Article type
Year
Abstract
Background: Search strategies for systematic reviews (SRs) are designed
to identify all relevant articles concerning a specific question. LILACS is a
database which contains articles from Latin-American and Caribbean
countries. Previous studies showed that searching LILACS could improve
SRs published in core medical journals, finding relevant articles in more
than 60% of them [1,2]. These conclusions, however, should be evaluated
in a context of medical specialties. High quality trials in complex specialties,
such as oncology, are less likely to be published in a journal indexed only in
LILACS. In these cases, some reviewers could consider searches in minor
databases not worthy, since no relevant trials would be identified, and
resources would be spent without significant results. Objectives: To
evaluate the use of LILACS as a source of articles that could be included in
Cochrane SRs about oncology. Methods: We analyzed all Cochrane SRs
related to cancer published in 2007. For each SR, a structured and highly
sensitive search strategy for identification of randomized controlled trials
(RCTs) in LILACS [3] was performed. The additional search strategy was
adapted from each SR’s reported strategies for MEDLINE and EMBASE. Full
text of potential RCTs were obtained and evaluated separately by the two
reviewers. Results: We found 18 SRs about treatment or clinical
management of patients with cancer. Only three of them originally
searched LILACS. The searches did not reveal pertinent articles that could
have data to be included in any of the 18 SRs. Conclusion: Searching
LILACS is probably irrelevant to SRs in oncology. This statement was based
on SRs published in 2007. Further evaluation of all Cochrane SRs published
in oncology may find more accurate results. LILACS does not need to be
considered as an indispensable source of articles. Future discussion about
importance of this database in SRs must consider the medical specialty
involved in the question. References: [1] Clark OA, Castro AA. Int J
Epidemiol. 2002 Feb;31(1):112-4. [2] Manrıquez J, et al. 15th
Cochrane Colloquium. Sao Paolo, Brazil. 2007. [3] Manriquez JJ. J Clin Epidemiol
2008;61(4):407-11.
to identify all relevant articles concerning a specific question. LILACS is a
database which contains articles from Latin-American and Caribbean
countries. Previous studies showed that searching LILACS could improve
SRs published in core medical journals, finding relevant articles in more
than 60% of them [1,2]. These conclusions, however, should be evaluated
in a context of medical specialties. High quality trials in complex specialties,
such as oncology, are less likely to be published in a journal indexed only in
LILACS. In these cases, some reviewers could consider searches in minor
databases not worthy, since no relevant trials would be identified, and
resources would be spent without significant results. Objectives: To
evaluate the use of LILACS as a source of articles that could be included in
Cochrane SRs about oncology. Methods: We analyzed all Cochrane SRs
related to cancer published in 2007. For each SR, a structured and highly
sensitive search strategy for identification of randomized controlled trials
(RCTs) in LILACS [3] was performed. The additional search strategy was
adapted from each SR’s reported strategies for MEDLINE and EMBASE. Full
text of potential RCTs were obtained and evaluated separately by the two
reviewers. Results: We found 18 SRs about treatment or clinical
management of patients with cancer. Only three of them originally
searched LILACS. The searches did not reveal pertinent articles that could
have data to be included in any of the 18 SRs. Conclusion: Searching
LILACS is probably irrelevant to SRs in oncology. This statement was based
on SRs published in 2007. Further evaluation of all Cochrane SRs published
in oncology may find more accurate results. LILACS does not need to be
considered as an indispensable source of articles. Future discussion about
importance of this database in SRs must consider the medical specialty
involved in the question. References: [1] Clark OA, Castro AA. Int J
Epidemiol. 2002 Feb;31(1):112-4. [2] Manrıquez J, et al. 15th
Cochrane Colloquium. Sao Paolo, Brazil. 2007. [3] Manriquez JJ. J Clin Epidemiol
2008;61(4):407-11.