Developing optimal search strategies for retrieval of clinical trials in LILACS database

Article type
Authors
Conceição MA1, Pinto RR2, Conceição MA1
1Universidade Federal de São Paulo (UNIFESP). São Paulo, Brazil
2BIREME. Centro Latino Americano e do Caribe em Informação em Ciências da Saúde, Brazil
Abstract
Background: The inclusion of LILACS (Latin American and Caribbean Health Sciences Literature database) in a search for clinical trials is important to identify studies not included in other major bibliographic databases such as MEDLINE or EMBASE. The identification of clinical trials in LILACS is a challenge, considering the lack of a standardized mechanism for search in this database and the quality of indexing. If the clinical trials are not properly indexed they will not be recovered, and thus, harm the results of important studies such as systematic reviews.
Objectives: To develop optimal search strategies for detecting clinical trials in LILACS taking in consideration the indexing process.
Methods: We have used a sample of 63 studies from LILACS, which were analyzed and classified as clinical trial by two indexers in a previous study with 721 articles from journals of cardiology. We developed the search strategies combining terms of clinical trials (Publication Type) available in DeCS (Health Science Descriptors) and used by indexers for LILACS with other terms present in the title and abstracts fields of the analyzed studies. The strategy has been run in the system and adjusted according to the analyzed studies and subsequently we have calculated the sensitivity and specificity of the search.
Results: This paper presents a specific and sensitive search strategy that has been developed taking for reference a case study that evaluated the clinical trials published in indexed journals of cardiology in LILACS.
Conclusions: From the proposed strategy, it will be possible to perform an optimal search for clinical trials in LILACS in different areas, with a high degree in relevant article retrieval. We expect that this study will contribute to the improvement of retrieval mechanisms of clinical trials in the LILACS database.