Peer reviewed AI-assisted

Hypertension detected from consumer smartwatch sensors without a cuff

A Google model reading photoplethysmography and accelerometer signals from a consumer smartwatch, with no cuff calibration, flagged hypertension in 196 prospectively enrolled participants with 65.8% sensitivity and 90.0% specificity against 24-hour ambulatory blood pressure monitoring, slightly ahead of initial office blood pressure screening at the same specificity.

Model
AI-PPG-ACC-HTN
Field
Medicine
Date
2026-11-05
Human collaborators
Paolo Di Achille, Stephen P. Juraschek, Ming-Zher Poh

What was found

More than half of adults with hypertension do not know they have it, and the condition produces no symptoms to prompt a measurement. The model takes the optical pulse and motion waveforms a smartwatch already records during ordinary wear, with no blood pressure cuff used to calibrate it to the individual, and was evaluated in a prospective multicentre study of 196 participants free of known cardiovascular disease and not taking antihypertensive medication, against 24-hour ambulatory monitoring as the reference standard, over seven days of real-world wear. It reached sensitivity 65.8% (95% CI 54.0 to 76.3), specificity 90.0% (95% CI 83.2 to 94.7) and positive predictive value 80.6% (95% CI 68.6 to 89.6), where initial office blood pressure screening in the same study reached sensitivity 55.3% at the same specificity. The authors frame the use as opportunistic screening that prompts people toward a clinical assessment and cuff-based confirmation, not as a measurement of blood pressure.

Novelty check

Europe PMC, 2026-11-09: TITLE:(smartwatch OR "smart watch" OR wearable OR photoplethysmography) AND TITLE:(hypertension) AND ("ambulatory blood pressure") returned 8 records, all reviews, position papers or cuffless blood pressure measurement devices rather than a cuffless hypertension classifier validated prospectively against ambulatory monitoring; a broader query without the title restriction returned 181 records dominated by cuffless blood pressure estimation, a different task, including a 2025 American Heart Association scientific statement on those devices. The distinction matters: the prior literature largely tries to output a blood pressure number, while this model outputs a hypertension flag. No entry in this registry covers wearable hypertension detection. Other vendors ship hypertension notification features on consumer watches, which were not evaluated in this paper and are not the subject of this entry.

Caveats and known objections

Developed, run and authored by Google Research, with one academic co-author at Beth Israel Deaconess Medical Center, so this is the developer's own prospective study and not an external validation. The cohort is small at 196 participants and deliberately unrepresentative of the people a screening tool would meet in practice: those with known cardiovascular disease or already on antihypertensive medication were excluded. Sensitivity of 65.8% means about a third of hypertensive participants were missed, and the absence of a flag carries little information. The endpoint is detection accuracy against a reference standard, a surrogate: no clinical outcome, no diagnosis rate and no treatment change was measured, and whether a flag leads anyone to a clinic was not studied. Neither the model weights nor the sensor data are available outside the company. Autonomy is ai-assisted: the model raises a screening flag that a clinician and a cuff-based measurement must confirm.

No independent check is recorded on this entry yet.

Reading the primary source closely enough to say whether it supports the claim counts as a check, and you are credited on the entry.

Or on GitHub: submit a check challenge the grade send a correction or send a pull request

Entry history (1 event)
  1. AddedEntered the registry graded Peer reviewed and AI-assisted.

Entries are never deleted. A grade that does not hold up is downgraded on the record, with the reason beside it.

Graded peer reviewed for verification and ai-assisted for autonomy. What these mean.

Cite this entry

Plain text
whataifound.org. (2026). Hypertension detected from consumer smartwatch sensors without a cuff. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2026-11-05-smartwatch-hypertension-detection
BibTeX
@misc{whataifound-googleresearch-2026-detection,
  title        = {Hypertension detected from consumer smartwatch sensors without a cuff},
  author       = {{whataifound.org}},
  year         = {2026},
  howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
  note         = {Result by Google Research. Verification: Peer reviewed. Autonomy: AI-assisted.},
  url          = {https://whataifound.org/finding/2026-11-05-smartwatch-hypertension-detection}
}

Related findings

← All medicine findings in the registry