Evidence Synthesis Methodologies in the Health Sciences

Study selection, or screening, is the process of eliminating references that do not meet your inclusion criteria. Best practice is to have two independent reviewers for each study, as well as a third reviewer to resolve conflicts. Elimination or retention of references at each stage should be based on your research question and inclusion and exclusion criteria as defined in your protocol.

Because Brown University subscribes to Covidence, this guide focuses on how to use that tool for screening and later stages of evidence synthesis projects.

To use Covidence as your screening tool, you'll need an account.

Once you have an account, you'll create a new review. This guide to using Covidence from the UNC Health Sciences Library will take you through the entire process; consult your librarians here at Brown if you have questions.

These FAQs from Covidence will also be helpful.

Covidence performs an automatic deduplication process as you import references. You can review these decisions for errors as needed, as well as marking duplicates manually. Deduplication should be completed before the screening process beings.

Pilot Screening (Optional)

Pilot or test screening a random sample of your references will help with consistency in screening decisions by all reviewers. Differences of opinion can be resolved before you begin screening large sets of results, and you can then add details about those consensus decisions to the Eligibility Criteria section of your Covidence project so they can be displayed during screening.

Title/Abstract Screening

This it the first formal phase of the screening process. During this phase, reviewers use the title and abstract of each reference to decide if it:

• Does not meet the inclusion criteria, and should be excluded (NO).
• Might meet the inclusion criteria, but lacks enough information to make a final decision (MAYBE).
• Does meet the inclusion criteria (YES).

Maybe and Yes studies move forward to the second and final phase, which is full text screening.

Full Text Screening

During the full text screening phase, the full text of all studies to be screened must be obtained and uploaded to Covidence. At this point, the decision is simply between whether a reference meets the inclusion criteria (INCLUDE) or does not (EXCLUDE). You'll need to document your reasons for exclusion during this phase.

Covidence will automatically generate a PRISMA flow diagram based on the references you import and screen. This may or may not be complete enough for final reporting, but it's a good start. For more details about what the flow diagram should contain and how it should be populated, see the Report Results page of this guide.