Article type
Year
Abstract
Background: The number of randomized controlled trials has increased exponentially over the past 30 years. Although the Cochrane Collaboration’s achievements have been exemplary, the Cochrane Challenge cannot be fulfilled because we are bogged down by unscientific methods of identifying quality randomized trials. This prohibits the periodic update of systematic reviews.
Objectives: To seek consensus on a search strategy to fulfil the Cochrane Challenge.
Methods: The search strategies recommended in the Cochrane Handbook are a combination of handsearching and electronic search. The former is time consuming, expensive and impossible to duplicate and validate the authors' findings. The electronic search, the Optimally Sensitive Search Strategy, is a 29-line algorithm shown to be highly sensitive in identifying randomized trials but lacks specificity. (Chow TKF et al. 2004) We present a simpler, single-line algorithm: (double blind or random).af. We applied this strategy to a systematic review pertaining to post-thoracotomy pain (Joshi GP et al. 2008) to demonstrate its ease of use, high sensitivity in identifying quality randomized trials, high specificity in excluding non-randomised trials and its advantages to updating systematic reviews in a timely fashion.
Results: Applying (double blind or random).af to Ovid EMBASE/MEDLINE®, over the past 30 years (1980 to 2009) demonstrated that the number of randomized trials increased at a near-exponential rate (2.5 times per decade). At the presentation of this paper, there will be over 1.5 million randomized trials listed. Adding the strategy to a post-thoracotomy pain relief algorithm, we demonstrated that comparable conclusions to Joshi’s results could have been achieved by reviewing 866 articles. Furthermore, this review can be updated currently (6 May 2011) by reviewing an additional 230 articles.
Conclusions: (double blind or random).af is reproducible, highly sensitive and specific. If adopted universally, it can manage the impending information overload and fulfil the Cochrane Challenge.
Objectives: To seek consensus on a search strategy to fulfil the Cochrane Challenge.
Methods: The search strategies recommended in the Cochrane Handbook are a combination of handsearching and electronic search. The former is time consuming, expensive and impossible to duplicate and validate the authors' findings. The electronic search, the Optimally Sensitive Search Strategy, is a 29-line algorithm shown to be highly sensitive in identifying randomized trials but lacks specificity. (Chow TKF et al. 2004) We present a simpler, single-line algorithm: (double blind or random).af. We applied this strategy to a systematic review pertaining to post-thoracotomy pain (Joshi GP et al. 2008) to demonstrate its ease of use, high sensitivity in identifying quality randomized trials, high specificity in excluding non-randomised trials and its advantages to updating systematic reviews in a timely fashion.
Results: Applying (double blind or random).af to Ovid EMBASE/MEDLINE®, over the past 30 years (1980 to 2009) demonstrated that the number of randomized trials increased at a near-exponential rate (2.5 times per decade). At the presentation of this paper, there will be over 1.5 million randomized trials listed. Adding the strategy to a post-thoracotomy pain relief algorithm, we demonstrated that comparable conclusions to Joshi’s results could have been achieved by reviewing 866 articles. Furthermore, this review can be updated currently (6 May 2011) by reviewing an additional 230 articles.
Conclusions: (double blind or random).af is reproducible, highly sensitive and specific. If adopted universally, it can manage the impending information overload and fulfil the Cochrane Challenge.
Images
