Article type
Year
Abstract
Objectives: To identify searching bias in the Cochrane reviews on peripartum antiretroviral therapies (ART) for preventing the risk of mother-to-child transmission (MTCT) of HIV.
Methods: The databases of Cochrane Library (2007 issue 1), PubMed, EMBASE, CINAHL, AIDSearch, AIDS LINE, AIDSTRIALS, AIDSDRUGS, AIDS info and the Chinese Databases (CBM, CNKI, VIP) were searched from the beginning of the databases to April 30, 2007. Cochrane reviews and randomized controlled trials (RCTs) which assessed the efficacy of ART for preventing MTCT were selected. We compared the sources, search strategies and the eligibility criteria of included Cochrane reviews with that of our own similar research to determine whether, and to what extent, the searching and selection bias in the Cochrane reviews existed.
Results: Two Cochrane reviews (Brocklehurst 2002 & 2005) were eligible, including 7 and 8 RCTs on ART for preventing MTCT of HIV respectively. They searched only the Pregnancy and Childbirth Group Specialised Register and Cochrane Controlled Trials Register. Conference abstracts from the International AIDS Conferences and Conference on Retroviruses and Opportunistic Infections were also searched in the second review. But some general databases (eg. MEDLINE, EMBASE, Psychlit, CINAHL, etc.) and some special databases (eg. AIDSearch, AIDSLINE, AIDSTRIALS, AIDSDRUGS, etc.) were not searched. Although the date of most recent amendment was June 2 2005, all the included trials were updated to 2001. Only one abstract was included in the trial, PACTG 316. When we searched and selected the trials using similar search strategies and eligible criteria, there were 15 RCTs published after that, including the full articles of PACTG 316. About 30 ongoing trials were registered in Clinical trial registers.
Conclusions: Searching bias was found in the Cochrane reviews on peripartum antiretroviral therapies for preventing the risk of mother-to-child transmission of HIV. Cochrane reviews should be updated carefully and substantively, especially in the important fields.
Methods: The databases of Cochrane Library (2007 issue 1), PubMed, EMBASE, CINAHL, AIDSearch, AIDS LINE, AIDSTRIALS, AIDSDRUGS, AIDS info and the Chinese Databases (CBM, CNKI, VIP) were searched from the beginning of the databases to April 30, 2007. Cochrane reviews and randomized controlled trials (RCTs) which assessed the efficacy of ART for preventing MTCT were selected. We compared the sources, search strategies and the eligibility criteria of included Cochrane reviews with that of our own similar research to determine whether, and to what extent, the searching and selection bias in the Cochrane reviews existed.
Results: Two Cochrane reviews (Brocklehurst 2002 & 2005) were eligible, including 7 and 8 RCTs on ART for preventing MTCT of HIV respectively. They searched only the Pregnancy and Childbirth Group Specialised Register and Cochrane Controlled Trials Register. Conference abstracts from the International AIDS Conferences and Conference on Retroviruses and Opportunistic Infections were also searched in the second review. But some general databases (eg. MEDLINE, EMBASE, Psychlit, CINAHL, etc.) and some special databases (eg. AIDSearch, AIDSLINE, AIDSTRIALS, AIDSDRUGS, etc.) were not searched. Although the date of most recent amendment was June 2 2005, all the included trials were updated to 2001. Only one abstract was included in the trial, PACTG 316. When we searched and selected the trials using similar search strategies and eligible criteria, there were 15 RCTs published after that, including the full articles of PACTG 316. About 30 ongoing trials were registered in Clinical trial registers.
Conclusions: Searching bias was found in the Cochrane reviews on peripartum antiretroviral therapies for preventing the risk of mother-to-child transmission of HIV. Cochrane reviews should be updated carefully and substantively, especially in the important fields.