9.2.9 Data collection

Data collection is the procedure for extracting relevant details and data from the included systematic reviews and meta-analyses for the Umbrella Review. To avoid risk of bias, the standardized JBI data extraction tools (see Appendix 10.3 of this chapter) should be used to extract the data from the included reviews. Reviewers should have discussed and piloted its use prior to launching into extraction of data for the Umbrella Review to maximize consistency and the likelihood that the relevant results are being identified and detailed sufficiently for the purposes of reporting in the Umbrella Review. Without some discussion and piloting, reviewers may interpret fields in the tool or their relevance to the Umbrella Review questions slightly differently; differences unearthed at the completion of extraction for the review will invariably create more, unnecessary work for the review team. Any additions or modifications to the data extraction tool that are demanded by the nature of review question should be reviewed through by all reviewers and discussed in detail before extracting the data independently. Any additions or modifications should be identified and submitted with the review protocol and approved for publication in the JBI Evidence Synthesis prior to use by any reviewer.

Guided by the data extraction tool, information regarding the citation details, the objectives of the included review, the participants, the setting and context, the number of databases sourced and searched, the date range of database searching, the date range of included studies that inform each outcome of interest, the number/types of studies/country of origin of primary research studies in the included research synthesis, the instrument used to appraise the primary studies in the research synthesis and the rating of their quality, the outcomes reported by the included reviews that are relevant to the Umbrella Review question, and the type of review and the method of synthesis/analysis employed to synthesize the evidence as well as any comments or notes the Umbrella review authors may have regarding any included study.

Importantly, specific details of the factor or issue of interest to the Umbrella Review; for example the range of interventions, phenomena of interest, population details or outcome differences should be extracted in detail with the key findings/results. Extraction for a JBI Umbrella Review should be conducted independently by two reviewers to further minimize the risk of error. The protocol must therefore describe how data will be extracted and include the JBI data extraction instruments for systematic reviews in the appendices of the protocol. Extraction and presentation of data for a JBI Umbrella Review should be limited to the results and findings presented by the included research syntheses; in this regard it is not recommended that the researchers conducting the Umbrella Review retrieve primary studies (original research) in an included systematic review, for example, to access extra data. It is unlikely that authors of a JBI Umbrella review will need to contact the authors of an included research synthesis as is often the norm when undertaking a JBI Systematic Review (see other Chapters of this Manual).