Commentary

AI Aids Medical Tricorder's Arrival -- 250 Years Early

I can’t decide which aspect of this story astounds me more — the fact that a consumer medical device with capabilities akin to Dr. McCoy’s tricorder has been successfully prototyped, or that it was done by two brothers from Pennsylvania, one an ER doctor and the other an engineer, who beat out a Taiwanese consumer electronics giant.

Since I’m told that some of you may be unfamiliar with “Star Trek” lore, the tricorder was (will be?) a two-piece mobile medical diagnostic tool used by Dr. Leonard “Bones” McCoy, ship’s surgeon of the United Star Ship (USS) Enterprise (NCC-1701), in “Star Trek: TOS” (The Original Series), which debuted on Sept. 8, 1966.

The two pieces of the tool included a roughly two-inch-long cylinder that looked a lot like a salt shaker (it was), and a rectangular box with a screen. Dr. McCoy would wave the salt shaker over a patient’s torso, and the box would diagnose any medical condition, right down to DNA mutations. Its use usually precipitated Dr. McCoy’s declaration of, “I’m a doctor, not a bricklayer,” or something similar.

advertisement

advertisement

In 2012, the Qualcomm Foundation decided to put up a $10 million purse to sponsor an Xprize contest to develop such a device. According to contest rules, the winning entry had to had to weigh no more than five pounds, diagnose at least 16 different conditions based on analysis of five vital signs, and be usable by anybody, without the help of a medical professional. Thus, it would “put the means for health awareness, metrics, and initial steps of care first in the hands of the person to whom the health belongs.”

Teams from all over the world competed. Earlier this month, the Harris brothers took home first prize of $2.5 million — plus millions in additional funding and partnerships with companies from Kimberly-Clark to Lowe’s — to transform their prototype tricorder into a consumer product.

Also kicking in $1.6 million to help productize the tricorder was The Roddenbery Foundation, as in the late great Gene “Great Bird of the Galaxy” Roddenberry, creator of “Star Trek.”

Side note: the tricorder Xprize contest launched in 2012, so it’s literally been a five-year mission.

The winners, Dr. Basil Harris, an ER physician at Lankenau Medical Center near Philadelphia, and his brother George, a network engineer, actually headed up a larger team -- of friends and family. “I’m a big fan of the whole ‘Star Trek’ universe, as is my team. Most of us missed the first run of “TOS,’ but we wore out a bunch of betamax tapes watching the episodes over and over again,” Dr. Harris told me in an email interview.

“It almost seems like after our sci-fi fi video marathons in the ’80s we all went out to gain the skills just so we could reconvene decades later and build tricorders!” Dr. Harris enthuses.

And that’s how things really happened. Roddenberry stalked real-life scientists to make certain “Star Trek”’s science was sound, and the kids who watched “Star Trek” grew up inspired to build what they saw. That’s why my past columns have exhorted you all to read sci-fi to understand the future rushing at you.

But I digress.

The Harrises’ tricorder (nicknamed DxtER) has two non-invasive sensors that must be worn, not waved (a chest sensor and a fingertip sensor), and an iPad that serves as their “rectangular box with a screen.” Together, they monitor five vitals: temperature, heart rate and heart rhythm (continuous EKG), respiratory rate, oxygen saturation, and blood pressure (without a cuff). In addition, the sensors get continuous readings of the user's glucose, hemoglobin, and white-blood-cell levels -- all without drawing any blood.

Using that information, DxtER’s AI-based app, operating entirely on the iPad, can diagnose whether you have one of 34 different conditions, including diabetes, atrial fibrillation, chronic obstructive pulmonary disease, urinary tract infection, sleep apnea, leukocytosis, pertussis, stroke, tuberculosis, or pneumonia, according to DxtER’s Web page.

Further, Dr. Harris says, “The user does not need to have any medical training to utilize the system. As they interface with the iPad, the app guides them in how to apply the sensors and gives feedback on how they're doing. It also guides them on how (the AI decides if and when) to use the other components in the kit. The top-level AI of our diagnostic engine makes the decisions and synthesizes all of the data to obtain a final diagnosis.”

A “next generation” version of the sensor, i.e., after the version that won the prize, is already in clinical trial at Harris’ home hospital.

(BTW, HTC is the consumer electronics company that sponsored the second-place team, which was awarded $1 million. Kudos to HTC for supporting and funding something that could really make a difference.)

But how did two guys from Pennsylvania without formal AI training pull this off? It would seem AI programming may not be all that hard, since you don’t actually write code. I’m told it’s more like raising a toddler -- training an unformed young mind.

Listen to Dr. Harris explain: “The AI follows a dynamic process that incorporates some subjective information that the user supplies with all of the objective measures. We developed it first with detailed algorithms based on my own clinical experience as an ER doc for over 12 years. Then we validated and refined those initial static algorithms through a chart review study... Then we added flavor by dynamically stressing the system with role-playing physicians.

“We created personas for the physician experts to follow, allowing them to role-play well-known scenarios, but restricted them from giving specific subjective clues to the system… We did this to build confidence in the system before letting it loose for our real clinical tests and the Xprize tests.”

The Harrises’ saga reminds me of a subgroup of the digital revolution that doesn’t get due consideration, something I think of as “tinkerers.” Over the past 10 or 15 years, the building blocks of extremely advanced technologies have become widely accessible at low cost (relatively speaking). Anyone can use them, though of course it helps if you have advanced degrees (and helps more if you have smart VCs).

Consequently, one of the less-well-understood reasons why technology advancement relentlessly accelerates is because there are so many more people who can do it nowadays. We’re not talking about a thousand flowers blooming, we’re talking millions. And when millions of people with millions of ideas start tinkering with advanced technology, it’s a cinch that amazing stuff will bubble up with greater frequency.

There’s no reason you or your team can’t achieve equivalent marketing accomplishments -- i.e., use advanced artificial intelligence to build great customer experiences. The Harrises show that passionate technology tinkerers can reach for, and attain, extraordinary achievements.

You may be a mere mortal marketer, not a medical doctor with an extra Ph.D. in engineering like Basil Harris. But after all, marketing isn’t brain surgery. Like Dr. Harris, all you really need is the desire to achieve, domain knowledge, and a little fearlessness regarding the technology.

All that will be a requisite in the AI-powered marketing era.

Next story loading loading..