WEARABLES & IOT CONNECTORS

Data Types, Mapping & Time-Series

5 of 7

Transform raw device measurements into standardized, interoperable longitudinal health data with FHIR mapping, terminology and unit normalization, provenance, quality controls and scalable time-series handling.

Supported Data TypesMapping to FHIRTerminologies & UnitsTime-Series HandlingData Quality & ProvenanceExamples

Data Normalization Flow

Device Measurement

Raw value, unit, source timestamp and device/provider identifier

Source & Context

Capture source, patient/device link, original code/unit and connection method

MSIS Normalization

Validated terminology/unit mapping, deduplication and quality flags

FHIR Resource

Observation / Device / Provenance with normalized and original source context

Mediloop Platform

Longitudinal record, analytics, alerts and care workflows

Supported Data Types

Category	Examples	FHIR resource	Typical units
Vital signs	Heart rate, BP, SpO₂, respiratory rate, temperature	Observation	/min, mmHg, %, °C
Activity & fitness	Steps, distance, calories, active minutes	Observation	count, km, kcal, min
Sleep	Stages, duration, sleep score	Observation	h/min, score
Glucose	CGM readings, trends, alerts	Observation	mg/dL, mmol/L
ECG / waveform	Single/multi-lead ECG, rhythm	Observation + attachment/derived representation	mV, Hz
Respiratory	Respiration rate, FEV1, FVC, peak flow	Observation	/min, L, L/s
Body composition	Weight, BMI, body fat, muscle mass	Observation	kg, kg/m², %
Device / environment	Device events, air quality, ambient sensors	Observation / DeviceMetric	domain-specific

Mapping to FHIR

Example mapping

Field	Source	FHIR mapping
Measurement	72 bpm	Observation.valueQuantity
Code	Heart rate	Observation.code — e.g. validated LOINC mapping
Unit	bpm	UCUM code /min
Timestamp	Device/provider UTC timestamp	Observation.effectiveDateTime
Device	Provider device id	Observation.device → Device
Source	HealthKit / provider API / BLE	Provenance / connector metadata
Quality	Provider/device quality	Observation interpretation/extension or provenance metadata as designed

FHIR Observation example

jsonCopy

{
  "resourceType": "Observation",
  "status": "final",
  "code": {
    "coding": [{
      "system": "http://loinc.org",
      "code": "8867-4",
      "display": "Heart rate"
    }]
  },
  "subject": { "reference": "Patient/pat_123" },
  "device": { "reference": "Device/dev_456" },
  "effectiveDateTime": "2026-08-31T08:00:00Z",
  "valueQuantity": {
    "value": 72,
    "system": "http://unitsofmeasure.org",
    "code": "/min"
  }
}

Terminologies & Units

LOINC for validated observation mappings where an appropriate code exists.

SNOMED CT for clinical concepts/device contexts where applicable.

UCUM for units of measure.

Preserve original manufacturer/provider codes and units alongside normalized mappings.

Version mappings and record provenance; do not silently replace uncertain mappings.

Support multi-mapping or unresolved-code states when no safe canonical mapping exists.

Time-Series Handling

Support high-frequency and periodic measurements with explicit sampling metadata.

Store raw/source time and normalized UTC time; keep time-zone context when relevant.

Detect missing samples, out-of-order arrivals and clock drift.

Support real-time streaming and batch ingestion with idempotent deduplication.

Separate raw data from calculated/aggregated views.

Apply configurable retention/downsampling policies appropriate to use case and governance.

Use scalable time-series storage without changing the external FHIR/API contract.

Data Quality & Provenance

Validate expected ranges, units and data types.

Detect duplicate, anomalous and implausible measurements.

Record device, provider, connection path and ingestion timestamp.

Keep original values alongside normalized values.

Maintain mapping version, transformation lineage and confidence/quality indicators.

Respect patient consent and retention policy.

Do not promote consumer-wellness data to clinical-grade status without appropriate evidence/validation.

Developer Integration Surface

Surface	Operation / resource	Use	Status
FHIR	Observation / Device / DeviceMetric / Provenance	Normalized longitudinal data exchange	FHIR surface
Bulk / streaming ingestion	Batch samples, MQTT/webhook events, local gateway upload	High-volume time-series ingestion	Mediloop contract planned
Query API	Patient/device/date/code filters and pagination	Application access to normalized measurements	Mediloop contract planned
Events	measurement.created, device.status.changed, data.gap.detected	Real-time workflows and monitoring	Planned event contract
SDK	ingest(), observations.list(), devices.get()	Type-safe developer convenience surface	SDK contract evolving

Examples

Illustrative time-series query

httpCopy

GET /v1/observations?patient_id=pat_123&code=8867-4&from=2026-08-01T00:00:00Z&to=2026-08-31T23:59:59Z
Authorization: Bearer <token>
X-Tenant-Id: <tenant-id>

# Planned Mediloop REST façade; FHIR search remains the standards surface.

Illustrative ingestion

typescriptCopy

await client.wearables.ingest({
  patientId: 'pat_123',
  deviceId: 'dev_456',
  measurements: [{
    sourceCode: 'heart_rate',
    value: 72,
    unit: 'bpm',
    observedAt: '2026-08-31T08:00:00Z'
  }],
  preserveSource: true,
});

Next Steps

Implement mappings

Target devices/providers, FHIR resources and terminology

Configure time series

Sampling, retention, aggregation and query patterns

Set quality rules

Ranges, gaps, duplicates, provenance and data classification

Test real data

Device clocks, missing data, provider delays and edge cases

Connect workflows

Patient dashboards, RPM, alerts and EHR interoperability