Di Achille, Paolo, Cai, Lawrence, Wu, Jiang, Gao, Mingwu, Daryani, Bhavna, Wang, Jonathan, Khanna, Utkarsh, Ko, Ho, Pathak, Anupam, Malhotra, Mark, Patel, Shwetak, Shreibati, Jacqueline B., Juraschek, Stephen P., Rudrapatna, Pramod, Thompson, Matthew, Poh, Ming-Zher · Nov 5, 2026
Authors of a Nature paper report smartwatch screening for signs of hypertension
Markdown · JSON · Original source
The authors of a paper report that artificial intelligence (AI) applied to photoplethysmography and accelerometry recorded by a consumer smartwatch detected hypertension without cuff calibration. They tested AI-PPG-ACC-HTN in a prospective study of 196 participants without known cardiovascular disease or antihypertensive medication, comparing it with 24-hour ambulatory BP monitoring. Across seven days of real-world monitoring, the authors report 65.8% sensitivity and 90.0% specificity; initial office BP screening had the same specificity and 55.3% sensitivity. Sensitivity is the share of people with hypertension the screening identifies; specificity is the share without it that it correctly excludes. The authors argue that a watch could prompt at-risk people to seek clinical assessment and confirmatory diagnosis using established cuff-based methods. These results describe detection in this study, not a confirmed diagnosis or demonstrated improvement in health outcomes. The abstract does not establish how performance would hold in other populations or how often watch prompts would lead to follow-up.
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
· Article
Source details
Original source date: Nov 5, 2026 (2026-11-05T00:00:00.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-11-13T21:35:23.399Z. First observed: 2026-11-05T17:09:47.518Z. Retrieved: 2026-11-13T21:34:23.768Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The abstract describes 196 participants free from known cardiovascular disease and antihypertensive medication; it does not establish performance in other populations. The reported monitoring covered seven days. The abstract does not report whether smartwatch prompts led to clinical follow-up or improved health outcomes. The supplied source is the article abstract and related page text, not the full study methods and results.
Open question
How well would AI-PPG-ACC-HTN perform in populations beyond the study’s participants?
Source-stated
The authors report using AI on smartwatch photoplethysmography and accelerometry, without cuff calibration, to detect hypertension.
applying artificial intelligence (AI) to multimodal waveforms (photoplethysmography and accelerometry) recorded using a widely available consumer smartwatch (AI-PPG-ACC-HTN), without any cuff calibration
Source-stated
The prospective study included 196 participants free from known cardiovascular disease and antihypertensive medication, and compared results with 24-hour ambulatory BP monitoring.
a prospective, multicenter study of 196 diverse participants free from known cardiovascular disease and antihypertensive medication against gold-standard 24-hour ambulatory BP monitoring
Source-stated
Over seven days, the authors report 65.8% sensitivity and 90.0% specificity for smartwatch detection; initial office BP screening had 55.3% sensitivity and 90.0% specificity.
Over seven days of real-world monitoring, AI-PPG-ACC-HTN detected hypertension with a sensitivity of 65.8% (95% CI, 54.0%-76.3%), specificity of 90.0% (95% CI, 83.2%-94.7%)
Source-stated
The authors propose smartwatch screening as a way to prompt at-risk individuals toward clinical assessment and confirmatory diagnosis using cuff-based methods.
prompting at-risk individuals toward clinical assessment and confirmatory diagnosis using established cuff-based methodologies
Source and contributor context
Di Achille, Paolo, Cai, Lawrence, Wu, Jiang, Gao, Mingwu, Daryani, Bhavna, Wang, Jonathan, Khanna, Utkarsh, Ko, Ho, Pathak, Anupam, Malhotra, Mark, Patel, Shwetak, Shreibati, Jacqueline B., Juraschek, Stephen P., Rudrapatna, Pramod, Thompson, Matthew, Poh, Ming-Zher · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Nature
Nature publishes research, news, and commentary across the sciences.
Editorial provenance
Version 1. Publication ID: 87912be5-3e6f-45e1-888b-dbe599a7c922.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-13T21:35:23.399Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 5077a8bb9c2c39e5c461ea8086818fd53055aa3d01c482fdd1f1086241a03753. Source text hash: f0bd804b1a4c2955ee970a62678e2df9a891db117610ff835ab646f6016de5e2.
Revision and source-check history
Version 1 · 2026-11-13T21:35:23.399Z
Authors of a Nature paper report smartwatch screening for signs of hypertension
The authors of a paper report that artificial intelligence (AI) applied to photoplethysmography and accelerometry recorded by a consumer smartwatch detected hypertension without cuff calibration. They tested AI-PPG-ACC-HTN in a prospective study of 196 participants without known cardiovascular disease or antihypertensive medication, comparing it with 24-hour ambulatory BP monitoring. Across seven days of real-world monitoring, the authors report 65.8% sensitivity and 90.0% specificity; initial office BP screening had the same specificity and 55.3% sensitivity. Sensitivity is the share of people with hypertension the screening identifies; specificity is the share without it that it correctly excludes. The authors argue that a watch could prompt at-risk people to seek clinical assessment and confirmatory diagnosis using established cuff-based methods. These results describe detection in this study, not a confirmed diagnosis or demonstrated improvement in health outcomes. The abstract does not establish how performance would hold in other populations or how often watch prompts would lead to follow-up.
