Article type
Year
Abstract
Background: systematic reviews (SRs) are widely considered as the best available evidence to inform health decisions. Because of their increasing recognition, and also for other reasons, the number of reviews is growing fast and the difficulty of identifying all existing reviews for a given question is increasing accordingly.
The Epistemonikos database compiles evidence from different sources and is now the largest database of SRs. However, it is not clear how comprehensive it is.
Objectives: to estimate the recall (sensitivity) of Epistemonikos to identify SRs.
Methods: in order to estimate the recall of Epistemonikos we used a sample of overviews of SRs retrieved from PubMed, published during the first trimester of 2019. We used all of the SR included in these overviews as our gold standard and evaluated whether they were part of the Epistemonikos database. For SRs not found in Epistemonikos, we explored the reasons for their absence.
Results: our search strategy retrieved 2311 records, of which 82 fulfilled our definition of an overview. We have already analyzed 54 overviews (the rest are awaiting assessment and will be presented at the Colloquium).
The total number of SRs included in the overviews was 843 (average 16.3). The Epistemonikos database had 778 (92.3%) of these included reviews. The main reason for the absence of a review in Epistemonikos was the provenance from a database not routinely searched by Epistemonikos (48 reviews), mostly from Chinese databases. Only 17 reviews (2%) were missed because of not being detected by the search strategies and algorithms of the Epistemonikos database.
Conclusions: Epistemonikos includes the vast majority of systematic reviews considered by overviews of systematic reviews, which is a good proxy for the comprehensiveness of this database. The added value of conducting searches in multiple databases to identify the few reviews not included in Epistemonikos needs to be balanced against the substantial amount of time and resources needed to do this. One exception might be topics where Chinese literature might be more prominent. We think it is reasonable to state that, for most questions, the Epistemonikos database constitutes a one-stop shop for systematic reviews.
The Epistemonikos database compiles evidence from different sources and is now the largest database of SRs. However, it is not clear how comprehensive it is.
Objectives: to estimate the recall (sensitivity) of Epistemonikos to identify SRs.
Methods: in order to estimate the recall of Epistemonikos we used a sample of overviews of SRs retrieved from PubMed, published during the first trimester of 2019. We used all of the SR included in these overviews as our gold standard and evaluated whether they were part of the Epistemonikos database. For SRs not found in Epistemonikos, we explored the reasons for their absence.
Results: our search strategy retrieved 2311 records, of which 82 fulfilled our definition of an overview. We have already analyzed 54 overviews (the rest are awaiting assessment and will be presented at the Colloquium).
The total number of SRs included in the overviews was 843 (average 16.3). The Epistemonikos database had 778 (92.3%) of these included reviews. The main reason for the absence of a review in Epistemonikos was the provenance from a database not routinely searched by Epistemonikos (48 reviews), mostly from Chinese databases. Only 17 reviews (2%) were missed because of not being detected by the search strategies and algorithms of the Epistemonikos database.
Conclusions: Epistemonikos includes the vast majority of systematic reviews considered by overviews of systematic reviews, which is a good proxy for the comprehensiveness of this database. The added value of conducting searches in multiple databases to identify the few reviews not included in Epistemonikos needs to be balanced against the substantial amount of time and resources needed to do this. One exception might be topics where Chinese literature might be more prominent. We think it is reasonable to state that, for most questions, the Epistemonikos database constitutes a one-stop shop for systematic reviews.
PDF