Investing in updating: how do conclusions change when Cochrane systematic reviews are updated?

BMC Med Res Methodol. 2005 Oct 14:5:33. doi: 10.1186/1471-2288-5-33.

Abstract

Background: Cochrane systematic reviews aim to provide readers with the most up-to-date evidence on the effects of healthcare interventions. The policy of updating Cochrane reviews every two years consumes valuable time and resources and may not be appropriate for all reviews. The objective of this study was to examine the effect of updating Cochrane systematic reviews over a four year period.

Methods: This descriptive study examined all completed systematic reviews in the Cochrane Database of Systematic Reviews (CDSR) Issue 2, 1998. The latest version of each of these reviews was then identified in CDSR Issue 2, 2002 and changes in the review were described. For reviews that were updated within this time period and had additional studies, we determined whether their conclusion had changed and if there were factors that were predictive of this change.

Results: A total of 377 complete reviews were published in CDSR Issue 2, 1998. In Issue 2, 2002, 14 of these reviews were withdrawn and one was split, leaving 362 reviews to examine for the purpose of this study. Of these reviews, 254 (70%) were updated. Of these updated reviews, 23 (9%) had a change in conclusion. Both an increase in precision and a change in statistical significance of the primary outcome were predictive of a change in conclusion of the review.

Conclusion: The concerns around a lack of updating for some reviews may not be justified considering the small proportion of updated reviews that resulted in a changed conclusion. A priority-setting approach to the updating of Cochrane systematic reviews may be more appropriate than a time-based approach. Updating all reviews as frequently as every two years may not be necessary, however some reviews may need to be updated more often than every two years.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Bibliometrics
  • Confidence Intervals
  • Data Interpretation, Statistical
  • Databases, Bibliographic*
  • Decision Making
  • Evidence-Based Medicine / standards*
  • Humans
  • Meta-Analysis as Topic*
  • Reproducibility of Results
  • Review Literature as Topic
  • Time Factors