Skip to main content
It looks like you're using Internet Explorer 11 or older. This website works best with modern browsers such as the latest versions of Chrome, Firefox, Safari, and Edge. If you continue with this browser, you may see unexpected results.

Stony Brook University

Systematic Reviews: Data Extraction

A guide to conducting systematic reviews.

Data Extraction

Once you have identified all studies to be included in the systematic review, the next step is to extract and analyze the data contained in those studies. For a qualitative (non-meta-analysis) systematic review, you will create Summary of Findings tables and Bias/Evidence Quality figures. A meta-analysis requires pooling of data and specialized statistical analysis. This page will focus on qualitative systematic reviews.

The data extraction should be based on the previously defined interventions and outcomes established during the research question, inclusion/exclusion criteria, and search protocol development. If those stages have been done properly, it should not be too difficult to identify the data elements that need to be extracted from each included study.

Create a data extraction form that will be filled in for each included study. Use a software program that will allow you to create a form/questionnaire/survey and then create statistics, tables, and figures out of that data. There are a variety of these available including Microsoft Access/Excel, Qualtrics, REDCap, Google Forms/Sheets, etc.


The Data Extraction Form

Once a proposed data extraction form has been created in a selected software program:

1. Pre-test the data extraction form - have at least two separate individuals use the form to collect data from ~5 included studies.

2. Use that experience to fix any problems or solve any issues with the form. Check their interrator reliability to see how valid your form is.

3. Use the revised data extraction on all studies.

4. The data extraction form can include your evidence grading/bias evaluation or that can be done in a separate form.

NOTE: It is best practice to have two separate individuals do data extraction from each included study. That decreases bias. However, at present, this is only a strongly recommended practice and is not always practiced. However you decide to conduct this, make certain that are fully transparent in reporting the process when you write up your review methodology.