Using it in your review
Using Cidexar in a published review
Cidexar is a screening aid. Your team makes every decision, and Cidexar gives you what you need to describe the step clearly in your methods and to answer the questions peer reviewers ask.
What to report
Describe the AI-assisted step as openly as any other part of your method. Cidexar records all of this for you.
The tool, and that it was AI-assisted.
Name Cidexar and say it was used as a first reader at the title and abstract stage.
The setting you used.
Cautious, standard or aggressive, and whether you changed it during the review.
Your criteria version.
The version of your eligibility criteria the screening used. Every version is saved and dated.
How set-aside papers were handled.
That they were kept, and how you checked them, for example by screening a random sample.
Who made the final decisions.
How many reviewers screened, and whether they worked independently.
The numbers.
Imported, duplicates, screened, set aside and included, ready for your flow diagram.
Methods text
A paragraph you can adapt
Cidexar gives you a ready-made paragraph for your review, filled in with your setting and numbers. Here is the shape of it, for a review with two independent reviewers:
"Titles and abstracts were first screened with Cidexar (cidexar.com), an AI-assisted screening tool, using the review's eligibility criteria at the standard setting. Records flagged as likely exclude were retained, and a random sample was checked by a reviewer. All inclusion decisions were made by two independent reviewers."
For a single-reviewer review, say so, and describe how the set-aside papers were checked. Adapt the wording to your review and to your journal's requirements.
Your flow diagram
Cidexar counts every record, so the boxes in your flow diagram add up.
Identified
Every record imported, from every database you searched, with its source file.
Duplicates
Records you marked as duplicates. Both originals stay on record.
Screened
Records screened at title and abstract, including those set aside by the AI and those without an abstract read by a person.
Excluded and included
Your team's decisions, kept separate from Cidexar's suggestions.
If a peer reviewer asks
Does the AI make the decisions?
No. Cidexar scores each title and abstract and suggests which papers clearly fail your criteria. People make every inclusion and exclusion decision, and Cidexar keeps its suggestions separate from your team’s decisions.
Could it have missed relevant studies?
It can, like any screening method. In our testing at the standard setting it kept about 98% of the papers that made it into the final reviews, and results vary by review. That’s why papers set aside are kept, the setting is your choice and you can screen a random sample of the set-aside pile. Report the setting and how you checked it.
Which AI model did you use?
Cidexar uses a combination of tools tested and tuned specifically for screening. The details aren’t public, but how it was tested and how well it performs are. See how it was tested
Is the screening step reproducible?
The parts you control are on record: the exact version of your criteria, the setting, every suggestion and every decision with its date. Report the criteria version and setting in your methods.
Responsible use of AI in evidence synthesis
Evidence-synthesis organisations have published guidance on using AI responsibly in reviews. Its common themes are being open about how AI was used, keeping people responsible for decisions, and protecting the rigour of the method. Cidexar is designed around the same ideas: it suggests and your team decides, nothing is hidden or deleted, and everything you need to report is on record.
Check your journal's and your field's current guidance as well. Requirements are changing quickly.
See it on your own review
Cidexar is coming soon. Leave your email and we'll let you know the moment it's ready.