7.6.4  Extracting data from multiple reports of the same study

Studies are frequently reported in more than one publication (Tramèr 1997, von Elm 2004). However, the unit of interest in a Cochrane Intervention review is the study and not the report. Thus, information from multiple reports needs to be collated. It is not appropriate to discard any report of an included study, since it may contain valuable information not included in the primary report. Review authors will need to decide between two strategies:

The choice of which strategy to use will depend on the nature of the reports and may vary across studies and across reports. For example, if a full journal article and multiple conference abstracts are available, it is likely that the majority of information will be obtained from the journal article, and completing a new data collection form for each conference abstract may be a waste of time. Conversely, if there are two or more detailed journal articles, perhaps relating to different periods of follow-up, then it is likely to be easier to perform data extraction separately for these articles and collate information from the data collection forms afterwards.


Drawing flow diagrams for participants in a study, such as those recommended in the CONSORT Statement (Moher 2001), can be particularly helpful when collating information from multiple reports.