Chapter 5: Data Management

5.1 IoMT Data Lifecycle

Medical IoT devices generate continuous streams of health data that must flow through collection, transmission, storage, processing, and ultimately integration with clinical systems. Managing this data lifecycle effectively determines whether IoMT delivers clinical value.

The data journey presents challenges at each stage: ensuring measurement accuracy at collection, maintaining integrity during transmission, enabling efficient storage at scale, supporting real-time processing for alerts, and achieving semantic interoperability for clinical use.

5.2 Data Interoperability Standards

Interoperability enables IoMT data to flow between devices, platforms, and clinical systems with shared meaning.

Key Standards for IoMT

Standard	Scope	Application
HL7 FHIR	Healthcare data exchange	EHR integration, clinical data
IEEE 11073	Personal health devices	Device data formats, PHD profiles
Continua (IHE PCD)	Connected health devices	End-to-end interoperability
LOINC	Observation codes	Standard measurement identifiers
SNOMED CT	Clinical terminology	Clinical concepts, conditions
HL7 v2	Legacy messaging	ADT, ORU messages to legacy systems

5.3 IEEE 11073 Personal Health Devices

IEEE 11073 defines data models and protocols for personal health devices:

IEEE 11073 PHD Specializations

Profile	Device Type	Measurements
11073-10407	Blood pressure monitor	Systolic, diastolic, MAP, pulse
11073-10408	Thermometer	Body temperature
11073-10415	Weighing scale	Body weight, BMI
11073-10417	Glucose meter	Blood glucose
11073-10441	Cardiovascular fitness	Heart rate, activity
11073-10404	Pulse oximeter	SpO2, pulse rate

5.4 FHIR Resources for IoMT

FHIR provides standardized resources for representing IoMT data:

Observation Resource Example

{
  "resourceType": "Observation",
  "id": "bp-reading-001",
  "status": "final",
  "category": [{
    "coding": [{
      "system": "http://terminology.hl7.org/CodeSystem/observation-category",
      "code": "vital-signs"
    }]
  }],
  "code": {
    "coding": [{
      "system": "http://loinc.org",
      "code": "85354-9",
      "display": "Blood pressure panel"
    }]
  },
  "subject": { "reference": "Patient/123" },
  "effectiveDateTime": "2025-01-15T08:30:00Z",
  "device": { "reference": "Device/bp-monitor-456" },
  "component": [
    {
      "code": { "coding": [{ "system": "http://loinc.org", "code": "8480-6" }] },
      "valueQuantity": { "value": 120, "unit": "mmHg" }
    },
    {
      "code": { "coding": [{ "system": "http://loinc.org", "code": "8462-4" }] },
      "valueQuantity": { "value": 80, "unit": "mmHg" }
    }
  ]
}

5.5 Data Quality Management

Ensuring IoMT data quality is essential for clinical decision-making:

Data Quality Dimensions

• Accuracy: Measurements reflect true physiological values
• Completeness: No missing data points or metadata
• Timeliness: Data available when clinically needed
• Consistency: Measurements comparable across devices/time
• Validity: Values within physiologically plausible ranges
• Provenance: Source, device, and collection context known

Quality Assurance Processes

Process	Description	Implementation
Validation Rules	Check plausible ranges	BP 60-260/30-150, HR 30-250
Device Calibration	Regular accuracy verification	Per manufacturer schedule
Patient Verification	Confirm measurement context	Activity level, position, timing
Artifact Detection	Identify erroneous readings	Motion, interference detection

5.6 Data Storage Architecture

IoMT data storage must handle high-volume time-series data efficiently:

Storage Patterns

• Time-Series Databases: Optimized for sequential sensor data (InfluxDB, TimescaleDB)
• FHIR Servers: Healthcare-native storage with search (HAPI FHIR, Azure FHIR)
• Data Lakes: Raw storage for analytics (S3, ADLS)
• Operational Stores: Current state for real-time access
• Archival Storage: Long-term retention with tiered cost

5.7 Analytics and Insights

Transforming raw IoMT data into clinical insights:

Analytics Layers

Layer	Function	Example
Real-Time Alerts	Threshold-based notifications	SpO2 < 90% → immediate alert
Trend Analysis	Pattern detection over time	Weight gain trend in HF patient
Predictive Models	Risk prediction	Decompensation risk score
Population Analytics	Cohort insights	Program outcomes, benchmarking