Post-exertional malaise (PEM) is the core symptom of ME/CFS - a disproportionate worsening after physical or cognitive exertion that often appears only 24 to 72 hours later. That delay is what makes PEM so deceptive: by the time you notice you overdid it, it is already too late.
The Challenge
The delay between trigger and symptom makes the connection extremely hard to detect intuitively. What felt manageable yesterday can cause a crash today. This pattern creates a vicious cycle of overexertion and collapse.
Studies suggest that 90% of ME/CFS patients exceed their limits regularly because they recognize warning signs too late.
Heart Rate Variability as an Early Warning System
Heart rate variability (HRV) measures the time differences between consecutive heartbeats. It is a direct marker of autonomic nervous system activity - and therefore a window into the physiological stress state of the body.
Our analysis of HRV data shows a consistent pattern:
RMSSD (short-term HRV) drops 12 to 24 hours before noticeable PEM
The HRV/HR ratio shifts toward sympathetic dominance
Nighttime HRV recovery is clearly reduced
Day-to-day variability decreases and begins to plateau
Our Multi-Metric Approach
HRV alone is a strong signal, but not a perfect one. That is why we combine several metrics into a weighted score:
Stress Score = (HRV component x 0.50) + (HR component x 0.30) + (Activity x 0.20)The weighting follows the literature: HRV shows the strongest correlation with fatigue (r = -0.40 to -0.60), followed by resting heart rate and activity level.
HRV Component (50%)
We analyze RMSSD against your personal baseline. A drop of more than 15% below the 7-day average significantly increases PEM risk.
Heart Rate Component (30%)
Elevated resting heart rate is a classic sign of physiological stress. We track the trend over several days to smooth out short-term fluctuations.
Activity Component (20%)
Both too much and too little activity can be problematic. We compare your current activity level with what is sustainable for you personally.
The Prediction Window
Our data shows that changes in HRV patterns typically appear 24 to 72 hours before PEM is subjectively noticeable. That window gives you a chance to intervene:
At "elevated risk": reduce activity by 30% to 50%
Schedule more recovery breaks
Identify and reduce stressors
Prioritize sleep quality
Beta users report that this early warning helped them reduce PEM episodes by an average of 60%.
Limitations
We are transparent about the limitations of the approach:
HRV accuracy varies between wearables
Caffeine, alcohol, and medication influence readings
Calibration to your personal baseline takes two to three weeks
Emotional stress can create false alarms
Even with these limitations, the research shows that physiological markers such as HRV provide valuable additional information beyond self-assessment alone.
What Comes Next
We continue to improve the algorithm. As the user base grows, we can use machine learning to recognize individual patterns even more precisely. The goal is a personal early warning system that helps you avoid PEM before it happens.
You can find more technical details in our whitepaper on PEM risk prediction.