Expected Outcomes

Primary Expected Outcome

Participants receiving active investigational therapy are expected to demonstrate greater improvement from baseline in WOMAC Pain and WOMAC Physical Function scores at Week 12 compared with placebo.

Secondary Expected Outcomes

The active treatment group may demonstrate:

• Reduced joint stiffness
• Improved overall WOMAC scores
• Improved ability to perform daily activities
• Reduced index-joint pain
• Improved participant global assessment
• Reduced reliance on rescue analgesics
• Sustained benefit through Week 24

Digital Assessment Expected Outcomes

Use of the conversational WOMAC Assessment Agent, sentiment layer, and Intervention Agent is expected to demonstrate:

• Consistent protocol-based assessment administration
• Correct assignment of knee versus hip context
• Question-level voice sentiment analysis
• Identification of uncertainty and hesitation
• Identification of potential response dissonance
• Real-time Intervention Agent activation
• Clarification of participant misunderstandings
• Confirmation of ambiguous responses
• Reduction in missing assessment data
• Reduction in invalid responses
• Improved assessment completion
• Standardized structured data capture
• Complete interaction audit trail
• Real-time transfer of completed assessment data
• Reduced manual coordinator burden

Data Captured Per WOMAC Question

For each assessment question, the system displays a record containing: assessment (WOMAC Hip), question number, question text, participant response, response value, voice captured, sentiment analysis status, sentiment provider (Hume AI), primary sentiment/affective signals, confidence/uncertainty signal, dissonance detected, intervention threshold met, intervention agent activated, clarification provided, original response, final confirmed response, timestamp, and audit record availability.

Demo Scenario

For the live ICON demonstration, the trial dashboard is populated with: Trial MOTION-OA 301, Participant MOT-301-0147 (Demo Participant), Visit Week 12 (Day 84), Blinded treatment assignment, Index Joint Right Hip, Assessment WOMAC Hip, Conversational AI method with voice input and Hume AI sentiment analysis enabled, question-level sentiment and dissonance detection enabled, Intervention Agent enabled, status Due Today, previous assessment Week 8 completed, next scheduled assessment Week 24.

Demo Workflow

When the participant launches the assessment, the Studio retrieves the participant's trial configuration and instantiates the required agents and rules: Trial → Participant → Visit → Index Joint → Assigned WOMAC Assessment → Primary Assessment Agent → Voice Capture → Hume AI Sentiment Analysis → Question-Level Dissonance Detection → Intervention Agent When Triggered → Participant Confirmation → Scoring → Structured Study Data → Audit Trail.

The clinical agent is instantiated from the study configuration and operates within protocol rules, participant context, visit schedule, validated assessment, response rules, sentiment signals, intervention rules, escalation rules, and structured clinical data capture.

Recommended Live Demo Moment

At least one WOMAC question is configured to generate an illustrative Intervention Agent trigger: the participant selects a response indicating little or no difficulty, but then verbally adds information suggesting the activity was actually difficult. The Studio visually shows: response captured → Hume AI sentiment signal received → possible dissonance detected → intervention threshold met → Intervention Agent activated → clarification requested → participant confirms final answer → final response stored. This demonstrates the closed-loop clinical AI architecture rather than only the conversational interface.