Baseline Assessment

Overview

The baseline assessment establishes a user's personal wellness baseline through a structured conversation that includes validated clinical screening questions. This is typically completed once when a user first joins the platform.

Clinical Questions

The baseline assessment includes five standardised questions:

PHQ-2 (Depression Screening)

1. "Over the last two weeks, how often have you been bothered by having little interest or pleasure in doing things?"

   • Not at all (0)
   • Several days (1)
   • More than half the days (2)
   • Nearly every day (3)

2. "Over the last two weeks, how often have you been bothered by feeling down, depressed, or hopeless?"

   • Not at all (0)
   • Several days (1)
   • More than half the days (2)
   • Nearly every day (3)

GAD-2 (Anxiety Screening)

3. "Over the last two weeks, how often have you been bothered by feeling nervous, anxious, or on edge?"

   • Not at all (0)
   • Several days (1)
   • More than half the days (2)
   • Nearly every day (3)

4. "Over the last two weeks, how often have you been bothered by not being able to stop or control worrying?"

   • Not at all (0)
   • Several days (1)
   • More than half the days (2)
   • Nearly every day (3)

Mood Scale

5. "On a scale of 1 to 10, where would you say your mood is today?"
   • 1 (Very low) to 10 (Excellent)

Score Calculation

Clinical Composite Score

The clinical component uses a composite formula:

// PHQ-2 Component (0-6 scale → 0-21 contribution)
const phq2Component = (6 - phq2Total) * 3.5;
 
// GAD-2 Component (0-6 scale → 0-21 contribution)
const gad2Component = (6 - gad2Total) * 3.5;
 
// Mood Component (1-10 scale → 0-50 contribution)
const moodComponent = moodScale * 5;
 
// Clinical Score (0-100)
const clinicalScore = phq2Component + gad2Component + moodComponent;

Multimodal Enrichment

For baseline assessments, the clinical score is enriched with multimodal features:

Final Score = (Clinical Score × 0.70) + (Audio Score × 0.15) + (Visual Score × 0.15)

If visual features fail to extract:

Final Score = (Clinical Score × 0.85) + (Audio Score × 0.15)

Implementation

Component: BaselineAssessmentSDK.tsx

The baseline assessment uses the ElevenLabs SDK:

import { useConversation } from '@elevenlabs/react';
 
const BaselineAssessmentSDK = ({ userId, onComplete }) => {
  const conversation = useConversation({
    onMessage: handleMessage,
    onDisconnect: handleComplete
  });
  
  const startAssessment = async () => {
    await conversation.startSession({
      agentId: 'agent_9301k22s8e94f7qs5e704ez02npe'
    });
  };
  
  // ... rest of implementation
};

Response Extraction

Clinical responses are extracted from the conversation transcript using pattern matching:

const extractMoodScore = (transcript: string): number => {
  // Handles numeric ("8") and word numbers ("eight")
  const patterns = [
    /(?:mood|score|rate).*?(\d+|one|two|three|four|five|six|seven|eight|nine|ten)/i,
    /(?:I(?:'d| would) (?:say|put|rate).*?)(\d+|one|two|three|...)/i
  ];
  // ... extraction logic
};

Data Stored

The baseline assessment stores:

{
  "assessment_type": "baseline",
  "clinical_scores": {
    "phq2_total": 1,
    "gad2_total": 1,
    "mood_scale": 7,
    "phq2_positive_screen": false,
    "gad2_positive_screen": false
  },
  "conversation_quality": "complete",
  "elevenlabs_session_id": "conv_xxxx",
  "mind_measure_composite": {
    "score": 77,
    "phq2_component": 21,
    "gad2_component": 21,
    "mood_component": 35
  },
  "multimodal_enrichment": {
    "enabled": true,
    "audio_features": { ... },
    "visual_features": { ... },
    "scoring_breakdown": {
      "clinicalScore": 77,
      "audioScore": 25,
      "visualScore": 40,
      "finalScore": 64
    }
  }
}

Clinical Interpretation

PHQ-2 Screening

• Score 0-2: Negative screen (low risk)
• Score 3-6: Positive screen (warrants further assessment)

GAD-2 Screening

• Score 0-2: Negative screen (low risk)
• Score 3-6: Positive screen (warrants further assessment)

Important: These are screening tools, not diagnostic instruments. A positive screen should prompt referral to appropriate clinical resources.

Last Updated: December 2025