When we started building Elara, one obvious research question was whether wearable signals such as HRV or resting heart rate could flag an upcoming episode of post-exertional malaise (PEM) early enough to serve as a warning system. With more longitudinal data, the answer is more nuanced today than it was at the start. This article, originally published on January 8, 2026, was updated accordingly on September 14, 2026.
In short: HRV and resting heart rate can add physiological context, but in our analyses so far they are not robust stand-alone precursors of a later PEM episode.
What HRV actually measures
Heart rate variability describes variation in the time between heartbeats and is often used as an indirect signal of autonomic nervous-system state. It is also influenced by many other factors, including sleep, infections, medication, hydration, psychological stress, measurement conditions, and the wearable itself.
A single HRV value is therefore neither a PEM test nor a universal exertion threshold. It is more useful when viewed against a personal baseline and alongside other information.
What we learned from Elara trajectories
Our current longitudinal analyses point to the individual trajectory as being more informative than any single wearable signal. A composite measure of actual daily demand covaries within the same person with symptom state more clearly than HRV or resting heart rate alone.
That matters because it differs from the earlier hypothesis of a fixed early-warning window: at present we do not see a basis for reliably predicting from HRV alone whether PEM will occur 24 to 72 hours later.
Why wearables can still be useful
They document activity, sleep, heart rate, and HRV with little additional logging effort.
They make changes from a personal baseline visible across days and weeks.
They add context to self-reported energy, symptoms, exertion, and recovery.
They can help with retrospective questions: what was different before a worse day?
How Elara uses these data today
Elara brings wearable data and self-report onto one timeline. The emphasis is on trajectories, exertion, and the personal baseline — not on a medical PEM prediction derived from HRV.
That distinction matters for pacing. A number should not create the impression of a guaranteed green light for the day. Personal experience, known limits, symptoms, and the context of the day remain central.
What we explicitly do not claim
No wearable metric diagnoses PEM or ME/CFS.
Elara currently has no validated 24–72-hour PEM prediction.
A particular HRV drop is not a universal PEM threshold.
Observational data cannot establish that an app signal prevents or reduces PEM episodes.
What needs to be tested next
A true prediction has to be tested prospectively: the model must make its estimate before the event, perform on previously unseen data, and demonstrate useful discrimination beyond simple baselines. Keeping descriptive pattern recognition separate from predictive validation is central to our next research steps.
Conclusion
Wearables remain valuable to Elara, but primarily as additional trajectory and context signals rather than a crystal ball. A restrained interpretation is more useful than a score that promises more certainty than the data can support.