Eligibility Criteria Rationalization Workbench
Uses RWD and NLP to identify criteria that can be safely broadened to improve enrollment Evidence basis: Trial Pathfinder found multiple restrictive oncology criteria had limited impact on treatment effect estimates while broader criteria increased eligible pools; later NLP and RWD studies support computable criteria simulation mainly in retrospective analyses
The Problem
“Rationalize restrictive trial eligibility criteria using RWD and NLP before protocol finalization”
Organizations face these key challenges:
Eligibility criteria are written in ambiguous free text and are hard to operationalize consistently
Sponsors lack quantitative evidence on which criteria truly affect safety or treatment effect estimates
Manual retrospective feasibility analysis is slow and expensive
Restrictive criteria reduce enrollment rates and site productivity
Underrepresentation of key patient subgroups weakens external validity
RWD sources are heterogeneous, incomplete, and noisy
Clinical and regulatory stakeholders need transparent rationale, not black-box recommendations
Impact When Solved
The Shift
Human Does
- •Review protocol eligibility criteria manually for restrictive language and inconsistencies
- •Coordinate criterion updates through spreadsheets, email, and review meetings
- •Assess enrollment impact using prior experience and retrospective evidence
- •Document rationale, approvals, and final wording changes for governance
Automation
- •No AI-driven analysis in the legacy workflow
- •No automated prioritization of criteria changes
- •No systematic simulation of broader eligibility options
- •No continuous monitoring of criteria-related enrollment risk
Human Does
- •Decide which proposed criterion changes are clinically and operationally acceptable
- •Approve final rationalization recommendations and supporting rationale
- •Review exceptions, edge cases, and criteria with uncertain evidence
AI Handles
- •Analyze eligibility criteria with RWD and NLP to flag overly restrictive requirements
- •Prioritize criteria changes by likely enrollment benefit and evidence strength
- •Generate standardized rationalization recommendations and supporting summaries
- •Monitor criteria patterns and surface items needing review or re-evaluation
Operating Intelligence
How Eligibility Criteria Rationalization Workbench runs once it is live
AI runs the first three steps autonomously.
Humans own every decision.
The system gets smarter each cycle.
Who is in control at each step
Each column marks the operating owner for that step. AI-led actions sit above the divider, human decisions and feedback loops sit below it.
Step 1
Assemble Context
Step 2
Analyze
Step 3
Recommend
Step 4
Human Decision
Step 5
Execute
Step 6
Feedback
AI lead
Autonomous execution
Human lead
Approval, override, feedback
AI handles assembly, analysis, and execution. The human gate sits at the decision point. Every cycle refines future recommendations.
The Loop
6 steps
Assemble Context
Combine the relevant records, signals, and constraints.
Analyze
Evaluate options, risk, and likely outcomes.
Recommend
Present a ranked recommendation with supporting rationale.
Human Decision
A human accepts, edits, or rejects the recommendation.
Authority gates · 1
The system must not change inclusion or exclusion criteria without approval from the clinical protocol team or study medical lead. [S1]
Why this step is human
The decision carries real-world consequences that require professional judgment and accountability.
Execute
Carry out the approved action in the operating workflow.
Feedback
Outcome data improves future recommendations.
1 operating angles mapped
Operational Depth
Technologies
Technologies commonly used in Eligibility Criteria Rationalization Workbench implementations:
Key Players
Companies actively working on Eligibility Criteria Rationalization Workbench solutions:
Real-World Use Cases
AI-assisted oncology clinical trial participant prescreening
An AI system reads patient chart information and quickly flags people who might qualify for oncology trials, instead of relying only on manual staff review.
AI first-reader triage for baseline low-dose CT lung cancer screening
An AI system looks at lung screening CT scans first and quickly clears scans that appear negative, so radiologists can focus on scans more likely to need follow-up.
AI simulation of trial population diversity impact on safety and effectiveness assessment
Use AI to test different trial designs on a computer first, so sponsors can see whether the study will include enough different kinds of patients to understand how the drug works for them.