Background and Motivation

Healthcare data exists in fragmented silos across thousands of incompatible systems. A typical patient's health data is distributed across:

• 3-5 different Electronic Health Record (EHR) systems
• 2-3 wearable devices with proprietary formats
• Multiple mobile health applications
• Research study databases
• Insurance claim systems

Current standards like HL7 FHIR address structured clinical data but lack:

1. Cryptographic verifiability of data provenance
2. Built-in privacy-preserving computation
3. Support for continuous sensor data streams
4. Regulatory compliance by design

Design Principles

High-Impact Use Cases

Core Data Structures

{
  "@context": [
    "https://www.w3.org/2018/credentials/v1",
    "https://w3id.org/uhds/contexts/health-credential-v1.jsonld"
  ],
  "id": "urn:uuid:98c5cffc-7b48-4d5a-9a8a-123456789abc",
  "type": ["VerifiableCredential", "HealthCredential", "LabResult"],
  "issuer": "did:web:lab.hospital.example",
  "issuanceDate": "2026-01-15T08:30:00Z",
  "credentialSubject": {
    "id": "did:example:patient123",
    "testType": "https://loinc.org/4548-4",
    "testName": "Hemoglobin A1c",
    "result": 5.8,
    "unit": "%",
    "referenceRange": {
      "low": 4.0,
      "high": 5.6,
      "population": "non-diabetic adults"
    }
  },
  "credentialSchema": {
    "id": "https://w3id.org/uhds/schemas/lab-result.json",
    "type": "JsonSchemaValidator2023"
  },
  "privacyLevel": "pseudonymous",
  "dataQuality": {
    "grade": "clinical",
    "confidence": 0.97,
    "deviceCertification": "FDA-510(k)-cleared"
  },
  "proof": {
    "type": "Ed25519Signature2020",
    "created": "2026-01-15T08:30:00Z",
    "verificationMethod": "did:web:lab.hospital.example#key-1",
    "proofPurpose": "assertionMethod",
    "proofValue": "z58DAdFfa9SkqZMVPxAQpic7ndSayn1PzZs6ZjWp1CktyGesjuTSwRdoWhAfGFCF5bppETSTojQCrfFPP2oumHKtz"
  }
}

Standard Credential Types

{
  "$schema": "http://json-schema.org/draft/2020-12/schema",
  "$id": "https://w3id.org/uhds/schemas/lab-result.json",
  "title": "UHDS Laboratory Result",
  "description": "Verifiable credential for clinical laboratory results",
  "type": "object",
  "required": [
    "testType", "testName", "result", "unit", 
    "collectionDateTime", "issuer", "subject"
  ],
  "properties": {
    "testType": {
      "type": "string",
      "format": "uri",
      "description": "LOINC code for test type",
      "examples": ["https://loinc.org/4548-4"]
    },
    "testName": {
      "type": "string",
      "description": "Human-readable test name",
      "examples": ["Hemoglobin A1c", "Complete Blood Count"]
    },
    "result": {
      "oneOf": [
        { "type": "number" },
        { "type": "string" },
        { 
          "type": "object",
          "properties": {
            "value": { "type": "number" },
            "comparator": { 
              "type": "string",
              "enum": ["<", "<=", ">", ">=", "="]
            }
          }
        }
      ]
    },
    "unit": {
      "type": "string",
      "pattern": "^[a-zA-Z0-9/\\-\\^]*$",
      "description": "UCUM unit of measure"
    },
    "referenceRange": {
      "$ref": "#/definitions/ReferenceRange"
    },
    "criticalValue": {
      "type": "boolean",
      "description": "Indicates life-threatening result"
    }
  }
}

JSON-LD Context Definition

The UHDS JSON-LD context provides semantic interoperability across healthcare systems:

{
  "@context": [
    "https://www.w3.org/ns/credentials/v2",
    {
      "@version": 1.1,
      "uhds": "https://w3id.org/uhds/vocab#",
      "health": "https://schema.org/MedicalEntity#",
      "loinc": "https://loinc.org/",
      "snomed": "http://snomed.info/id/",
      
      // Core UHDS Types
      "HealthCredential": "uhds:HealthCredential",
      "LabResult": "uhds:LabResult",
      "VitalSigns": "uhds:VitalSigns",
      "GenomicData": "uhds:GenomicData",
      
      // Properties
      "testType": {
        "@id": "health:testType",
        "@type": "@id"
      },
      "result": {
        "@id": "health:resultValue",
        "@type": "xsd:decimal"
      },
      "unit": {
        "@id": "health:measurementUnit",
        "@type": "xsd:string"
      },
      "referenceRange": {
        "@id": "health:normalRange",
        "@type": "uhds:Range"
      },
      "privacyLevel": {
        "@id": "uhds:privacyLevel",
        "@type": "uhds:PrivacyLevel"
      },
      "dataQuality": {
        "@id": "uhds:dataQuality",
        "@type": "uhds:DataQuality"
      },
      
