Breakfast Bytes Blogs

One of the most interesting things I saw at MWC was a meeting I had with Mark-Eric Jones. Full disclosure: I've known Mark-Eric for over twenty years and he is a friend. He is CEO of a company called Léman Micro Devices. Lac Léman is a lake on the French-Swiss border. The company is based in Lausanne (also the home of DATE this year). The name is not significant, they named the company when they founded it and were not yet sure exactly what they would do.

After 6 years, they came out of stealth mode a couple of weeks ago. What they do is make a little sensor to go on smartphones that makes medically accurate measurements of vital signs. See the picture above, complete with jamón in the background for Barcelona authenticity, or the sensor attached to Mark-Eric's business cards below it. The five vital signs that are measured are:

• Temperature
• Heart rate
• Respiration rate
• Blood oxygen
• Blood pressure

It is the last of these, blood pressure, that has everyone interested since it makes the measurement without a cuff. The reason that this is so significant is that the World Health Organization (WHO) estimates that over 9M people a year die from hypertension (high blood pressure), mainly because they don't know that they have it—and high blood pressure is the #1 indicator of early death. In the US, the Center for Disease Control (CDC) estimates that over 400K Americans die from it, half by heart-attacks, half by strokes. Medication to treat hypertension is cheap, since the most common drugs are out of patent. Potentially, this little device could save one life every three seconds.

All the measurements (except temperature) are made by putting the side of your finger on the sensor. Temperature is measured by pointing the sensor at what you want to measure the temperature of; for medical usage you swipe it across your forehead (without touching it). At lunch, we verified that the water was cold and the food was hot. Heart rate and respiration rate are measured using signal processing to detect the changes and pulse oximetry for blood oxygen. These are actually very old techniques brought into the modern semiconductor era. Infra-red bolometry for non-contact temperature was invented in 1878, photoplethysmography (PPG) for heart and respiration rates was invented in 1938, and Riva-Rocci arterial occlusion for blood pressure was invented in 1896. Pulse oximetry for blood oxygen is the baby, not even 50 years old, having been discovered in 1972.

The blood pressure is the most important of these from a business point of view. In fact, I have heard "cuffless blood pressure measurement" being described as the killer app for health devices. The way it works is that you put the side of your finger on the sensor and you play a little video game. You have to direct balls into an opening by moving the opening: increase the pressure of your finger for the left, decrease for the right. This cunningly makes you press and release the side of your finger which is somewhat analogous to the nurse pumping up and releasing the cuff in a doctor's office (with a wonderfully-named sphygmomanometer). After 30 seconds, you get a pretty good reading; if you play for a minute it is medically accurate.

It turns out that one challenge in blood pressure measurement is hydrostatic pressure. That's why nurses like to put the cuff on your upper arm so that it is essentially level with your heart, and why they don't trust wrist-based measurements, since your wrist can easily be above or below your heart, and adds (or subtracts) a constant amount of pressure to all readings. The module and software solve this problem by using the front-facing camera on the phone to detect your eyes, and thus triangulates where the phone is, in relation to your heart, and make appropriate static pressure adjustments.

Considering the size of the module, there is a lot in there. It combines two LEDs (one red and one infrared), a photodiode, a pressure and temperature sensor, a thermopile (with cold junction temperature sensor) and a mixed signal ASIC that performs data acquisition. Each module has a unique ID number stored into a non-volatile OTP memory that can be read by the application software. The ASIC is a mixed-signal design. It was done with Cadence tools on a Hosted Design Environment supported by the team in Munich. The big win for them was simplicity: no IT people, no servers, no complicated license mixing.

The ASIC is 180nm. The module interfaces to the phone through an I2C bus and all the processing is done on the smartphone processor. There is no processor nor any memory on the ASIC. Why try and do complex calculations on the module when you've got a multi-core ARM a couple of inches away? So there are tens of megaabytes of software that go along with the device.

One big issue with the sale of medical devices is the medical regulators, such as the FDA. The module itself is regulated as a medical device, so, for example, it is actually made of plastic that is non-toxic enough the sensor could be implanted in your body (although it wouldn't work there). However, the smartphone to which it is attached is treated as an "accessory" (that is a technical term in this context), meaning that it doesn't require any medical certification. The first phone manufacturer that Mark-Eric talked to thought it was too good to be true, and he had to show them the letters from the FDA (and its European and Chinese equivalents) confirming this. This may sound trivial but is a major "feature" of the module. If phone manufacturers had to certify all their phones as medical devices, nobody would go near the module. Getting through the certification process also serves as a major business barrier, not to mention that they already have several patents, including in China, which is the #1 smartphone market (in units) and so potentially the biggest market for this module. As Mark told me, "I've been spending a lot more time in China than the US, where there is only one smartphone manufacturer."

Nobody truly knows the size of the market, but potentially it is the number of smartphones sold, which today is about 1.4B/year. Mark-Eric has a 3 question elevator pitch too:

1. What is the most important thing to you? Answer most people give after a bit of thought, my health
2. How much do you spend on medical electronics? Err...nothing
3. How much do you spend on electronics in general? Big number, perhaps $1000/year

Or, as Tim Cook, Apple's CEO said last year: "health care may make the smartphone market look small."

Medical electronics is in the mainframe era, where you need expensive equipment only available in hospitals and, sometimes, doctors' offices. It needs to move into the smartphone era. Also, in an era where smartphones are pretty undifferentiated since most of us don't really care about an extra 100 MIPS on our GPU or a few more megapixels on our camera, this has the potential to make some phones a lot more attractive than others. They have not focused on cost-reducing the module since the biggest challenge is to get it adopted. But Mark-Eric knows that approximately 1 microsecond after adoption, people who were advising him to just get it working and not worry about the cost will be wanting it cheaper. In high volume after a cost-reduction redesign, the module could be as little as $1.

More details are on the Léman website.