부산대학교 도서관

Background: Reporting bias in randomised clinical trials (RCTs) arises from a tendency among researchers to report and journals to publish results that are positive, statistically significant, or that coincide with the hypothesis of the researchers. The most common type of reporting bias is dissemination bias, defined as "the publication or non-publication of the results of research studies depending on the nature and/or direction of the results". Objectives: To develop strategies to prevent and control the dissemination bias of RCTs. Methods: Four studies with different methodologies have been carried out: 1) A survey among European agencies that fund biomedical research to explore their opinion and possible solutions about dissemination bias; 2) An observational and descriptive study to identify and describe the journals that publish original research in Spain and Latin America in Spanish language; 3) An observational and descriptive study to identify, via hand and electronic searching, RCTs published in Spanish Physiotherapy journals, as well as to provide their main characteristics; and 4) The development of a database to facilitate the design, management and analysis of results of handsearching projects, as well as the submission of references of identified RCTs to CENTRAL, the Cochrane Collaboration global database of RCTs. There were four additional supplemental studies: 1) A report that explains the early beginnings and progress of the handsearching of RCTs initiative in Spain and Latin America; as well as 2-4) The results of two observational and descriptive studies for identifying RCTs published in Spanish Gynaecology and Obstetrics journals and in Spain (one publication) and Latin American Dermatology journals (two publications), respectively. Results: 1) Through the survey of European agencies that finance biomedical research, it was found that more than two thirds adopted measures to prevent or control the dissemination bias. However, less than a quarter were aware of the publication status of projects funded in 2005. Based on the overall results of this survey and of other studies carried out within the OPEN project, specific recommendations for preventing and controlling dissemination bias were developed for these agencies. 2) A total of 1498 biomedical journals that publish original research in Spain and Latin America in Spanish were identified. 3) A total of 78 RCTs were retrieved from ten Physiotherapy Spanish journals, none of which were indexed in MEDLINE or EMBASE. 4) BADERI (Database of Iberoamerican Clinical Trials and Journals, by its initials in Spanish) has been developed. BADERI is a free tool that allows the tracking of different aspects of handsearching projects, as well as generating reports for descriptive studies and files specifically designed for exporting references to CENTRAL. To date, there are over 6000 references to RCTs in BADERI. Lastly, through the supplemental studies, 1) the early beginnings, scope and progress of handsearching activities coordinated by the Iberoamerican Cochrane Centre in Spain and Latin America are explained, 2) 230 RCTs were identified in 16 Gynaecology and Obstetrics journals from Spain, and 3-4) 144 RCTs were identified in 21 Dermatology journals from Spain and Latin America. Conclusions: Most funding agencies of biomedical research in Europe do not implement rigorous strategies to verify the dissemination of funded RCTs. Additionally; there is a significant number of biomedical journals that publish original research in Spain and Latin America. However, most of these journals do not have an impact factor or are not indexed in literature databases. Moreover, the number of RCTs published in Spanish Physiotherapy journals is limited; these RCTs present a low methodological quality. Lastly, BADERI has become an invaluable tool for coordinating handsearching projects of RCTs, as well as for facilitating the submission of references to the identified RCTs to CENTRAL.