Search filter accuracy to LILACS (via iAHx 2.6-5 interface) for retrieving Randomized Controlled Trials by Brazilian Cochrane Centre

Article type
Authors
Puga MES1, Silva V1, Soares BGDO1, Atallah A1
1Brazilian Cochrane Centre - Discipline of Emergency Medicine and Evidence-Based Medicine, Federal University of Sao Paulo (UNIFESP), São Paulo, Brazil
Abstract
Background: The translation and validation of a clinical trial filter for identifying studies scientifically correct and clinically relevant in the bibliographic index LILACS is critical to use in the systematic reviews.

Objective: To assess the accuracy of the new translation of the search filter for retrieving clinical trials in the LILACS in the new interface iAHx 2.6–5.

Methods: Stages were completed to translate and identificate the new terms found in other works by the gold standard handsearch and therefore incorporated into this new version. The chosen methodology as an alternative to handsearch was the Relative Recall. Analysis was done in the references of Cochrane Systematic Reviews that used the LILACS bibliographic index. The search using the words ‘LILACS and Cochrane’ was performed in PubMed and the full text of Systematic Reviews was retrieved in the Cochrane Library through the free version for Brazilians at http://cochrane.bireme.br/.

Results: After searching the literature, we found that none of the versions available are translated to the new interface and also are not validated. The stages of identifying terminology which formed a set ‘gold standard’ for the new filter were performed considering the development and assessment based in performance (sensitivity), specificity, and accuracy.

Conclusions: We established a new accurate search filter for retrieving clinical trials in the LILACS (via iAHx 2.6–5 interface).