Skip to main content

Registry project to case finding to curation to export

This scenario demonstrates how to create a registry project, attach the relevant cohorts, review potential registry cases, start curation, track abstraction progress, and export registry results.

Best for: registrar workflows, cancer registry demos, and human-in-the-loop abstraction.

Step 1: Create a registry project​

  1. In the left navigation, go to Patient Registry > Registry Projects.
  2. Click New Registry Project to start the project creation workflow.
  3. In Project Details, select the registry mode: Standard Registry or Custom Registry.
  4. Enter the registry project title and description.
  5. Configure the reporting context by selecting the registry standard, jurisdiction, compliance program, and registry.
  6. Configure the validation metafile and validation rule sets if specific validation settings are required.
  7. Continue to Cohorts and select one or more cohorts to include in the registry project.
  8. Continue to Collaborators and assign the team members who should participate in abstraction, validation, or project review.
  9. Continue to Review and confirm the registry project details, selected cohorts, assigned team, registry configuration, and validation settings.
  10. Click Save Registry Project to create the registry project.

Step 2: Review potential registry cases​

  1. Open the created registry project from Patient Registry > Registry Projects.
  2. Review the project summary and case status cards, including total cases, pending abstraction, in-progress cases, complete cases, and submitted cases.
  3. Review the Potential Cases list.
  4. Use search or filters to find specific patients or case statuses when needed.
  5. Use Add Cases if additional cases should be added to the registry project.
  6. Review case details such as patient profile, clinical profile, document count, assignee, tumor ID, and last activity.
  7. Mark cases as not reportable where appropriate, or keep them in the potential case list for abstraction.

Step 3: Start curation for selected cases​

  1. In the case table, identify cases that require abstraction.
  2. Click Start Curation for the selected case.
  3. Confirm the patient, clinical profile, available documents, tumor ID, and assignee.
  4. Start the curation workflow.
  5. Monitor the case as it moves into abstraction or in-progress status.
  6. Repeat the process for additional cases that require curation.

Step 4: Review extracted fields and evidence​

  1. Open a case that is in abstraction or review.
  2. Review the extracted registry fields, confidence indicators, and linked evidence.
  3. Open supporting source documents when field-level validation is needed.
  4. Compare extracted values with the supporting evidence.
  5. Correct extracted values where needed.
  6. Mark fields or the case according to the available review workflow.
  7. Track project progress from the registry project dashboard.

Step 5: Complete cases and export registry results​

  1. Confirm that reviewed cases have the required fields completed.
  2. Confirm that key registry values have supporting evidence.
  3. Resolve outstanding review items before marking the case complete.
  4. Mark cases as complete when they are ready.
  5. Return to the registry project detail page and review the final project status counts.
  6. Click Exports to prepare the registry results.
  7. Select the required export option or format.
  8. Share the exported results with the registry, reporting, or data operations team according to the approved workflow.

Recipe reference​

This scenario is also a standalone recipe you can reuse in other workflows.

Curate a registry project with human review​

When to use on its own: For cancer registry abstraction, registry-specific data collection, or demos of human-in-the-loop curation.
Features involved: Patient Registry Projects, case finding, Standard or Custom Registries, Data Curation, case review, validation, and export.
Edge cases / limitations: Registry project setup options vary by deployment. External submission behavior should be confirmed before customer demos. Some fields may be calculated or read-only.
Value: Combines AI-assisted abstraction with registrar control and audit-friendly review.