A cardiologist read everything published on wearable heart rate variability and readiness scores, expecting to find a few good studies. He found none. Run the same search against every HRV trial ever registered — 3,801 of them, one dot each — and the field empties the same way.
Every dot below is a study that told the registry it would measure heart rate variability. Press play. Each step applies one condition that would have to be true before the number on your wrist could be called evidence.
Every one of them promised to record HRV, RMSSD, SDNN or pulse rate variability as a declared outcome. The median study enrolls 60 people.
Named consumer wearableGeneric wearable / PPGECG or HolterInstrument not stated
Named consumer wearable74Generic wearable / PPG117ECG or Holter192Instrument not stated3,418
Stage 0 of 4 — nothing filtered yet.
what survives all four
6
trials out of 3,801. They enrolled 738 people between them; the largest is 389. None of the six used a named consumer wearable. Five never said what measured the HRV at all.
80/20
The test that matters for a build isn’t “is HRV real?” — it is. The test is whether anyone has connected this device’s number to an outcome a clinician would act on. Four registry filters answer that in a minute, and for every consumer brand on the market today the answer is no. Write that sentence into your data dictionary next to the field, and you have done the one thing that keeps a useful integration from becoming a claim you can’t defend.
“If they believed and invested in the products they are selling, we’d not be in this position of not knowing.”
— Eric Topol, Ground Truths, September 19, 2026, on the wearable companies
Argue with this chart — here is where it is weakest
Absence in a registry is not absence of evidence. This shows that the study has not been registered in the United States under words that name it. Observational cohorts, non-US registries and company-run analyses never enter this file. Topol himself notes a WHOOP study of 2,700-plus runners that is unpublished, and an 8-million-user Fitbit analysis that is a descriptive study, not an outcome trial.
90% of the field never says what measured the HRV. 3,418 of 3,801 trials name no instrument anywhere in their title, intervention or outcome text. The grey band is not “ECG studies” and not “wearable studies” — it is the registry’s biggest hole, and it is 90% of the picture.
The classes come from regular expressions, not from reading the protocols. A trial that used an Oura ring but wrote “wrist-worn device” lands in the generic band; a trial that lists mortality only as a safety note is counted as naming a hard outcome, which makes stage 2 generous rather than strict.
“Results posted” is a legal obligation for some trials and not others. A study can be fully published in a journal and never post to ClinicalTrials.gov. Stage 4 is therefore a floor on reporting, not a measure of what is known.
Small n is the field, not a tail. 2,814 of the 3,801 trials enroll 100 participants or fewer; only 55 reach 1,000. Any trend you think you see across this field is being carried by studies too small to support it — including the six survivors, whose median enrollment is 79.5.
A dot is a plan. Enrollment is the sponsor’s target wherever the trial has not finished, and 813 of the 3,801 — more than a fifth — have a start year of 2025 or later.
The honest version of the claim
Not “wearable HRV is worthless.” Topol is explicit that some wearable metrics earned their evidence — activity, atrial fibrillation detection, resting heart rate — and that HRV from a consistent device, overnight, trended across weeks, is the defensible ceiling. The claim the registry supports is narrower and harder to dodge: as of August 2026 nobody has registered the trial that would turn a readiness score into a clinical finding, and the companies with a hundred million wrists and the budget to run it are the ones who have not.