Every device reading, in the chart and in the study.

The CGM, monitor or wearable reports a value once. Kymolog codes it to FHIR, files it in Epic, Oracle Health or MEDITECH, and keeps the same reading, with its device and timestamp, for research.

Request a demo How it works

FHIR R4 · LOINC · UCUM · on premises or cloud

Fig. 1. One reading, recorded once, filed twice.

Medical device data analytics starts with a coded reading

A continuous glucose monitor (CGM) reports 132 mg/dL at 08:02. The Kymolog™ edge agent writes that value as one FHIR R4 Observation. FHIR, short for Fast Healthcare Interoperability Resources, is the HL7 standard for exchanging health records, and an Observation is its record of one measurement.

This Observation carries a LOINC code for what was measured, a UCUM unit for the value, the device that took the reading and when. Kymolog then adds the patient and the encounter. No one retypes or recodes the value later, so the chart and the research dataset both use this record.

The anatomy of one ObservationA FHIR R4 Observation for one glucose reading, set as JSON on a card, with six fields called out: the LOINC code 99504-3 says what was measured, glucose in interstitial fluid; the patient reference Patient/p-041 is matched to the patient by bed or by the sensor’s assignment; the effective time 08:02 is when the device took the reading; the value 132, in red, is the reading itself; the UCUM unit mg/dL is its coded unit; and the device reference Device/7f3a names the device that sent it.FHIR R4 OBSERVATION{  "resourceType": "Observation",  "status": "final",  "code": { "coding": [{    "system": "http://loinc.org",    "code": "99504-3" }] },  "subject": { "reference": "Patient/p-041" },  "effectiveDateTime": "2026-10-03T08:02:00-04:00",  "valueQuantity": {    "value": 132,    "unit": "mg/dL",    "system": "http://unitsofmeasure.org",    "code": "mg/dL" },  "device": { "reference": "Device/7f3a" }}LOINC 99504-3glucose in interstitial fluidpatient referencematched by bed or assignmenteffective time 08:02when the device took itvalue 132the reading itselfUCUM mg/dLits unit, codeddevicewhich device sent it
The anatomy of one ObservationA FHIR R4 Observation for one glucose reading, set as JSON on a card, with six fields called out: the LOINC code 99504-3 says what was measured, glucose in interstitial fluid; the patient reference Patient/p-041 is matched to the patient by bed or by the sensor’s assignment; the effective time 08:02 is when the device took the reading; the value 132, in red, is the reading itself; the UCUM unit mg/dL is its coded unit; and the device reference Device/7f3a names the device that sent it.FHIR R4 OBSERVATION{  "resourceType": "Observation",  "status": "final",  "code": { "coding": [{    "system": "http://loinc.org",    "code": "99504-3" }] },  "subject": {    "reference": "Patient/p-041" },  "effectiveDateTime":    "2026-10-03T08:02:00-04:00",  "valueQuantity": {    "value": 132,    "unit": "mg/dL",    "system": "http://unitsofmeasure.org",    "code": "mg/dL" },  "device": {    "reference": "Device/7f3a" }}1234561LOINC 99504-3glucose in interstitial fluid2patient referencematched by bed or assignment3effective time 08:02when the device took it4value 132the reading itself5UCUM mg/dLits unit, coded6devicewhich device sent it
Figure 2 as a table
FieldValueWhat it records
codeLOINC 99504-3glucose in interstitial fluid
subjectPatient/p-041the patient, matched by bed or device assignment
effectiveDateTime2026-10-03T08:02:00-04:00when the device took the reading
valueQuantity.value132the reading itself
valueQuantity.unitmg/dL (UCUM)its unit, coded
deviceDevice/7f3athe device that sent it
Fig. 2. One reading as Kymolog files it: a FHIR R4 Observation that carries its LOINC code, UCUM unit, device, effective time and patient. The chart and the study receive this same record.

Into the chart

The edge agent sits beside the devices and picks up each value as the device reports it. Kymolog matches the reading to the patient and encounter by bed, then files it in the EHR the hospital already runs.

  • Epic
  • Oracle Health
  • MEDITECH
  • athenahealth
  • eClinicalWorks
  • NextGen
  • Veradigm

How medical device integration works

Into the study

Kymolog keeps the record it files in the chart for research too. Each reading keeps its device and timestamp and stays linked to its patient and encounter, so medical device analytics can trace every value to its source.

Sites that want to share data can add the data network: consent, de-identification, dataset licensing and royalties to contributing sites, built from the readings Kymolog already keeps.

How Kymolog records provenance

A CGM reading in the chart and in the study datasetOne CGM reading, 132 mg/dL taken at 08:02 by Device/7f3a, leaves its recording strip and lands in two places. Panel a, into the chart: delivered over FHIR R4 or HL7 v2, it fills the 08:02 cell of the "Glucose, CGM" row in the flowsheet of patient p-041 in bed 12, after 128 at 07:52 and 130 at 07:57. Panel b, into the study: the same record is kept as a dataset row with patient p-041, effective time 08:02:00, LOINC 99504-3, value 132 mg/dL and device 7f3a, under the rows for 07:52 and 07:57. The value 132 is in red in both places.132 mg/dL08:02:00, CGMDevice/7f3aone readingaInto the chartFHIR R4 or HL7 v2bInto the studythe same recordFLOWSHEETp-041 · bed 1207:5207:5708:02Glucose, CGM mg/dL128130132SpO₂ %979696STUDY DATASETdevice readingsPATIENTEFFECTIVELOINCVALUEDEVICEp-04107:52:0099504-3128 mg/dL7f3ap-04107:57:0099504-3130 mg/dL7f3ap-04108:02:0099504-3132 mg/dL7f3a
A CGM reading in the chart and in the study datasetOne CGM reading, 132 mg/dL taken at 08:02 by Device/7f3a, leaves its recording strip and lands in two places. Panel a, into the chart: delivered over FHIR R4 or HL7 v2, it fills the 08:02 cell of the "Glucose, CGM" row in the flowsheet of patient p-041 in bed 12, after 128 at 07:52 and 130 at 07:57. Panel b, into the study: the same record is kept as a dataset row with patient p-041, effective time 08:02:00, LOINC 99504-3, value 132 mg/dL and device 7f3a, under the rows for 07:52 and 07:57. The value 132 is in red in both places.132 mg/dL08:02:00, CGMDevice/7f3aaInto the chartFHIR R4 or HL7 v2bInto the studythe same recordFLOWSHEETp-041 · bed 1207:5207:5708:02Glucose, CGM mg/dL128130132SpO₂ %979696STUDY DATASETdevice readingsPATIENTEFFECTIVELOINCVALUEDEVICEp-04107:52:0099504-3128 mg/dL7f3ap-04107:57:0099504-3130 mg/dL7f3ap-04108:02:0099504-3132 mg/dL7f3a
Figure 3 as a table
Where it landsWhat it recordsValue
the chartflowsheet row "Glucose, CGM", 08:02 column, patient p-041 in bed 12132 mg/dL
the studydataset row: patient p-041, 08:02:00, LOINC 99504-3, device 7f3a132 mg/dL
Fig. 3. Kymolog files the CGM value in the patient’s chart and keeps the same record, with its device, time and LOINC code, as a row in the study dataset.

What Kymolog records, and where it runs

The edge agent reads CGMs, electrocardiogram (ECG) monitors, blood pressure cuffs, thermometers and pulse oximeters, which report oxygen saturation (SpO₂). Any other device class joins by configuration: a manifest of its codes and units, with no custom build. See the devices Kymolog records.

Each part of the medical device data platform runs where its operator puts it. A hospital runs the install on its own servers or private cloud, or a device maker runs it in the maker's cloud. The research store sits beside the install or in our cloud, which only ever receives de-identified readings. The agent sends device values without patient details, and patients are matched inside the install. Read what the install exposes, and to whom.

Who uses Kymolog

  • Device makers

    Deliver your device's readings to hospital EHRs and keep their provenance for post-market surveillance.

  • Hospitals

    Chart bedside device readings without retyping, each one linked to the patient in that bed.

  • Researchers

    Study continuous physiological data from ICU and ward beds, coded at capture and linked to the encounter.

  • Trial sponsors

    Collect wearable and home-device readings with the device and time recorded on each one.

Request a Kymolog demo

Tell us which devices you use and which EHR you run. We reply within one business day with times to talk. In the demo, a reading from one of your device classes goes into a test chart and a research dataset.

Request a demo