Can reviews be both quick and evidence-based?
Posted on 5th August 2013 by Alice Buchan
Like all things, EBM can be regarded as a work in progress. Cochrane reviews are considered by many to be the gold standard in methodologically sound meta-analysis [1]. Unfortunately, as pointed out by Jon Brassey of Trip database (reviewed here), they are both expensive and time-consuming (to the point where a review may be out of date by the time it is published) [2]. Brassey argues that these factors mean current Cochrane methodology, however good, is simply unsustainable.
Brassey isn’t the only person asking whether there is a quick way of getting evidence-based answers – so has Mona Nasser on her blog [3]. One of her key points is that whilst protocols for ‘full’ Cochrane and other systematic reviews are well established, this is not the case for rapid reviews. Evidence-based answers can (in many cases) be found if you use her five scenarios, and make appropriate use of the existing systematic reviews and literature where available. These methods are likely to be really useful for answering defined clinical questions you may have.
The time and money issues raised earlier are more tricky to solve than the need for a quick way of finding an evidence-based answer, as instead of using the existing systematic reviews, it requires a change in the way reviews themselves are done. This poses the issue of how to get the methodological quality of a Cochrane review as quickly and cheaply as possible. Some of the methods suggested include using the largest trial, and including only a selection of new trials in an updated review [2]. One exciting possibility, though in its very early stages, is using statistics to create an overall ‘score’ for included studies, with scores corresponding to a narrative conclusion, as seen here in the Trip blog [4]. These ultra-rapid reviews all took 5 minutes or less, and several corresponded fairly well to the appropriate Cochrane review. Personally, one of my major concerns with this approach is that the hand selection of relevant trials is unlikely to be as systematic (or to accurately include and exclude studies) as current methods (though there is no claim that this is the case).
Nonetheless, I think that this project is very exciting, and will be watching as it progresses. These 5-minute or less reviews might be a useful tool to use when a quick answer is needed, and in scenario 1 in Nasser’s blog (lots of studies, but no systematic review), might be a useful tool to provide guidance in the absence of a Cochrane or other systematic review. If methodologically sound methods are used to choose the included studies, something like this tool may be useful to speed up the writing or updating of systematic reviews, or provide an interim result at an early stage of the review process.
Links:
[1] University of Virginia Health System: Claude Moore Health Sciences Library. Cochrane Database of Systematic Reviews: Navigating the maze. https://www.hsl.virginia.edu/collections/ebm/aboutcochrane.cfm (source no longer available August 2019).
[2] Brassey, J. A critique of the Cochrane collaboration. April 2013.
[3] Nasser M. Can quick answers be evidence-informed? July 2013. http://monanasser.wordpress.com/2013/07/05/can-quick-answers-be-evidence-informed/
[4] Brassey J. Ultra-rapid reviews, first test results. August 2013. http://blog.tripdatabase.com/2013/08/ultra-rapid-reviews-first-test-results.html
No Comments on Can reviews be both quick and evidence-based?