What does this whitepaper answer first?
This page should answer the core questions behind the PEM risk model first, before moving into the deeper technical and research-oriented framing.
Short answer
What does the PEM risk model do in one sentence?
The model combines HRV, resting heart rate, and activity patterns against the personal baseline to surface elevated PEM risk earlier.
Usefulness
Why are wearable signals useful here at all?
Because PEM often appears with delay, and symptom awareness alone may come too late. Physiological signals can add earlier clues before the full worsening becomes obvious.
Limit
What does the model explicitly not do?
It does not replace diagnosis. The model is a structured risk and observation layer that still needs to be interpreted with symptoms, course, and context.
Abstract
This whitepaper outlines an algorithm for predicting post-exertional malaise (PEM) risk using HRV, resting heart rate, and activity data. The goal is earlier overload detection in ME/CFS, Long COVID, and related fatigue conditions.
1. Why PEM Forecasting Is Hard
PEM often appears with a delay after exertion and is experienced subjectively by patients. That makes purely symptom-based early warning difficult.
Wearable signals add another layer by capturing physiological deviations earlier and more consistently than point-in-time self-report alone.
2. Model Design
The approach combines three signal groups: heart rate variability, resting heart rate, and activity patterns. Rather than interpreting them separately, they are aggregated against each user’s personal baseline.
The resulting stress indicator brings autonomic dysregulation and unusual exertion patterns together in a single score.
3. Use in Elara
Elara uses the score to monitor exertion trends longitudinally and surface elevated-risk phases. The model is not intended as a diagnostic substitute, but as a structured observation and decision-support layer.
Combined with symptom tracking, it creates a stronger basis for pacing decisions, trajectory review, and research analysis.
Key Points
Multiple biomarkers are more robust than any single signal alone.
Comparison against the individual baseline is essential.
The model targets risk assessment and early warning, not diagnosis.
References
Davenport TE, Stevens SR, VanNess JM, Stevens J, Snell CR. (2019)
"Conceptual model for post-exertional malaise in myalgic encephalomyelitis/chronic fatigue syndrome"
Work, 62(3): 505-513
Davis HE, McCorkell L, Vogel JM, Topol EJ. (2023)
"Long COVID: major findings, mechanisms and recommendations"
Nat Rev Microbiol, 21(3): 133-146
Rao N, et al. (2020)
"Heart rate variability as a biomarker for fatigue in chronic fatigue syndrome: A pilot study"
Journal of Translational Medicine, 18: 412
Komaroff AL, Lipkin WI. (2021)
"Insights from myalgic encephalomyelitis/chronic fatigue syndrome may help unravel the pathogenesis of postacute COVID-19 syndrome"
Trends Mol Med, 27(9): 895-906
Wirth K, Scheibenbogen C. (2021)
"A unifying hypothesis of the pathophysiology of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS)"
Autoimmun Rev, 20(1): 102527
The complete internal whitepaper includes additional references and implementation details.