Blood oxygen (SpO2) and respiratory rate are two metrics Apple Watch captures that get less attention than HRV and sleep, but they’re meaningfully informative — particularly for spotting illness onset, altitude effects, and sleep quality indicators. Here’s how to query and interpret them.
Blood Oxygen (SpO2)
Apple Watch Series 6 and later measures blood oxygen saturation. The HealthKit identifier is HKQuantityTypeIdentifierOxygenSaturation. Values are expressed as a fraction (0.98 = 98% SpO2).
Measurement context: Apple Watch measures SpO2 in the background during sleep and when you manually take a reading via the Blood Oxygen app. During sleep, it measures every few minutes. The background measurements are automatic when Sleep Mode is active.
What’s normal: Healthy resting SpO2 is typically 95-100%. Below 90% is clinically significant. For Apple Watch’s consumer-grade sensor, expect ±1-2% accuracy — it’s useful for trend monitoring and flagging unusually low readings, not for clinical decisions.
Querying SpO2:
Show me my blood oxygen levels over the last 30 days.Claude calls query_metric(metric_type="HKQuantityTypeIdentifierOxygenSaturation", days=30).
For a 30-day window, this returns daily aggregates — the day’s average SpO2, min, and max. The min value is often the most informative, as it captures the lowest reading during the night.
Checking for nocturnal dips:
Were there any nights in the last 90 days where my blood oxygen dropped below 94%?Claude uses search_records(metric_type="HKQuantityTypeIdentifierOxygenSaturation", max_value=0.94, days=90).
Note: SpO2 in HealthKit is stored as a decimal fraction, not a percentage — 0.94 = 94%, 0.98 = 98%.
Altitude effects:
If you travel to high altitude (ski trips, mountain hiking), SpO2 often drops as the body acclimates.
I went skiing in Colorado in early March at around 9,000 feet. Can you see any SpO2 changes during that trip?Claude pulls query_metric(metric_type="HKQuantityTypeIdentifierOxygenSaturation", days=30) (or uses compare_periods with the travel dates) and looks for SpO2 dips in the relevant window.
Illness correlation:
SpO2 can drop slightly during respiratory illness (cold, flu, COVID).
Did my blood oxygen show any patterns during January when I was sick?Claude uses compare_periods with pre-illness, during-illness, and recovery dates.
Respiratory Rate
Respiratory rate is measured by Apple Watch Series 3 and later during sleep. The HealthKit identifier is HKQuantityTypeIdentifierRespiratoryRate. Unit: breaths per minute.
When it’s measured: Apple Watch measures respiratory rate during sleep — typically every few minutes while you’re lying still. You won’t see readings during your active day; this is a sleep-time metric.
What’s normal: Resting respiratory rate for adults is typically 12-20 breaths/minute. During sleep, it’s often in the lower end of that range. Elevated respiratory rate during sleep can indicate illness, stress, or sleep-disordered breathing.
Querying respiratory rate:
What's my respiratory rate trend over the last 60 days?Claude calls query_metric(metric_type="HKQuantityTypeIdentifierRespiratoryRate", days=60).
For a 60-day window, this returns daily averages. You’ll see your baseline (typically 14-17 breaths/min for most healthy adults) and any deviations.
Illness detection:
Elevated respiratory rate during sleep is one of the earliest indicators of respiratory illness — often appearing before you feel sick.
Show me days in the last 90 days where my nighttime respiratory rate was above 17 breaths per minute.Claude uses search_records(metric_type="HKQuantityTypeIdentifierRespiratoryRate", min_value=17, days=90).
Correlation with sleep and HRV:
On nights where my respiratory rate was elevated (above 16), how was my HRV the next morning?Claude identifies high-respiratory-rate nights from query_metric, then checks the HRV for the following morning from get_hrv_trend. Elevated respiratory rate and suppressed HRV often co-occur during illness or high stress — seeing both simultaneously strengthens the signal.
Combining SpO2, Respiratory Rate, and HRV
The three metrics together form a respiratory health picture:
- Normal SpO2 + normal respiratory rate + normal HRV = everything is fine
- Low SpO2 + elevated respiratory rate + suppressed HRV = potential respiratory illness or sleep disruption
- Normal SpO2 + elevated respiratory rate + suppressed HRV = could be stress, anxiety, or early illness
- Low SpO2 + normal respiratory rate = possible altitude, or the reading was an artifact
Look at my SpO2, respiratory rate, and HRV over the last 7 days.
Is there any pattern that would suggest I might be getting sick?Claude pulls all three metrics and looks for the co-occurrence pattern. If SpO2 has been creeping down (98% → 97% → 96%) and respiratory rate has been elevated for 2 consecutive nights, that’s a flag worth noting — even if you feel fine.
This kind of early-warning analysis from biometric trends is genuinely useful, even with the caveats about Apple Watch measurement accuracy. The trend direction is more informative than any single absolute reading.
health4ai: free on the App Store.
Download on the App Store →