Bob Hauser is fun at dinner parties. He entertains, he tells old stories, and occasionally the retired cardiologist will diagnose a friend's atrial fibrillation right there at the dinner table. It's happened more than once.

Hauser, who is retired from regular practice at the Minneapolis Heart Institute, has been using a wireless, consumer-oriented device called the Kardia Mobile to test his friends when they complain about unsettling sensations in their chest like a racing heart beat or palpitations. He has used the smartphone-enabled device to detect quivering or potentially dangerous fibrillation in several acquaintances who went on to get care from their regular doctors.

"I've had a number of people, whether here or down in Florida, say, 'Hey Bob, I know you're retired but I've got these palpitations,' " said Hauser, who says he has no financial stake in the device. "And I have now picked up I think seven in total people who came to me with this concern, and I subsequently took their EKG with this Kardia device and found that they were in atrial fibrillation."

Now doctors at Mayo Clinic in Rochester are teaming up with Silicon Valley-based AliveCor on two medical research projects to see whether the Kardia Mobile cardiac monitor and the artificial intelligence system it links to in California can reveal evidence of low potassium levels and irregularly long ­heartbeats.

Both conditions can contribute to major cardiac problems. The Mayo research involving the detection of low potassium levels has been ongoing since last year under the direction of Dr. Paul Friedman. Low potassium is tracked because it can cause bad changes in a person's heart beat and can indicate kidney problems.

In July, Mayo and AliveCor announced their second collaboration, in which the device will be used in a clinical trial to see if the Kardia system can detect irregularly long heartbeats as effectively as a standard in-office test. That health condition is called "Long QT syndrome," which is treatable if detected early, and potentially fatal if not.

"If we can create a tool that accurately reflects the patient's Q-T status, then I think this is going to be a game-changer," said Dr. Mike Ackerman, who is directing the trial. "My ultimate goal is that we would have a time when every person, everywhere, knows their Q-T health, and that we would have begun universal screening for long Q-T syndrome."

The Kardia Mobile is a slim device about the size of a small chocolate bar, and it can be stuck to the back of a smartphone or used alongside a phone or tablet. The user places four fingers on the device sensors, and the device reads a person's electrocardiogram (ECG) data and sends it to the phone, which displays it on an app that the user has downloaded. The data can be sent to a doctor or stored for future reference. In the U.S., a board-certified cardiologist will automatically review the first recording within 24 hours, and after that, the user can record unlimited readings.

Kardia's original test, already cleared by the FDA, either confirms a normal heart rhythm or detects signs of atrial fibrillation. AF is a relatively common condition that can lead to blood clots if not treated with blood thinners.

"Sadly, the first indication that people have heart issues is usually a very bad situation — they have a stroke, or worse. And then they get the ECG done," said AliveCor CEO Vic Gundotra. "Now technology allows you to put that in your pocket and do an ECG. It's an extraordinary thing that we live in a world where this kind of technology is available inexpensively, and you can carry it in your purse or wallet."

The Kardia Mobile is intended to be used by consumers and is not covered by insurance, though it may be eligible for FSA and HSA programs. The device itself costs $100 with a 30-day return policy and a one-year limited warranty, with an optional subscription for premium services available for another $99 a year.

Gundrota said the Kardia Mobile is priced as low as possible while still guaranteeing quality manufacturing. That's a different way to think about marketing than traditional med-tech, where manufacturers often base list prices for new devices and therapies on the current cost that the market will bear for the existing treatments, regardless of manufacturing costs.

Gundrota used lofty language to explain the rationale for the pricing model — "this is about the radical democratization of information," he said — but he also acknowledged that AliveCor has a selfish motive in having the device used as widely as possible. That reason is artificial intelligence.

When the Kardia device takes a reading, the data are beamed from the user's smartphone to the AI systems at AliveCor's headquarters, which have access to 18 million past ECG readings and deidentified medical records. The "machine learning" system scans those older data and tries to discern subtle clues in the wiggly ECG readings that correlate to disease states like atrial fibrillation.

The more data available, the better the AI predictive systems will become. Today the AI system examines 300,000 different aspects of each new ECG reading, giving it far deeper insight than the dozen or so features a cardiologist might consider on the same chart, Gundrota said.

He dismissed the idea that such devices may one day "replace" cardiologists, saying that doctors will always use them as tools to supplement their existing skills.

A Mayo spokeswoman noted that the health system has a financial stake in AliveCor, and is contributing the intellectual property in the form of expertise from Drs. Ackerman and Friedman for which it may receive future payments. Mayo said it would use any revenue from the deals to support its not-for-profit missions in patient care, education and research.

Hauser, the Minneapolis cardiologist, said he's confident using the device because of the data that the FDA reviewed before granting clearance to the device. The data showed that doing twice-a-week tests with the device was associated with a fourfold increase in detection of atrial fibrillation compared to a control group, and resulted in more patients who needed anti-coagulation drugs getting them.

But if AliveCor was his company to run, Hauser said he would want to run a longer-term clinical trial to see whether the increased rates of detection actually result in longer life spans for the patients. So far AliveCor has no plans for such a trial.

"As our platform scales, we do expect that AliveCor data, and the insights it yields, will contribute to our collective goal of improving heart health around the world," Gundrota said in an e-mail response.

Joe Carlson • 612-673-4779