Article type
Year
Abstract
Background: In low and middle income countries resources available for health care are limited, and it is of particular importance that they are used for things that work. With growing awareness of the importance of systematic reviews great efforts have been made to identify all randomised, or potentially randomised, studies. Much of this effort has been co-ordinated through the Cochrane Collaboration, and the Cochrane Central Register of Controlled Trials (CENTRAL) is acknowledged as the most complete source of randomised trials (RCTs).
Objectives: To estimate what proportion of trials on CENTRAL were conducted in low and middle income countries and to assess how good a source of such studies CENTRAL is.
Methods: Countries were grouped as low or middle income using the World Bank classification. All 427807 citations on The Cochrane Library's CENTRAL (Issue 4, 2004) were transferred to ProCite and searched using country names in English, French and Spanish, where relevant, and tropical disease terms. 39,170 citations (9%) were identified. All citations published in 1991, 1995, and 2000 were selected for electronic and manual checking. In addition the bibliographic databases Chinese Medical Journals (China), IndMED (India), KoreaMED (Korea), LILACS (Latin America), Magyar Orvosi Bibliografia (Hungary), Panteleimon (Ukraine) were searched for these 3 years, and checked whether studies identified were in CENTRAL. Databases of Eastern Mediterranean, Arabic and Thai journals have been identified and will also be searched.
Results: Of 7948 citations in 1991, 1995 and 2000, 1008 (13%) were definitely not an RCT, based on manual inspection of the CENTRAL record. 1289 (16%) had no participants from a low or middle income country. In the target years we found 780 citations on the Chinese Medical Journals database that were not in CENTRAL, 301 on LILACS, 97 on KoreaMED, 65 on IndMED, 57 on Panteleimon, and 4 on Magyar Orvosi Bibliografia. Therefore, of the 5848 citations identified to date, 4544 (78%) are on CENTRAL.
Conclusions: Based on these preliminary data, 5% of citations on CENTRAL are of RCTs or possible RCTs from low and middle income countries. CENTRAL is a considerable source of such trials, but is far from comprehensive.
Objectives: To estimate what proportion of trials on CENTRAL were conducted in low and middle income countries and to assess how good a source of such studies CENTRAL is.
Methods: Countries were grouped as low or middle income using the World Bank classification. All 427807 citations on The Cochrane Library's CENTRAL (Issue 4, 2004) were transferred to ProCite and searched using country names in English, French and Spanish, where relevant, and tropical disease terms. 39,170 citations (9%) were identified. All citations published in 1991, 1995, and 2000 were selected for electronic and manual checking. In addition the bibliographic databases Chinese Medical Journals (China), IndMED (India), KoreaMED (Korea), LILACS (Latin America), Magyar Orvosi Bibliografia (Hungary), Panteleimon (Ukraine) were searched for these 3 years, and checked whether studies identified were in CENTRAL. Databases of Eastern Mediterranean, Arabic and Thai journals have been identified and will also be searched.
Results: Of 7948 citations in 1991, 1995 and 2000, 1008 (13%) were definitely not an RCT, based on manual inspection of the CENTRAL record. 1289 (16%) had no participants from a low or middle income country. In the target years we found 780 citations on the Chinese Medical Journals database that were not in CENTRAL, 301 on LILACS, 97 on KoreaMED, 65 on IndMED, 57 on Panteleimon, and 4 on Magyar Orvosi Bibliografia. Therefore, of the 5848 citations identified to date, 4544 (78%) are on CENTRAL.
Conclusions: Based on these preliminary data, 5% of citations on CENTRAL are of RCTs or possible RCTs from low and middle income countries. CENTRAL is a considerable source of such trials, but is far from comprehensive.
PDF