The process of data extraction for a systematic review involves recording specific information from each of the studies. This information is entered into a paper or electronic form that has been carefully constructed.
The construction of the data extraction form can be done in one of many ways. However, it should be designed to contain information that is directly related to the research question being posed.
Once the data extraction form has been constructed, it’s recommended that it be pilot tested by several reviewers using a sample group of research studies.1,2 This will aid in ensuring the right amount of data is being collected. Too much information can be time consuming to record, and too little can exclude important information requiring the reviewers to re-extract the data.
Although forms can be similar across different reviews, they should be adapted and modified to fit your specific review. In general, the data extraction forms should gather some basic information concerning who is doing the data extraction, bibliographic information of the study being reviewed, and the source of the information.
The following example outlines what could be included in a data extraction form for a review of effectiveness studies adapted from the Centre for Reviews and Dissemination2:
Coding format and instructions for coders
UHN Library and Information Services - For information about this web site, contact Ronald MacPherson at 416-340-4800 x4279 or by e-mail Ronald.MacPherson@uhn.ca