Data Quality: What to Trust
Overview
Not all health metrics in HealthKit are equally reliable. Understanding data quality helps you interpret your health information appropriately.
Reliability Tiers
High Reliability (When Properly Collected)
Metrics with strong accuracy when sensors are worn correctly:
| Metric | Notes |
|---|---|
| Step count trends | Very good for typical walking/running |
| Heart rate trends | Accurate with good sensor contact |
| Workout duration | Directly measured time |
| GPS distance (outdoor) | Satellite-based, generally accurate |
| Sleep duration | Good estimate with consistent wear |
Moderate Reliability
Useful for trends but take individual values with caution:
| Metric | Notes |
|---|---|
| Resting heart rate | Depends on measurement conditions |
| VO₂ max estimate | Algorithm-based; trends more useful than absolutes |
| Active calories | Estimates vary 20-40% from actual |
| Sleep stages | Less accurate than polysomnography |
| HRV | Highly variable; personal baseline matters |
Use with Caution
Often noisy or prone to outliers:
| Metric | Notes |
|---|---|
| Single SpO₂ readings | Outliers common; use averages |
| Night-to-night sleep stages | High variability |
| Calorie burn estimates | Can be significantly off |
| Body fat % (BIA scales) | Affected by hydration, timing |
Best Practices
Focus on Trends
- Single readings can mislead - One data point isn't meaningful
- 7-28 day averages - Better represent true status
- Look for patterns - Consistent changes over weeks
- Compare to yourself - Your baseline is your reference
Measurement Consistency
For reliable trends: 1. Measure at same time of day 2. Use same device/method 3. Similar conditions (hydration, recent activity) 4. Consistent device placement/fit
When Single Values Matter
Some metrics are meaningful as single readings: - Blood pressure (with proper technique) - Blood glucose (for diabetics) - Temperature (during illness) - ECG (when checking for arrhythmia)
Red Flags for Data Quality
Likely Sensor Issues
- Heart rate reading very high/low at rest
- SpO₂ consistently <90% in healthy person
- Huge day-to-day swings in stable metrics
- Values far outside physiological range
User/Environmental Factors
- Device not worn properly
- Measurement during activity
- Extreme temperatures
- Wet or dirty sensors
Interpreting Changes
| Change Pattern | Likely Meaning |
|---|---|
| Gradual trend over weeks | Real physiological change |
| Sudden single-day spike | Measurement artifact likely |
| Consistent shift after lifestyle change | Probably real |
| Erratic daily variation | Normal noise; focus on averages |
When to Investigate
Trust your data more when: - Trend is consistent over 1-2+ weeks - Multiple metrics tell same story (HR up, HRV down = stress) - Correlates with how you feel - Measurement conditions were good
Question your data when: - Single dramatic reading - Contradicts how you feel - Device wasn't worn properly - Environmental factors present
Clinical vs Consumer Data
- Consumer wearables are screening tools, not diagnostic devices
- Concerning trends warrant professional evaluation
- Clinical measurements have different standards
- Don't replace medical testing with consumer devices
Summary
- Trends > single points
- Consistency improves reliability
- Know each metric's limitations
- Use multiple metrics together
- Seek clinical evaluation for concerns