Apple explains how its new Watch can measure how hard your heart is beating

When the latest Apple Watch models were unveiled last month, the announcement of longer battery life, a bigger display and satellite connectivity for Apple Watch Ultra 3 were eye-catching. But it was a new health feature that was the real draw: notifications for hypertension, that is, high blood pressure.

Apple has placed health and the heart at the forefront for its Watch for years, with ECG readings and blood oxygen measurements among recent highlights.

Sumbul Desai, Apple’s vice president of health, spoke to The Independent soon after the announcement to explain the new feature and the thinking behind it.

“We’ve been wanting to work on hypertension for many years, to be candid. Hypertension affects more than a billion people worldwide, but less than half those cases are diagnosed. We wanted to raise awareness and to give people more power to avoid some complications that can happen down the line,” says Dr Desai.

But how to measure it? Conventional methods, where a clinician straps a cuff to your arm may not be the best.

“Often, when I used to see people in the clinic, they would come in,” Dr Desai explains, “and they'd be really nervous, so their blood pressure would be elevated, or they just ran from parking their car and, again, it's elevated. But does that truly reflect what their blood pressure is as they live their everyday life?”

The new feature is not like heart rate, where you can initiate a reading instantly. Here, the feature works in the background by measuring blood pressure over a 30-day period. “We wanted to get a sense of your blood pressure as you're just living your life,” Desai says.

At the end of that period, if it’s spotted what it thinks are high blood pressure readings, the Watch will notify you and encourage you to log your blood pressure.

Other wearables can measure your blood pressure, such as the Hilo band and Samsung smartwatches. They usually require calibration with a traditional cuff, but that’s not necessary here — again, Apple wants a simple process.

“We think about health as being holistic at Apple, and one of the keys to managing hypertension is exercise. I always say, if I could prescribe anything, it would be movement, because that's key to so many conditions,” she adds.

While the heart rate monitoring on Apple Watch shows you beats per minute, there are no figures revealed for hypertension. Why is that?

“It was a few things, such as keeping it more simple and friendly. The way our algorithm works is that we did compare it to ground truth with a cuff, but we did it over a period of 30 days. Your blood pressure, one minute, can be higher, then you sit down, and it's lower,” Desai explains. “So, we decided to not fixate on a number: because of so many variations we were having a lot of outliers. And so it was better to do an aggregate over 30-day periods. The way the algorithm works is it looks at a signal that is indicative of hypertension, but isn't necessarily measuring the actual number but it correlates with the blood pressure number. We are not measuring systolic and diastolic directly in the traditional sense.

“What we're measuring is how the blood is flowing and what the response of the blood flow is, to the beats of the heart, and that correlates with blood pressure, which is why we didn't put an exact number in, for one reason. We wanted to start with how do we get the true sense of what your blood pressure is as you're living your life without a fixation on the number? And so that was the reason we decided to approach it more from this vantage point given the technology we have.”

Though no number is shown, the algorithm knows what the range is. It compares your individual readings over 30 days and then resets. “We had people take their blood pressure at various points during the day, and that's how we correlated the signal. We're looking at the trace pattern of the signal, that correlated with elevated blood pressure,” Dr Desai says.

She also explains that the sensitivity of the analysis is on the low side – Desai says it will detect four out of ten cases – compared to specificity which is very high, about 92 per cent.

“The reason we did that is, for those that get a notification, we wanted to feel confident that they will have a positive result. We didn't want to create a situation where, if the number was lower, say, we had false positives, and we wanted to make sure there was confidence in the algorithm when someone is using it. So, we made the trade-off of not being able to capture everyone, because if you look at the numbers of hypertension, it's still significantly a large number. But those that actually get a notification, we feel very confidently it will yield a stage one or stage two diagnosis. If you get notified, you're more than likely to have a condition.”

The 30-day system means it’ll assess your data for 30 days and if it sees nothing it will reset and start checking again over the next 30 days. “If you do receive a notification, it’s not that the process stops, we still keep checking in the background. I think it has a potential of shortening kind of the time frame that people get diagnosed with hypertension,” Dr Desai hopes.

The assessments take place multiple times a day, though not when you’re on a vigorous run, for instance, because your heart rate would naturally be elevated. There’s no set number of readings, but there’s a minimum across the 30 days for Apple to be confident in the data. Each reading takes just seconds.

The feature has a future, Desai thinks: “We do the appropriate validation testing to get the regulatory approval, because the regulators have to feel like we're not providing anybody with inaccurate information. But I think this area is ripe to understand more. This is very novel system in the way it does it, and we think we will learn that there may be other signals that this may be also indicative of, but we started with hypertension. And I think that's what's so remarkable.”