Article type
Year
Abstract
Background: The quest to find gold nuggets with a gold pan is a time-consuming but rewarding task. Imagine the value of a pan that found all gold nuggets without having to sift through as much dirt. The current pan used to find gold nuggets for a systematic review is the Highly Sensitive Search Strategy (HSSS) developed by The Cochrane Collaboration to detect randomized controlled trials. This "gold pan" is sensitive but this comes at the cost of precision, producing expansive results which require time and resources to screen, a lot of dirt to be sifted.
Objective: Our aim was to devise an RCT filter that retains sensitivity but increases precision to reduce title searching during a systematic review (a better "gold pan").
Methods: Our Newcastle RCT filter was developed and tested against 4 published systematic reviews in the Cochrane library that used the Cochrane filter (set 1). After revision, the filter was re-tested on these reviews and a further 4 reviews (set 2). The Newcastle RCT filter was tested against the Cochrane filter in Medline and Embase and against "no filter" in Central.
Results: Our new gold pan missed three trials from set 1. Indexing was checked and the "gold pan" was revised. The filter was re-tested on set 2 where it identified all included trials. All trials from set 1 were also identified with the revised version of the gold pan. Preliminary results showed increases in precision of 81% (N=6) for Medline, 166% (N=6) for Embase and 12% (N=5) for Central.
Conclusions: The process of identifying RCTs for a systematic review can be lengthy. The Cochrane filter identifies both RCTs and CCTs, but it is sometimes desirable to capture only RCTs for a review. We have developed a filter that when compared with the Cochrane filter retains sensitivity to RCTs with increased precision, preserving time and resources of the reviewer when conducting a systematic review. The Newcastle RCT filter "gold pan" has been developed to ensure you find your gold with less sifting. Gold miners are advised to keep watch as this new pan is tested further.
Objective: Our aim was to devise an RCT filter that retains sensitivity but increases precision to reduce title searching during a systematic review (a better "gold pan").
Methods: Our Newcastle RCT filter was developed and tested against 4 published systematic reviews in the Cochrane library that used the Cochrane filter (set 1). After revision, the filter was re-tested on these reviews and a further 4 reviews (set 2). The Newcastle RCT filter was tested against the Cochrane filter in Medline and Embase and against "no filter" in Central.
Results: Our new gold pan missed three trials from set 1. Indexing was checked and the "gold pan" was revised. The filter was re-tested on set 2 where it identified all included trials. All trials from set 1 were also identified with the revised version of the gold pan. Preliminary results showed increases in precision of 81% (N=6) for Medline, 166% (N=6) for Embase and 12% (N=5) for Central.
Conclusions: The process of identifying RCTs for a systematic review can be lengthy. The Cochrane filter identifies both RCTs and CCTs, but it is sometimes desirable to capture only RCTs for a review. We have developed a filter that when compared with the Cochrane filter retains sensitivity to RCTs with increased precision, preserving time and resources of the reviewer when conducting a systematic review. The Newcastle RCT filter "gold pan" has been developed to ensure you find your gold with less sifting. Gold miners are advised to keep watch as this new pan is tested further.