The Cochrane Empty Review Project: Prevalence and characteristics of systematic reviews without included studies in The Cochrane Database of Systematic Reviews

Article type
Authors
Shepard L1, Yaffe J1, Montgomery P2, Hopewell S3
1University of Utah, USA
2University of Oxford, UK
3UK Cochrane Centre and University of Oxford, UK
Abstract
Background: Little is known about systematic reviews that report no included studies. These so-called 'empty reviews' have been primarily discussed by clinicians and policy-makers expressing frustration with the lack of implications for decision-making in these reviews.

Objectives: The Cochrane Empty Review Project sought to describe the incidence, prevalence, and characteristics of systematic reviews without included studies in the Cochrane Database of Systematic Reviews (CDSR).

Methods: Empty reviews within the CDSR as of 15 August 2010 were identified, extracted, and coded for analysis. We examined the numbers and proportions of empty reviews, date of first publication, and recency of the last update by Cochrane Review Group (CRG). Number of excluded studies, studies awaiting assessment, and on-going studies listed within empty reviews were also examined.

Results: The Cochrane Empty Review Project identified 376 empty reviews in The CDSR, representing 8.7% of all active reviews. Forty-five (86.5%) of the 52 CRGs sustain at least one empty review, ranging from 0 to 35 (0 to 26.9%) per CRG. The average time since first publication is 4.2years (SD 3.4). These empty reviews were updated within an average of 2.8years (SD 2.2), with a range of 0 to 11years since update. The identified reviews listed an average of 9.6 excluded studies (SD 14.5), but varied from 0 to 124 excluded studies listed. Empty reviews typically contained few, if any, studies awaiting assessment or on going studies, with respective averages of 0.1 (SD 0.68) and 0.14 (SD 0.49) per review.

Conclusions: Reporting and updating of empty reviews appears to vary widely. This may or may not be explained by differences in editorial policy or practice across CRGs. Further analysis is warranted and development of more explicit guidance for the reporting of systematic reviews without included studies in The CDSR is recommended.