Original source
· Article
Source details
Original source date: Nov 5, 2026 (2026-11-05T00:00:00.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-11-13T21:35:23.399Z. First observed: 2026-11-05T17:09:47.518Z. Retrieved: 2026-11-13T21:34:23.768Z.
Limits
The abstract describes 196 participants free from known cardiovascular disease and antihypertensive medication; it does not establish performance in other populations. The reported monitoring covered seven days. The abstract does not report whether smartwatch prompts led to clinical follow-up or improved health outcomes. The supplied source is the article abstract and related page text, not the full study methods and results.
Open question
How well would AI-PPG-ACC-HTN perform in populations beyond the study’s participants?
Source-stated
The authors report using AI on smartwatch photoplethysmography and accelerometry, without cuff calibration, to detect hypertension.
applying artificial intelligence (AI) to multimodal waveforms (photoplethysmography and accelerometry) recorded using a widely available consumer smartwatch (AI-PPG-ACC-HTN), without any cuff calibration
Source-stated
The prospective study included 196 participants free from known cardiovascular disease and antihypertensive medication, and compared results with 24-hour ambulatory BP monitoring.
a prospective, multicenter study of 196 diverse participants free from known cardiovascular disease and antihypertensive medication against gold-standard 24-hour ambulatory BP monitoring
Source-stated
Over seven days, the authors report 65.8% sensitivity and 90.0% specificity for smartwatch detection; initial office BP screening had 55.3% sensitivity and 90.0% specificity.
Over seven days of real-world monitoring, AI-PPG-ACC-HTN detected hypertension with a sensitivity of 65.8% (95% CI, 54.0%-76.3%), specificity of 90.0% (95% CI, 83.2%-94.7%)
Source-stated
The authors propose smartwatch screening as a way to prompt at-risk individuals toward clinical assessment and confirmatory diagnosis using cuff-based methods.
prompting at-risk individuals toward clinical assessment and confirmatory diagnosis using established cuff-based methodologies
Source and contributor context
Di Achille, Paolo, Cai, Lawrence, Wu, Jiang, Gao, Mingwu, Daryani, Bhavna, Wang, Jonathan, Khanna, Utkarsh, Ko, Ho, Pathak, Anupam, Malhotra, Mark, Patel, Shwetak, Shreibati, Jacqueline B., Juraschek, Stephen P., Rudrapatna, Pramod, Thompson, Matthew, Poh, Ming-Zher · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Nature
Nature publishes research, news, and commentary across the sciences.
Editorial provenance
Version 1. Publication ID: 87912be5-3e6f-45e1-888b-dbe599a7c922.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-13T21:35:23.399Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 5077a8bb9c2c39e5c461ea8086818fd53055aa3d01c482fdd1f1086241a03753. Source text hash: f0bd804b1a4c2955ee970a62678e2df9a891db117610ff835ab646f6016de5e2.
passed · 2026-11-13T21:35:23.399Z · anthropic/claude-opus-5.5
Publication: 87912be5-3e6f-45e1-888b-dbe599a7c922
attribution — pass: The reading consistently attributes the method, results and screening proposal to the authors of the Nature paper. The page lists authors from Google Research, the University of Washington and Beth Israel Deaconess. Omitting the Google affiliation is not a misattribution.
Di Achille, P., Cai, L., Wu, J. et al. Opportunistically Detecting Signs of Hypertension on a Consumer Smartwatch.
summary — pass: The summary leads with the development: AI on smartwatch sensor signals, without cuff calibration. It accurately conveys the cohort, the reference standard and the figures, and explains sensitivity and specificity in plain language. Photoplethysmography and accelerometry are kept as exact terms suitable for tooltips. Minor fidelity note: the 95% CIs for watch sensitivity (54.0-76.3%) and office sensitivity (43.4-66.7%) overlap. The draft does not claim superiority, so this is acceptable. The title and question follow from the evidence.
Over seven days of real-world monitoring, AI-PPG-ACC-HTN detected hypertension with a sensitivity of 65.8% (95% CI, 54.0%-76.3%)
date — pass: originalPublished is 2026-11-05, which matches the article's stated publication date. The received date of 1 December 2025 is the journal submission, not a public release. The evidence shows no earlier publication.
Received: 01 December 2025 Accepted: 20 October 2026 Published: 05 November 2026
support — pass: All four claims are supported by verbatim excerpts. The screening use is presented as the authors' proposal. The reading explicitly states that the results are not a confirmed diagnosis or a demonstrated outcome benefit. The stated gaps (other populations, follow-up rates, abstract-only access) are accurate and not overclaimed.
prompting at-risk individuals toward clinical assessment and confirmatory diagnosis using established cuff-based methodologies
media — pass: The selected media is the verified direct link to the original Nature article page, which is the most direct way to access the source.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
Nature publishes research, news, and commentary across the sciences.
Explore NatureNature publishes research, news, and commentary across the sciences.
Explore NatureNature publishes research, news, and commentary across the sciences.
Explore NatureNature publishes research, news, and commentary across the sciences.
Explore Nature