Building a Cohort

Use Oracle AI Agents for Research to build a cohort to develop research and data plans.

To build a cohort using AI Agents for Research:

1. Ask AI Agents for Research to build a cohort, specifying the inclusion and exclusion criteria and any other relevant requirements. The AI Agent answers with a cohort definition and patient counts in the cohort.
2. Review the proposed cohort logic. Respond to the AI Agent with clarifying perimeters as needed.
   1. As part of building the cohort, AI Agents for Research maps clinical concepts in your criteria to standardized codes, code lists, or ontologies. You may need to review concept mappings before saving the cohort. See the Review Concept Mappings section below for more information,
3. Repeat Steps 1 and 2 until the cohort meets your research requirements.
4. To save the cohort, ask AI Agents for Research to save it to the specified location.

See the sample cohort prompts below for examples of perimeters to ask the AI Agent:

Create a cohort of adults with type 2 diabetes without complications. 
Include patients with at least one diagnosis in the last 24 months. 
Exclude patients with type 1 diabetes.

Create a cohort in Oracle RWD of adults age 18 or older with type 2 diabetes without complications. 
Include patients with at least one diagnosis in the last 24 months. 
Exclude patients with type 1 diabetes. 
Show which codes you are using and explicitly state your assumptions.

Cohort logic may be mapped to standardized codes. Review concept mappings for the following types:

• Diagnosis codes
• Procedure codes
• Medication concepts
• Laboratory or observation concepts
• Included and excluded terms
• Unmapped or ambiguous concepts

If a mapping if incomplete or inaccurate, ask AI Agents for Research to revise the mapping and review the updated mapping before proceeding.

Remove gestational diabetes codes from the concept list and show me the revised mapping.