SRDR+ 4. Screening: When to Use the SRDR+ Screener?

For many systematic reviews, a thorough search will result in hundreds or thousands of potential articles. The Institutes of Medicine guidelines indicate that title/abstract screening should be carried out by two or more members of the review team working independently (see Standards for Finding and Assessing Individual Studies Standards for Finding and Assessing Individual Studies – Finding What Works in Health Care – NCBI Bookshelf).

Double-screening thousands of titles and abstracts can be difficult and time-consuming.

However, SRDR incorporates a screening tool that uses advanced machine learning (ML) models to help your team more efficiently identify studies appropriate for your project.

How does it work?

The ML screener uses the inclusion/exclusion decisions of the team members to create a prediction (calculate a probability) score for every article. It then rearranges the articles to pull the ones most likely to be included to the front of the screening queue.

Because the ML model relies on team member screenings to formulate the model, n=100 initial screens are needed to generate the model. So, you won’t see results right away. However, any “pearls” (articles that you have already pre-identified for inclusion) that are manually marked for inclusion in the Project Dashboard (include link to this), are factored into the ML model by default. So, when beginning a project, if you already have a set of “pearl” articles, you will want to manually mark those for inclusion in the Project Dashboard.

In this tutorial, we will walk you through the different screening setup options, how to screen, and how to use the ML model results to make a decision regarding when to stop screening.

Yes, that’s right, you may not need to screen thousands of articles! Once you meet your stopping criteria (below), your team can stop screening (letting the ML model “screen” the rest for you).

Organization of This Section

This part of the tutorial is set up in two sections:

4.1 Resources for Team Leaders: This section provides information for how to set up the screening team, how to set up the screener so that it meets your team’s needs (there are many options for this), and how to track progress.

4.2 Resources for Screeners: This section provides basic instructions for members of the screening team–how to effectively use the tool.