Letters

Searching for reliable research evidence need not be difficult

EDITORSamanta and Beardsley summarise their search for and use of research evidence to inform their treatment of a patient with low back pain due to sciatica that did not respond to conservative treatment.¹ They are to be applauded for their explicit and conscientious use of research, but we are surprised by their method of searching for relevant systematic reviews and trials.

Instead of searching Medline they could have accessed the relevant data more rapidly and directly by searching the Cochrane Library. This CD Rom contains a suite of databases including the Cochrane Database of Systematic Reviewssystematic reviews carried out under the auspices of the Cochrane Collaboration, abstracts of quality appraised systematic reviews (DARE) prepared by the NHS Centre for Reviews and Dissemination, and probably the world's largest register of controlled clinical trials.

A search of the Cochrane Library 1999 issue 2 using the single term "sciatica" identifies two completed Cochrane reviews and six protocols for systematic reviews. The review of surgery for lumbar disc prolapse concludes that chemonucleosis is more effective than placebo and that it is less invasive but also less effective than discectomy.² The DARE database has references to several other reviews, which come to similar conclusions. In addition there were reviews of the effects and side effects of epidural steroids; although supporting the general safety of this treatment, they are less confident about its longer term benefits.³ There are also 105 trials identified in the controlled trials register, several published in the past few years.

Increasingly we would recommend a more efficient strategy for identifying systematic reviews and clinical trials which starts with a search of the Cochrane Library and Best Evidence⁴ and, since its recent launch, Clinical Evidence.⁵ Only if this strategy does not identify relevant and useful material is it worth investing the considerable time that can be needed to carry out a search of Medline and retrieve the relevant papers.

Trevor A Sheldon, professor. 
Department of Health Studies, University of York, York YO10 5DD tas5{at}york.ac.uk

Julie Glanville, information manager. 
NHS Centre for Reviews and Dissemination, University of York, York YO10 5DD