      // Differential Privacy
      "differentialPrivacy": "uhds:differentialPrivacy",
      "epsilon": {
        "@id": "uhds:epsilon",
        "@type": "xsd:decimal"
      },
      "delta": {
        "@id": "uhds:delta", 
        "@type": "xsd:decimal"
      },
      
      // Zero-Knowledge Proofs
      "zkProof": "uhds:zeroKnowledgeProof",
      "circuitId": {
        "@id": "uhds:circuitIdentifier",
        "@type": "@id"
      }
    }
  ]
}

Signature Suite Requirements

UHDS implementations MUST support the following cryptographic suites to ensure interoperability:

Zero-Knowledge Proof Circuits

UHDS defines standard ZKP circuits for common healthcare scenarios:

# R1CS Constraints for Range Proof
# Proves: min ≤ value ≤ max
# Without revealing: value

circuit RangeProof {
    // Public inputs
    input min;
    input max;
    
    // Private input
    private input value;
    
    // Constraints
    value - min >= 0;
    max - value >= 0;
    
    // Output proof
    output proof;
}

circuit BMIProof {
    // Private inputs (cm, kg)
    private input height_cm;
    private input weight_kg;
    
    // Public output ranges
    input bmi_min;
    input bmi_max;
    
    // Compute BMI
    // BMI = weight_kg / (height_m)^2
    // height_m = height_cm / 100
    var height_m = height_cm / 100;
    var bmi = weight_kg / (height_m * height_m);
    
    // Prove BMI in range
    bmi - bmi_min >= 0;
    bmi_max - bmi >= 0;
    
    output proof;
}

Differential Privacy Framework

For aggregate data analysis and AI training, UHDS specifies differential privacy mechanisms:

Credential Issuance API

POST /api/v1/credentials/issue HTTP/1.1
Host: api.healthsystem.example
Content-Type: application/json
Authorization: Bearer <issuer_token>

{
  "type": "LabResult",
  "credentialSubject": {
    "id": "did:example:patient123",
    "testType": "https://loinc.org/4548-4",
    "result": 5.8,
    "unit": "%",
    "collectionDateTime": "2026-01-15T07:00:00Z"
  },
  "options": {
    "privacyLevel": "pseudonymous",
    "proofFormat": "Ed25519Signature2020",
    "credentialDuration": "P90D"
  }
}

HTTP/1.1 201 Created
Content-Type: application/json
Location: /api/v1/credentials/urn:uuid:98c5cffc-7b48-4d5a-9a8a-123456789abc

{
  "credential": {
    "@context": ["..."],
    "id": "urn:uuid:98c5cffc-7b48-4d5a-9a8a-123456789abc",
    "type": ["VerifiableCredential", "LabResult"],
    "issuer": "did:web:api.healthsystem.example",
    "issuanceDate": "2026-01-15T08:30:00Z",
    "credentialSubject": {...},
    "proof": {...}
  },
  "metadata": {
    "issued": "2026-01-15T08:30:00Z",
    "expires": "2026-04-15T08:30:00Z",
    "privacyLevel": "pseudonymous",
    "revocationIndex": "12345"
  }
}

Credential Verification API

UHDS supports multiple verification modes depending on privacy requirements:

Zero-Knowledge Proof API

POST /api/v1/zkp/generate HTTP/1.1
Host: wallet.patient.example
Content-Type: application/json

{
  "credentialId": "urn:uuid:98c5cffc-7b48-4d5a-9a8a-123456789abc",
  "circuitType": "range_proof",
  "circuitParameters": {
    "field": "result",
    "min": 6.5,
    "max": 8.0
  },
  "verifier": "did:web:research.trial.example",
  "purpose": "clinical_trial_screening",
  "validityDuration": "P7D"
}

HTTP/1.1 200 OK
Content-Type: application/json

{
  "proof": {
    "type": "R1CSRangeProof2023",
    "circuitId": "https://w3id.org/uhds/circuits/range_proof_v1",
    "publicInputs": {
      "min": 6.5,
      "max": 8.0,
      "timestamp": "2026-01-15T09:00:00Z"
    },
    "proofValue": "z3t4F...abC9X",
    "verificationKey": "https://w3id.org/uhds/keys/range_proof_vk"
  },
  "metadata": {
    "generated": "2026-01-15T09:00:00Z",
    "expires": "2026-01-22T09:00:00Z",
    "privacyLevel": "zero_knowledge"
  }
}

Differential Privacy Query API

POST /api/v1/query/dp HTTP/1.1
Host: api.researchportal.example
Content-Type: application/json
Authorization: Bearer <researcher_token>

