Article type
Year
Abstract
Introduction/Objective: Unbiased systematic reviews (SRs) should include as many relevant articles as possible. However, the majority incorporate only articles in English and published data, and employ only the MEDLINE database to locate studies. The LILACS database, which indexes Latin American and Caribbean literature, is rarely consulted, yet only 41 of the 670 journals in LILACS are also indexed in either MEDLINE or EMBASE. LILACS could and should be an important source of articles for SRs.
Objectives: To evaluate if a search in LILACS can identify articles that should be part of SRs published in five core medical journals, thereby improving their precision.
Methods: Survey of the published literature. Sample: Systematic reviews published in 1997 in five core medical journals: JAMA, Lancet, NEJM and Annals of Internal Medicine. Procedures: The inclusion criteria of each SR were extracted and a LILACS search performed in order to verify if the database contained any reference that should have been also an alyzed in the SR. If a reference was found, we checked whether it was unknown to the SR. Main outcomes measures: Number of SR for which we could identity LILACS references that should have been analyzed, but were not.
Results: Of the 81 articles were identify, 68 were considered SRs and retrieved. Five reviews were excluded: two included LILACS in the searched databases and three reviewed individual patient data. Of the 63 remaining SRs, for 39 (62%) we identified LILACS articles that could fit the inclusion criteria; for five (8%) our analysis was not conclusive. We found no articles pertaining to the 19 remaining SRs (30%). For 38 SRs, a LILACS search revealed pertinent articles unknown to the SR. In only one SR could the articles identified by LIL ACS be found through methods other than a LILACS consult. Therefore, in 60% (38/63) of the cases, the review could have been improved by a LILACS database search.
Discussion: Searching LILACS improves SRs of the literature. This database must be regarded and employed as an indispensable source of articles
Objectives: To evaluate if a search in LILACS can identify articles that should be part of SRs published in five core medical journals, thereby improving their precision.
Methods: Survey of the published literature. Sample: Systematic reviews published in 1997 in five core medical journals: JAMA, Lancet, NEJM and Annals of Internal Medicine. Procedures: The inclusion criteria of each SR were extracted and a LILACS search performed in order to verify if the database contained any reference that should have been also an alyzed in the SR. If a reference was found, we checked whether it was unknown to the SR. Main outcomes measures: Number of SR for which we could identity LILACS references that should have been analyzed, but were not.
Results: Of the 81 articles were identify, 68 were considered SRs and retrieved. Five reviews were excluded: two included LILACS in the searched databases and three reviewed individual patient data. Of the 63 remaining SRs, for 39 (62%) we identified LILACS articles that could fit the inclusion criteria; for five (8%) our analysis was not conclusive. We found no articles pertaining to the 19 remaining SRs (30%). For 38 SRs, a LILACS search revealed pertinent articles unknown to the SR. In only one SR could the articles identified by LIL ACS be found through methods other than a LILACS consult. Therefore, in 60% (38/63) of the cases, the review could have been improved by a LILACS database search.
Discussion: Searching LILACS improves SRs of the literature. This database must be regarded and employed as an indispensable source of articles