How to Interpret Wearable Health Data
A practical guide to getting meaningful insights from your Apple Watch and iPhone health metrics.
The Four Principles
1. Focus on Trends, Not Single Points
Wearable signals have noise. A single data point is just a sample—what matters is the trend.
Recommended visualizations: - 7-day moving average - Short-term trend - 28-day moving average - Long-term baseline - Change from baseline - Today vs your 28-day average - Percentile bands - Your typical range (5th to 95th percentile)
Example: If your resting heart rate is 58 bpm one morning but your 7-day average is 62 bpm, the single reading doesn't indicate a problem—it's normal variation. If your 7-day average drops from 62 to 55 over a month, that's a meaningful signal.
2. Context Is Non-Optional
The same number can mean completely different things depending on:
| Factor | Effect |
|---|---|
| Sleep debt | Elevates HR, lowers HRV |
| Caffeine/alcohol | Affects HR, HRV, sleep |
| Hydration | Affects HR, HRV |
| Stress | Elevates HR, lowers HRV |
| Illness | Affects almost everything |
| Medications | Beta blockers, stimulants, etc. |
| Altitude/temperature | Affects SpO₂, HR |
| Sensor contact quality | Affects accuracy |
Best practice: When you see an unusual reading, ask "What else was different today?"
3. Different Metrics Have Different Reliability
Not all wearable measurements are equally accurate. A living systematic review (82 studies, >430,000 participants) found:
| Metric | Accuracy |
|---|---|
| Heart rate | Small mean bias, moderate variability |
| Sleep duration | Moderate accuracy |
| Step count | Moderate accuracy |
| Energy expenditure | Large, inconsistent error |
Implications: - Trust more: Heart rate trends, step count trends, sleep duration - Trust less: Absolute calorie burns, single SpO₂ readings, exact sleep stage minutes
4. Use Clinical Thresholds Carefully
Guideline thresholds (blood pressure, glucose, etc.) were developed using: - Clinical-grade devices - Standardized measurement protocols - Controlled conditions
Wearables can support awareness and tracking, but confirmation should use validated clinical devices when decisions depend on the result.
Recommended Derived Metrics
Beyond raw data, these calculations add value:
Baselines
- 7-day average - Recent trend
- 28-day average - Personal baseline
- Personal percentiles - What's normal for you
Deltas
- Today vs baseline - Am I different today?
- This week vs last week - Short-term change
- This month vs last month - Longer-term change
Consistency Metrics
- Day-to-day variability - Is the metric stable?
- Completeness - Are there gaps in data?
- Measurement conditions - Were conditions consistent?
When Single Readings Matter
Some metrics are meaningful as single readings when properly measured: - Blood pressure (with correct technique) - Blood glucose (for diabetes management) - Temperature (during suspected illness) - ECG (checking for arrhythmia) - Weight (though trends still preferred)
Red Flags in Wearable Data
Likely Measurement Issues
- Heart rate extremely high/low during rest
- SpO₂ consistently <90% when feeling fine
- Huge day-to-day swings in stable metrics
- Values outside physiological possibility
Potentially Real Signals
- Gradual trend over 1-2+ weeks
- Multiple metrics telling same story
- Correlates with symptoms
- Persists across measurement conditions
Practical Workflow
- Review weekly, not daily - Reduces noise anxiety
- Note lifestyle factors - Sleep, stress, illness, travel
- Compare to yourself - Your baseline is your reference
- Look for patterns - Multiple metrics changing together
- Share concerns - Discuss with healthcare provider
Summary
| Do | Don't |
|---|---|
| Focus on trends | Obsess over single readings |
| Compare to your baseline | Compare to population averages |
| Consider context | Interpret numbers in isolation |
| Use multiple metrics together | Rely on one number |
| Share concerning patterns | Self-diagnose |