3. Abstract screening
This process requires going through the abstracts uploaded on Covidence and screening to include or exclude each abstract for full text screening. Be as inclusive as possible at this stage so that you don’t miss any records for full text screening (but not so inclusive that you end up screening over an unreasonable amount of papers).
This stage of the process can typically last from 1-2 months.
3.1. Setup and Steps Using Covidence
Because the screening process involves multiple people, here are the highly recommended settings to initiate before screening.
- Create include and exclude tags for each person screening. We’ve typically used Include_[initials of screener here] and Exclude_[initials]. This will help to differentiate between abstracts screened by different people.
- To do this, go to Settings > Study tags. Then, create a new tag.
- Make the settings to just 1 vote needed for each abstract record. This will help with sorting through abstracts already screened and those left.
- To do this, go to Settings > Review Settings, and then change “Reviewers required for screen” to 1.
Now, to begin screening the abstract records, go back to the project home page and click Continue under the section Title and abstract screening.
To tag a study, on the abstract screening page, just complete the following steps:
- Click the checkmark next to an abstract record.
- Click Tags and select the appropriate tag. You will see a tag at the bottom of an abstract record if you’ve tagged it. You can always untag a record by clicking the X on the tag.
- After tagging a study as Include_[Initials] or Exclude_[Initials], click No or Yes. No-voted abstracts will be shown on the Irrelevant references section while Yes-voted abstracts” will be shown in Full Text screening.
3.2. Recommended heuristics for efficient screening
You can screen more efficiently by doing the following:
- Keeping an eye out on papers with the population of interest (e.g., chronic pain) and MRI.
- Checking that they don’t meet any of the “quick exclusion” criteria below.
Here are recommendations for quick exclusions:
- Review papers. You can typically identify them by the journal they’re published in. However, if a review paper is also a meta-analysis on a similar topic, you can still exclude it but save the title of that paper for later (jot it down in a text file or somewhere to reference later).
- Examples of journals with reviews: “Current Opinion in XXXXX”, “Trends in XXXXX”, “Expert Reviews in XXXXX”, “Neuroscience and Biobehavioral Reviews”, “Nature review in XXXXX”
- Clinical trial registrations.
- Animal studies.
- Studies that do not use MRI at all (e.g., EEG, MEG, fNIRS, etc.). Make sure to look at the abstract to make sure there aren’t cases where they use both MRI and another modality.
- Sample size is less than 10.
- (if doing an adult study) sample is characterized by participants less than 18 years of age
When you are done, you can go on to Full text screening.