Skip to Main Content

Data extraction

Latest guidance from JBI

Video (JBI, 2023) 1h 11min - 'Prof Zachary Munn and Dr Danielle Pollock discuss recommendations for the extraction, analysis and presentation of results in scoping reviews'

What is data extraction?

For scoping reviews, traditionally the term 'data charting' has been used to collectively describe the processes of extraction, analysis, and presentation.

Data items relevant to the review questions are extracted, analysis undertaken (usually via frequency counts and/or basic qualitative content analysis), and data presented (typically using tables, visualisations, and narrative summary) to adress the scoping review's objectives, and ideally answer the questions of the review.

'For ease of reference and tracking, it is suggested that reviewers keep careful records to identify each source. As reviewers chart each source, it may become apparent that additional unforeseen data can be usefully charted. Charting the results can therefore be an iterative process whereby the charting table is continually updated.' - JBI Manual for Evidence Synthesis, Chapter 11: Scoping reviews, 11.2.7 Data extraction

  • The information that you extract from the evidence sources included in your scoping review will depend on the purpose of your review and your question.
  • Data extraction forms are invaluable. These should be designed to target the objectives of the review, and usually include the population, concept and context. Develop an accompanying extraction guidance form.
  • Data extraction forms should ideally be piloted by the review team for each evidence source type to ensure that all required data is captured.
  • Assessment of eligibility of studies, and extraction of data from sources, should be done by at least two people, independently.
  • A draft data extraction form should be developed with your protocol, but it is acceptable to refine this at the review stage.
  • Data may be reported in diverse formats.

What data should be collected for each included study?

'...scoping review authors should only extract data items that are relevant to the scoping review questions...potential data items of interest can be structured around the PCC framework.' - (Pollock et al., 2023)

According to the JBI Manual for Evidence Synthesis, Chapter 11: Scoping reviews, 11.2.7 Data extraction, key information to chart for each paper can include:

  • Author(s)
  • Year of publication
  • Origin/country of origin (where the study was published or conducted)
  • Aims/purpose
  • Population and sample size (if applicable)
  • Methodology/methods
  • Intervention type/duration, comparater
  • Outcome measures (if applicable)
  • Key findings that relate to the scoping review question/s

Learn more:

The PRISMA-ScR includes some further guidance:

Data extraction templates/examples

Data extraction using Covidence

Covidence

If you are intending to use Covidence to complete your data extraction, the software will take you through two linked processes:

  • quality assessment
  • data extraction/table building
Note: You do not have to use a critical appraisal tool to assess your references for a scoping review, unless you have stated that you will do so as part of your protocol.

  • If not performing quality assessment, skip this step and proceed to data extraction.
  • If you are undertaking a critical appraisal, the PRISMA-ScR offers some guidance.
Note: after completing your data extraction in Covidence, you must export your data

How to present the results

Your protocol should include a plan for how you will present your results.

Your PCC inclusion criteria will assist you in choosing how the data should be mapped most appropriately, but you can refine this toward the end of the review, when you have a better picture of the sort of data available in your included studies.

The results of a scoping review may be presented in your final paper in a variety of ways, including:

  • tables and charts, featuring distribution of studies by year or period of publication, countries of origin, area of intervention (clinical, policy, educational, etc.) and research methods; and/or
  • in a descriptive format that aligns with the review objective/s and scope.

The latest guidance (Pollock et al. 2023) encourages 'creative approaches...to convey results to the reader in an understandable way' such as word clouds, honeycombs, heat maps, tree graphs, iconography, waffle charts and interactive resources.

Note: If you present your data in a table/chart, also include a narrative summary to explain how the results relate to your review objectives and questions. 

 

JBI advise (11.3.8.1 Search results) results can be classified under main conceptual categories, such as:

  • intervention type
  • population (and sample size, if it is the case)
  • duration of intervention
  • aims
  • methodology adopted
  • intervention type
  • key findings (evidence established)
  • gaps in the research

'For each category reported, a clear explanation should be provided.'

Navigate to report broken links and provide feedback page