Skip to main content

Supporting Material Matchmaking

Supporting Material Matchmaking is the process enabled by an API that takes ePI, IPS, and PV as input and returns composed Supporting Material that matches the terms, conditions, and languages present in the input data.

Purpose​

Selects relevant Supporting Material (aRMM/HEM) based on:

  • Patient conditions from IPS
  • Demographics and preferences from PV
  • Medication context from ePI
  • Language and literacy level

Matching Algorithm​

Step 1: Extract Context​

From inputs:

  • IPS: Conditions, allergies, medications
  • PV: Age, gender, pregnancy, language
  • ePI: Drug name, therapeutic class

Step 2: Query Candidates​

Search Supporting Material metadata:

  • Filter by standard terminology codes
  • Match language preference
  • Check content type (aRMM/HEM)
  • Verify target demographics

Step 3: Score Relevance​

Calculate scores based on:

  • Exact matches: High score (condition in IPS matches SM tag)
  • Related concepts: Medium score (hierarchical relationships)
  • Demographics: Age/gender appropriateness
  • Literacy level: Content complexity
  • Language: Preferred vs. available

Step 4: Rank & Filter​

  • Sort by relevance score
  • Apply threshold (minimum score)
  • Limit results (top N)
  • Deduplicate similar content

Step 5: Compose Response​

Return structured list:

  • SM resource references
  • Relevance scores
  • Matching reasons
  • Display recommendations

API Specification​

Request​

{
  "epi": { "resourceType": "Bundle", ... },
  "ips": { "resourceType": "Bundle", ... },
  "pv": { "resourceType": "Bundle", ... },
  "options": {
    "maxResults": 10,
    "minScore": 0.6,
    "contentTypes": ["aRMM", "HEM"],
    "languages": ["en", "es"]
  }
}

Response​

{
  "matches": [{
    "smReference": "DocumentReference/sm-123",
    "score": 0.95,
    "reasons": ["pregnancy", "language-match"],
    "displayPriority": "high"
  }]
}

Integration Points​

Lenses​

Lenses use matchmaking to:

  • Add hyperlinks to relevant SM
  • Embed supplementary content
  • Suggest educational materials

Applications​

Frontend apps use matchmaking to:

  • Display "Related Information" sections
  • Recommend educational content
  • Provide context-sensitive help

SM Tool​

SM Tool uses matchmaking to:

  • Preview relevance
  • Test tagging effectiveness
  • Identify content gaps

Performance Optimization​

  • Caching: Pre-compute common matches
  • Indexing: Fast terminology lookups
  • Batch processing: Multiple requests
  • CDN: Deliver SM content efficiently

Quality Metrics​

Track:

  • Match precision/recall
  • User engagement with matched SM
  • Content gap analysis
  • Language coverage