Detecting signs of hypertension using wearable sensors
Web page · https://research.google/blog/detecting-signs-of-hypertension-using-wearable-sensors
Health & Bioscience
First seen by Oriel on .
AI summary
Google Research presents AI-PPG-ACC-HTN, a cuff-calibration-free smartwatch system that aggregates sensor data over multiple days to screen for persistent hypertension, with performance comparable to cuff-based workflows in a clinical study.
- The model combines photoplethysmography (PPG) and accelerometry (ACC), using self-supervised learning on 492,347 hours of data from 2,109 participants before person-level aggregation and logistic regression.
- In a prospective study of 196 participants across five US sites, initial screening achieved 65.8% sensitivity, 90.0% specificity, and 80.6% positive predictive value against 24-hour ambulatory blood pressure monitoring.
- Among 48 participants with elevated office readings, confirmatory screening achieved 78.4% sensitivity, 90.9% specificity, and 96.7% positive predictive value.
- The system detects persistent risk rather than estimating exact blood pressure values, aggregating 15-second sensor segments across days to reduce transient noise.
- The approach is exploratory and intended to prompt clinical assessment, not replace arm cuffs or provide a standalone diagnosis; it is not an existing or announced commercial feature.
#wearables #hypertension #machine learning #blood pressure
Generated by gpt-5.6-sol. Check the original for details.
Aliases
| Source | External ID | URL |
|---|---|---|
| Google Research Blog | https://research.google/blog/detecting-signs-of-hypertension-using-wearable-sensors/ | https://research.google/blog/detecting-signs-of-hypertension-using-wearable-sensors/ Read discussion |