Limits...
Time to update and quantitative changes in the results of cochrane pregnancy and childbirth reviews.

Jaidee W, Moher D, Laopaiboon M - PLoS ONE (2010)

Bottom Line: Potential factors associated with time to update were assessed using the Cox proportional hazard model.Most reviews were updated less frequently than recommended by Cochrane policy, but few updates had revised conclusions.Prescribed time to update should be reconsidered to support improved decision-making while making efficient use of limited resources.

View Article: PubMed Central - PubMed

Affiliation: Department of Public Health Foundation, Faculty of Public Health, Burapha University, Chonburi, Thailand.

ABSTRACT

Background: The recommended interval between updates for systematic reviews included in The Cochrane Library is 2 years. However, it is unclear whether this interval is always appropriate. Whereas excessive updating wastes time and resources, insufficient updating allows out-of-date or incomplete evidence to guide clinical decision-making. We set out to determine, for Cochrane pregnancy and childbirth reviews, the frequency of updates, factors associated with updating, and whether updating frequency was appropriate.

Methodology/principal findings: Cochrane pregnancy and childbirth reviews published in Issue 3, 2007 of the Cochrane Database of Systematic Reviews were retrieved, and data were collected from their original and updated versions. Quantitative changes were determined for one of the primary outcomes (mortality, or the outcome of greatest clinical significance). Potential factors associated with time to update were assessed using the Cox proportional hazard model. Among the 101 reviews in our final sample, the median time before the first update was 3.3 years (95% CI 2.7-3.8). Only 32.7% had been updated within the recommended interval of 2 years. In 75.3% (76/101), a median of 3 new trials with a median of 576 additional participants were included in the updated versions. There were quantitative changes in 71% of the reviews that included new trials (54/76): the median change in effect size was 18.2%, and the median change in 95% CI width was 30.8%. Statistical significance changed in 18.5% (10/54) of these reviews, but conclusions were revised in only 3.7% (2/54). A shorter time to update was associated with the same original review team at updating.

Conclusions/significance: Most reviews were updated less frequently than recommended by Cochrane policy, but few updates had revised conclusions. Prescribed time to update should be reconsidered to support improved decision-making while making efficient use of limited resources.

Show MeSH
Flow diagram of selection of reviews for analysis.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2903478&req=5

pone-0011553-g001: Flow diagram of selection of reviews for analysis.

Mentions: From the retrospective cohort of 381 PCG reviews we excluded 105: of these, 37 had been withdrawn from the Cochrane database, and 68 were still protocols. There were 276 completed PCG reviews in the 2007 volume, Issue 3. Of these, 111 were updates according to our criteria, i.e. the latest versions of the study cohort were different from those of the first published version. However, examination of the full texts revealed errors in 10 reviews in the dates of first publication and/or last amendment. Our analysis was therefore limited to 101 updated reviews (Fig. 1).


Time to update and quantitative changes in the results of cochrane pregnancy and childbirth reviews.

Jaidee W, Moher D, Laopaiboon M - PLoS ONE (2010)

Flow diagram of selection of reviews for analysis.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2903478&req=5

pone-0011553-g001: Flow diagram of selection of reviews for analysis.
Mentions: From the retrospective cohort of 381 PCG reviews we excluded 105: of these, 37 had been withdrawn from the Cochrane database, and 68 were still protocols. There were 276 completed PCG reviews in the 2007 volume, Issue 3. Of these, 111 were updates according to our criteria, i.e. the latest versions of the study cohort were different from those of the first published version. However, examination of the full texts revealed errors in 10 reviews in the dates of first publication and/or last amendment. Our analysis was therefore limited to 101 updated reviews (Fig. 1).

Bottom Line: Potential factors associated with time to update were assessed using the Cox proportional hazard model.Most reviews were updated less frequently than recommended by Cochrane policy, but few updates had revised conclusions.Prescribed time to update should be reconsidered to support improved decision-making while making efficient use of limited resources.

View Article: PubMed Central - PubMed

Affiliation: Department of Public Health Foundation, Faculty of Public Health, Burapha University, Chonburi, Thailand.

ABSTRACT

Background: The recommended interval between updates for systematic reviews included in The Cochrane Library is 2 years. However, it is unclear whether this interval is always appropriate. Whereas excessive updating wastes time and resources, insufficient updating allows out-of-date or incomplete evidence to guide clinical decision-making. We set out to determine, for Cochrane pregnancy and childbirth reviews, the frequency of updates, factors associated with updating, and whether updating frequency was appropriate.

Methodology/principal findings: Cochrane pregnancy and childbirth reviews published in Issue 3, 2007 of the Cochrane Database of Systematic Reviews were retrieved, and data were collected from their original and updated versions. Quantitative changes were determined for one of the primary outcomes (mortality, or the outcome of greatest clinical significance). Potential factors associated with time to update were assessed using the Cox proportional hazard model. Among the 101 reviews in our final sample, the median time before the first update was 3.3 years (95% CI 2.7-3.8). Only 32.7% had been updated within the recommended interval of 2 years. In 75.3% (76/101), a median of 3 new trials with a median of 576 additional participants were included in the updated versions. There were quantitative changes in 71% of the reviews that included new trials (54/76): the median change in effect size was 18.2%, and the median change in 95% CI width was 30.8%. Statistical significance changed in 18.5% (10/54) of these reviews, but conclusions were revised in only 3.7% (2/54). A shorter time to update was associated with the same original review team at updating.

Conclusions/significance: Most reviews were updated less frequently than recommended by Cochrane policy, but few updates had revised conclusions. Prescribed time to update should be reconsidered to support improved decision-making while making efficient use of limited resources.

Show MeSH