Identificação de estudos na base de dados Lilacs para revisão sistemática de estudos de acurácia de teste diagnóstico

• Ano de defesa: 2016
• Autor(a) principal: Pereira, Rogerio Aparecido [UNIFESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Paulo (UNIFESP)
• Palavras-chave em Português: Review; Diagnosis; Search engine; Databases; Bibliographic; Acess to information; Revisão; Diagnóstico; Ferramenta de busca; Bases de dados bibliográficas; Acesso à informação
• Link de acesso: https://sucupira.capes.gov.br/sucupira/public/consultas/coleta/trabalhoConclusao/viewTrabalhoConclusao.jsf?popup=true&id_trabalho=4478626; http://repositorio.unifesp.br/handle/11600/47727

reviews of diagnostic test accuracy. Methods: We conducted many steps to build this strategy for LILACS: mapping the available literature about database structure and functionality; adaptation of the search strategy for MEDLINE/Pubmed, published on Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy for LILACS. After, we assessed all search strategies developed for MEDLINE regarding SR of diagnostic test accuracy published on Cochrane. For each SR of diagnostic test accuracy, we built and applied new search strategies, as proposed for LILACS, to test the ability to recover studies in this database. Results: The search strategy construction for SR of diagnostic test accuracy that recovered most studies available at LILACS was done through adaptation of the strategy designed for MEDLINE; adding terms in three languages (Spanish, Portuguese and English); with updates for the search tool and TAGs available in the database, as well as the controlled vocabulary (DeCS), to specify and broaden the result. Forty-three strategies for SR of test diagnostic accuracy were tested and all of them recovered studies through the search in the database. We noticed that 64% of analyzed reviews did not include LILACS in the methods for identification of studies. From the 16 reviews searching LILACS, 87% limited the search terms description to English with the sub-utilization of database resources. Conclusions: The construction of a sensitive search strategy for diagnostic test for LILACS is complex and should be built using TAGs, DeCS, three languages, beyond the base resources. Omission of search in this base or not performing the steps to build a strategy can lead to potential losses of relevant studies.