¶ Extracting Data from a Study
With the form built, reviewers fill it in for each included study. Each
extraction is one reviewer's data for one citation, and it moves through a simple review workflow.
Who can do this: Leaders, Consolidators, and Contributors can create and edit extractions.
¶ Open an Extraction
From the project's Extractions ▸ dropdown, choose Data Extractions
(/projects/<id>/extractions) to see the list, then open one to Work on it. (Citations become available for extraction once they've cleared screening — see
Citation lifecycle.)
The work screen: assignment and status at the top, the form's section tabs below.
The header shows Assigned to (the extractor), the citation and its Refman ID, and the
extraction status controls.
¶ Fill in the Form
- Work through the section tabs. Type 2 questions are answered inline; for Type 1 sections you
add an item per arm/outcome and fill its grid. - Use Select Key Question(s) to tag which key questions this extraction addresses.
- Your entries save automatically as you type or choose (there's no separate save button for
fields).
¶ The Status Workflow
The Extraction Status buttons track progress:
| Status | Color | Meaning |
|---|---|---|
| awaiting_work | orange | being worked on / not yet submitted |
| awaiting_review | blue | submitted, waiting for a reviewer |
| work_returned | red | a reviewer sent it back for changes |
| work_accepted | green | a reviewer approved it |
- When you've completed every section, VADRR asks "Would you like to submit this extraction for
review?" — accepting moves it to awaiting_review. - A reviewer (someone able to review) can then set work_accepted to approve, or work_returned
to send it back. On returning, VADRR prompts "Please enter a rejection reason if any. (Can be left
blank)." - A returned extraction shows "Rejection Reason: …" in red so the extractor knows what to fix.
¶ Deleting Type 1 Items
Removing a Type 1 item that already has data (or is used in comparisons) requires typing CONFIRM,
since it causes data loss.
¶ Next Steps
- When two or more reviewers have extracted the same study, reconcile them with the
Comparison Tool and consolidation. - Speed things up with LLM-assisted extraction.