When a company sets out to bring a cuffless, continuous blood pressure monitor to market, the wristband is the easy part. The hard part is proving that a cloud algorithm can read blood pressure accurately from an optical signal, across every kind of person who might wear it. That was the throughline of “Under Pressure,” Castor’s 2024 conversation with Jay Shah, Chief Medical Officer at Aktiia, and Derk Arts, CEO of Castor.
It matters because the problem is enormous. The World Health Organization estimates that about 1.28 billion adults aged 30 to 79 have hypertension, and only about one in five have it under control.[1] A device that tracks blood pressure continuously, rather than once or twice a year in a clinic, is a genuinely different tool against that gap.
What the session covered
Jay’s most useful point for anyone building a device is that the regulator does not separate the hardware from the software. He explained that Aktiia’s band uses off-the-shelf components, optical sensors and an accelerometer, and that the defensible technology is a cloud algorithm his team refined over two decades at the Swiss research institute CSEM. In his telling, the whole system is assessed as a single submission, which raises the evidence bar in a way most teams underestimate:
“You have to prove accuracy across different skin tones, from Fitzpatrick scale one to six, follicle density, hair density, BMI, obesity, gender. All those things you have to prove accuracy in every single cohort of users. If you thought you’re gonna rush a PPG based device to the market, think again.”
Jay Shah, Aktiia
Continuous measurement also changes what blood pressure is. Jay said the device takes roughly 200 readings a week, which turns a single annual threshold into a pattern a clinician can act on. He used salt as the example of why that matters:
“Only fifty percent of people with hypertension are salt sensitive. So for fully half the people we’re making that recommendation to, it’s not gonna help. And we have no idea which fifty percent that person is.”
Jay Shah, Aktiia
He was candid about the discipline that takes. Jay described a deliberate choice to withhold a reading rather than publish a bad one, so the accelerometer suppresses measurement during wrist motion, when the optical signal is too noisy to trust. Castor’s part in the story was the data layer. Derk described Castor as an API-first platform that lets a device company bring sensor data together with clinical and patient-reported data in one place, and Jay pointed to the platform’s flexibility and reliability as what made it fit the work.
What’s in the full recording
The conversation gets more concrete than a summary can:
- Jay opens up the device and explains why “nothing in here is proprietary,” and where the real IP sits (16:29).
- The roadmap hint that the algorithm has already been tested on six other devices, including a discussion of running it on third-party wearables (19:05).
- A longitudinal cohort study out of Liechtenstein, run with a McMaster University investigator, that Jay said follows about 1,800 people for 20 to 30 years with biomarker, imaging, record, and genomic data, now adding continuous blood pressure (26:37).
- The audience question on data access, and whether a study sponsor gets the raw optical signal. The answer, “generally, no,” and the reason why, draw the IP boundary clearly (32:27).
Watch the full conversation
Jay and Derk get specific about the evidence a cuffless device has to generate and how the data pipeline supports it. Watch the session on demand. Since the 2024 recording, Aktiia has rebranded as Hilo, and in 2025 its G0 Blood Pressure Monitoring System, the Hilo Band, became the first cuffless blood pressure monitor cleared by the FDA for over-the-counter use in the United States, with US availability following in 2026 and the device already available across Europe, the UK, and Australia.[2] If you are generating the evidence base for a medical device, see how Castor supports medical device clinical trials.
Frequently asked questions
Why is a cuffless blood pressure device a software challenge, not a hardware one?
The sensors in a cuffless monitor are largely off-the-shelf. The defensible technology is the algorithm that converts an optical signal into an accurate blood pressure reading, and regulators assess the hardware and that software together as a single submission. So the evidence burden sits on the algorithm’s accuracy, not the wristband.
What does it take to validate a cuffless blood pressure device?
According to Jay Shah of Aktiia, a device has to prove accuracy across the full clinical range of blood pressure and across every user subgroup, including Fitzpatrick skin tones one to six, hair and follicle density, BMI, and gender. That subgroup-by-subgroup bar is what makes bringing a PPG-based device to market slow and demanding.
What was Castor’s role for Aktiia?
Castor was the data platform. Derk Arts described it as an API-first system that lets a device company bring sensor data together with clinical and patient-reported data in one place, and Jay Shah pointed to its flexibility and reliability as the reason it fit Aktiia’s work.
References
- World Health Organization, Hypertension fact sheet: an estimated 1.28 billion adults aged 30-79 have hypertension, and around one in five (21%) have it under control. https://www.who.int/news-room/fact-sheets/detail/hypertension
- MassDevice: FDA clears Aktiia’s over-the-counter G0 cuffless blood pressure monitor (the Hilo Band), a 510(k) clearance and the first cuffless BP monitor indicated for OTC use in the US (2025). https://www.massdevice.com/fda-clears-aktiia-over-the-counter-g0-cuffless-blood-pressure-monitor/
