomes and attitudes toward risks, which will vary considerably among users of evidence.

These are quibbles, and one would expect these limitations of a work in progress produced by a collaboration convened across national and disciplinary boundaries and sustained by a common view of professional responsibility. Furthermore, to cite limitations of The Cochrane Library begs a comparison, and there is none to make. Use today's library despite its gaps. Better yet, find a way to help the Cochrane Collaboration make tomorrow's physicians and patients realise the promise of effectiveness and efficiency in health care.

Albert G Mulley, associate professor of medicine, Harvard Medical School, Boston, Massachusetts, USA