Why this cardiologist is cautious about the Apple Watch's blood pressure feature

[lostincalifornia]

"As a physician in the age of wearables, glancing at a patient’s wrist has become a natural, unspoken part of the physical exam. In most cases, finding an Apple Watch doesn’t mean much. But on occasion, it can offer a glimpse into a patient’s life — a heart rate trend, a sleep pattern, or a measure of activity.

..........................

Apple’s validation paper for the feature, alongside its FDA clearance letter, paints a more sobering story than its announcement did. As the company has acknowledged, its watches won’t offer users a direct, numerical blood pressure. Instead, it gathers data inputs from optical light sensors on the back of the watch trained on blood vessels, then uses a proprietary algorithm to infer blood pressure trends over time. This approach is notably different from how blood pressures are normally measured, with a cuff on a person’s arm.

........

The algorithm itself went through a rigorous process of design and testing, involving more than 100,000 patient cases for training it and over 2,000 for validating it against traditional blood pressure measurements. Trained on this large data set, the algorithm is reasonably designed to be specific for cases of high blood pressure, purposefully working to limit the number of false positive cases. It performed well across both sex and skin color — which has posed a problem for other light-based devices generating biomarkers, like pulse oximeters measuring blood oxygen levels.

But there’s a tradeoff: Across all tested groups, the algorithm was not sensitive, meaning that many users with high blood pressure would be falsely reassured with no notifications. The algorithm seemed to also perform worse among younger participants and participants with lower BMIs — the seemingly healthy-appearing use cases that Apple notes could benefit from this feature the most."

https://www.statnews.com/2025/10/13/apple-watch-hypertension-notifications-blood-pressure-public-health/

[Jarqui]

1. I tried a smart watch - not Apple

They claimed 99% accuracy
I compared its medical test results with calibrated equipment in my doctors office.
I returned the watch scolding them. NONE of their test results were remotely close.
My complaint was that they were wrongfully assuring some people they're ok when they're not. I told them that I hoped someone would sue the crap out of them.
They wouldn't give me the full refund they promised. I only got part of it ..