{
  "query": {
    "type": "average",
    "field": "result",
    "filter": {
      "testType": "https://loinc.org/4548-4",
      "unit": "%"
    }
  },
  "privacyParameters": {
    "epsilon": 1.0,
    "delta": 1e-5,
    "mechanism": "laplace"
  },
  "accessToken": "<consent_token>"
}

GDPR Compliance

UHDS is designed with GDPR principles as foundational requirements:

HIPAA Compliance

For US healthcare applications, UHDS addresses HIPAA requirements:

FDA Regulatory Considerations

For applications that may be considered Software as a Medical Device (SaMD), UHDS provides:

• Clinical Validation Framework: Documentation requirements for algorithm validation studies
• Change Control: Versioning and update mechanisms for regulated software
• Adverse Event Reporting: Integration with FDA reporting systems for safety monitoring
• Precertification Support: Alignment with FDA Digital Health Software Precertification Program

International Regulations

UHDS supports compliance with global healthcare regulations:

Conformance Classes

UHDS defines multiple conformance classes to support different deployment scenarios:

Test Suite Requirements

All UHDS implementations MUST pass the official test suite, available at https://github.com/UHDS-Standards/test-suite.

Reference Implementations

The following reference implementations are available:

Threat Model

UHDS addresses the following threat actors and vectors:

Cryptographic Security

UHDS employs defense-in-depth cryptographic protections:

Privacy Security Considerations

Beyond traditional security, UHDS addresses privacy-specific threats:

• Linkage Attacks: Preventing correlation of multiple credentials from the same individual through pseudonym rotation and unlinkable proofs
• Background Knowledge Attacks: Mitigating inference attacks using public information through differential privacy guarantees
• Statistical Disclosure: Preventing reconstruction of individual data from aggregate statistics using formal privacy proofs
• Metadata Analysis: Protecting timing, frequency, and other metadata that could reveal sensitive information

Privacy by Design Principles

UHDS implements the 7 Foundational Principles of Privacy by Design:

1. Proactive not Reactive; Preventative not Remedial: Privacy protections are built into the architecture, not added as an afterthought
2. Privacy as the Default Setting: Systems default to maximum privacy; users must opt-in to reduced privacy
3. Privacy Embedded into Design: Privacy is an integral component of core functionality
4. Full Functionality – Positive-Sum, not Zero-Sum: Privacy and functionality are mutually reinforcing, not trade-offs
5. End-to-End Security – Full Lifecycle Protection: Privacy extends throughout the entire data lifecycle
6. Visibility and Transparency – Keep it Open: All privacy measures are transparent and verifiable
7. Respect for User Privacy – Keep it User-Centric: Privacy controls are understandable and accessible to users

Data Minimization Techniques

UHDS employs multiple layers of data minimization:

Consent Management

UHDS implements granular, auditable consent mechanisms:

{
  "consentId": "urn:uuid:consent-12345",
  "subject": "did:example:patient123",
  "purpose": "clinical_research_study_abc",
  "dataTypes": ["LabResult", "VitalSigns"],
  "fields": ["result", "unit", "testType"],
  "processing": {
    "analysis": true,
    "aggregation": true,
    "sharing": false
  },
  "constraints": {
    "validUntil": "2026-12-31T23:59:59Z",
    "maxQueries": 100,
    "privacyBudget": {
      "epsilon": 2.0,
      "delta": 1e-5
    }
  },
  "revocable": true,
  "auditRequired": true
}

Privacy Impact Assessment Framework

UHDS includes a structured Privacy Impact Assessment (PIA) framework for implementations:

Language Support

UHDS supports multilingual healthcare contexts through:

{
  "testName": {
    "@value": "Hemoglobin A1c",
    "@language": "en"
  },
  "interpretation": {
    "@value": "正常範圍內",
    "@language": "zh"
  },
  "instructions": {
    "@value": "禁食8小時後檢測",
    "@language": "zh",
    "@container": "@set"
  }
}

Cultural Considerations

UHDS addresses cultural variations in healthcare:

• Reference Ranges: Population-specific normal ranges based on ethnicity, geography, and demographics
• Units of Measure: Automatic conversion between SI units and local conventional units
• Date/Time Formats: Support for different calendar systems and time representations
• Privacy Expectations: Cultural variations in privacy preferences and data sharing norms

AI and Algorithmic Ethics

UHDS supports ethical AI development through:

• Bias Detection and Mitigation: Tools for identifying and correcting algorithmic bias in health AI
• Explainability: Requirements for interpretable AI models in clinical decision support
• Human Oversight: Mechanisms for human review of AI recommendations in clinical settings
• Accountability: Audit trails for AI training data and decision processes

Research Ethics

For research applications, UHDS implements: