Editorial illustration of overnight oxygen report fields with estimates, time window, events, and signal quality

Reference data dictionary · version 1.0 · 11 September 2026

Short answer: do not read a headline overnight oxygen number without its definition, time window, event rule, and signal-quality information. SpO2 is an estimate, ODI is not AHI, and a minimum or time-below value cannot explain its own cause.

This reference is designed to be copied into a report, shared with a reader, or used as a checklist when two apps use similar labels for different calculations.

Download the data dictionary

The CSV maps common fields to a plain-language definition, a method question, and a limitation.

Download CSV

The field-by-field dictionary

FieldPlain-language definitionQuestion to askCommon limitation
SpO2An optical estimate of peripheral oxygen saturation.Where, how, and under what conditions was it measured?Motion, low perfusion, temperature, skin pigmentation, nail polish, and fit can affect the estimate.
PulseThe pulse rate detected by the same sensor.Was the pulse signal stable at the same moment?A plausible pulse does not validate every oxygen data point.
Recording timeThe period during which the device was collecting data.Does it represent time worn, recording time, or estimated sleep time?Recording time is not automatically sleep time.
Minimum SpO2The lowest displayed value in the selected period.How long did the value last and was the signal usable?A single minimum can be artifact, movement, or a short event.
Average or median SpO2A summary of the values over a period.Which values were included or excluded?An average can hide short repeated drops or long invalid sections.
ODI 3% or ODI 4%Desaturations meeting a selected percentage-drop rule per analyzed hour.Which drop definition, baseline, duration, and denominator were used?ODI is not AHI and is not calculated identically by every device.
Time below thresholdMinutes or a percentage below a displayed saturation threshold.Which threshold and usable-time window are shown?The number needs patient, altitude, device, and symptom context.
Signal qualityUsable, weak, missing, or artifact-affected data.What did the software exclude?A long recording with poor usable signal may be less informative than a shorter clean one.
Event markerA software flag created when a rule is met.What rule triggered the flag?An event marker is a classification, not a confirmed cause.
Visual data dictionary for SpO2, pulse, time, quality, and symptoms in an overnight oxygen report
A number becomes more portable when its method and limitations travel with it.

Why ODI and AHI must stay separate

ODI counts oxygen desaturations that meet a selected drop rule, often 3% or 4% from a baseline, per hour of analyzed time. AHI counts apneas and hypopneas per hour of sleep using a sleep-study method. The metrics may correlate in some settings, but they are not the same measurement and should not be swapped in a summary.

The systematic review of ODI for adult obstructive sleep apnea found substantial variation in study populations, thresholds, and diagnostic performance. A practical report should therefore include the rule behind the ODI rather than presenting a bare number. If a page says “ODI 12,” the reader still needs to know whether it means 3% or 4%, whether the denominator was recording time or estimated sleep time, and how artifact was handled.

The AASM also cautions that diagnosis, treatment efficacy, and treatment decisions should not be based only on automatically scored home sleep apnea data. The safe interpretation is that a report can provide structured evidence for clinical review, not that it can independently name the cause.

Comparison showing why two overnight oxygen reports with the same label may use different methods
Two reports can use the same field name while applying different thresholds, denominators, sampling, or artifact rules.

A method card to attach to every report

  1. Device and software: record the device model, sensor location, and app or firmware version where available.
  2. Recording window: write start time, end time, wear time, and whether the report estimates sleep time.
  3. Event rule: keep the ODI drop threshold, duration rule, and baseline definition.
  4. Quality: note usable signal, artifact handling, missing sections, and sensor-fit issues.
  5. Context: add symptoms, illness, altitude, sleep position, and the reason monitoring was performed.
  6. Comparison rule: do not compare two devices as though they were equivalent unless their methods are actually aligned.

Estimate

SpO2 is a sensor estimate. Preserve the reading but do not confuse it with a direct blood measurement.

Rule

ODI and event markers are rule-based outputs. Keep the threshold and denominator visible.

Window

Recording time, wear time, and sleep time are not automatically interchangeable.

Quality

Invalid or uncertain segments should remain visible instead of disappearing from the story.

Reuse and citation

Suggested citation: NightlyVitals, “Overnight Oxygen Report Field Definitions: A Reusable Data Dictionary,” September 11, 2026. Definitions are educational and do not replace clinical interpretation.

FAQ

What does SpO2 mean on an overnight report?

It is an estimate of peripheral oxygen saturation from an optical sensor. Review it with pulse, signal quality, time window, symptoms, and the device instructions.

Is ODI the same as AHI?

No. ODI counts selected oxygen drops; AHI counts breathing events under a sleep-study method. They are related in some settings but are not interchangeable.

Does time below a threshold prove a diagnosis?

No. It is a report metric whose meaning depends on the threshold, duration, usable signal, patient context, and clinical assessment.

How to quote a report without stripping away the method

A shareable report excerpt should carry more than one number. A compact format is: “ODI 3%: 8.2 events per hour, analyzed recording time 6 h 42 min, usable signal 93%, device and software version noted.” If the report does not expose one of those fields, say that it is unavailable instead of filling the gap with an assumption.

The same rule applies to minimum SpO2 and time below a threshold. Write the value, the duration or percentage, the analysis window, and the quality flag. “Minimum 88% for one sample, fair signal” is a different statement from “spent 4 minutes below 90% with good usable signal.” A careful reference page helps readers see that these are not interchangeable summaries.

Method card template

QuestionRecord thisIf missing
What device made the measurement?Model, sensor site, app or firmwareDo not compare it as if it were identical to another device
What period was analyzed?Start, end, wear time, sleep-time estimateLabel the denominator as unknown
What counts as an event?Drop threshold, duration, baseline, recoveryCall the metric device-defined
What data was excluded?Artifact, missing, weak, or invalid timeKeep the quality limitation beside the result

Why definitions are a linkable asset

Publishers often need a neutral reference they can cite without endorsing a product or diagnosing a reader. A field-level data dictionary serves that need because it answers a narrow question, links to primary sources, and makes its limitations visible. It can be quoted in a patient education page, a research explainer, or a device-support article while preserving the difference between an estimate and a clinical conclusion.

Keep this page stable when the definitions do not change. If a device vendor changes its event rule, add the new rule and the date rather than silently rewriting the old explanation. That gives readers a clean citation trail and makes future updates easier to audit.

Related reading

sources

Sources and verification

Last checked: September 11, 2026