iRhythm Stock – A Play on Cardiac Monitoring and AI
If you took all the blood vessels in an adult human and laid them out in a single line it would stretch to almost 100,000 miles, enough to circle the entire globe four times. Imagine designing a pump to push liquid through that massive nest of blood vessels and you’ll end up with a newfound appreciation for how much work your heart does throughout your lifetime. Understanding the complexities of the human heart is the job of cardiologists who study for about twelve years before being qualified enough to practice. That’s why they make the big bucks. According to Glassdoor, the average salary for a Cardiologist is $358,646 in the United States.
Problems with the heart often manifest themselves through early signs such as slow, fast, or irregular heartbeats. In order to diagnose potential heart problems, cardiologists will often utilize an ElectroCardioGram (ECG) which is a device that measures the heart’s activities. Since observations are generally needed beyond just an hour or two spent in the doctor’s office, cardiologists will often deploy what’s called “mobile cardiac telemetry” which is simply the ability to monitor the heart over a sustained period of time outside of the doctor’s office. One such apparatus is the “Holter monitor” which is a battery-powered portable medical device that measures and records your heart’s activity while you wear it.
That device is good for about 24-48 hours’ worth of observations. It’s certainly not the sort of streamlined wearable you’d expect to see in today’s day and age. There has to be a better way to capture the data over longer periods of time and analyze it using the powers of artificial intelligence. That’s exactly what a company called iRhythm Technologies is doing.
Founded in 2006, iRhythm Technologies (IRTC) took in $122.8 million in funding from some big names like Stanford University, Kaiser Permanente, and St. Jude Medical before filing for an IPO in October of 2016 which raised an additional $107 million. Since then, shares of iRhythm have almost tripled giving the company a present-day market cap of around $1.8 billion. The primary value proposition on offer is a medical device that’s much more streamlined, portable, and long-lasting. iRhythm has built a small medical device called Zio that can capture up to two weeks of ECG data while allowing the patient to conduct their life in a perfectly normal fashion.
Once the data has been gathered, then the AI algorithms step in and offer up an interpretation that – get this – matches the physicians’ interpretation 99.9% of the time. Unlike other companies that tout the superiority of artificial intelligence over those subpar humans, iRhythm takes a cleverer approach by simply matching what the humans can do. That’s far more palatable for all stakeholders involved. It also means that the company can spend more time focusing on breadth instead of depth.
Bigger Data Means Better Algorithms
With data from over one million patients, iRhythm can now create algorithms with increasingly sophisticated capabilities. In order to build better algorithms with all that big data, iRhythm partnered with the Stanford Machine Learning Group and released a study this year titled, “Cardiologist-Level Arrhythmia Detection in Ambulatory Electrocardiograms with Deep Neural Networks” in which they showcased “a deep learning model capable of arrhythmia detection at a level comparable to a panel of expert cardiologists for a total of 12 output classes.” While accuracy is important, other companies have learned the importance of adding value to the entire process as opposed to developing the perfect algorithm.
Adding Value vs. Adding Accuracy
iRhythm isn’t the only company looking at using machine learning to analyze ECGs. In our article on 10 French Startups Using AI in Healthcare we noted a startup called Cardiologs which is using AI to detect arterial fibrillation. Even with a $350K-a-year incentive structure in place, the best humans can get is detecting the condition in 59% of the cases. The Cardiologs’ neural network, trained using more than 500,000 recordings, predicts the condition 91% of the time. However, achieving that level of accuracy isn’t enough to sell the solution. The entire solution needs to add value at an attractive price point – and that’s what Zio does.
Dr. Amine Korchi, a Venture Partner at Fusion, wrote a great post on Medium which dissected a presentation by Cardiologs’ CEO, Yann Fleureau, in which he discussed the challenges of implementing an AI-based solution in clinical practice. Both of these highly experienced medical professionals arrived at the same conclusion; “precision and sexiness of an algorithm are pointless if it does not translate to scalability, improved patient outcomes and integration into existing clinical workflows.” In other words, having an algorithm means very little if you can’t implement it and derive value from it. Since iRhythm owns the data and the devices that generate the data, they’re in an excellent position to implement their algorithms as a “feature” which compliments the existing process without challenging it too much.
The Importance of Intellectual Property
Speaking of being challenged, this brings us to the topic of intellectual property. Since the idea of applying AI to something like cardiology is so new, just how do you go about securing some intellectual property around your technology that provides some barrier to entry for the other startups out there that are trying – or will try in the future – to run with the same idea? The aforementioned Medium post by D. Korchi elaborates upon Cardiologs’ approach to securing some defensible intellectual property.
IP in the field of Algorithmic Clinical Decision Support remains a gray zone. Cardiologs patents of Deep-Learning models are still pending according to Yann Fleureau, and they used a singular approach to protect their intellectual assets. They did not submit an application to patent only the Neural Networks architecture, but they are looking to patenting the end-to-end Cardiologs approach from EKG to Diagnosis for the Doctors. This holistic approach of Algorithmic Decision Support experience includes also patents for Human Interactions between clinicians and algorithms, the way to display the EKG and the user workflow.
A cursory look at the intellectual property secured by iRhythm shows that they’ve secured intellectual property around this space from patents that were filed by Stanford University as far back as 2007. Intellectual property appears to be front and center for iRhythm, and here’s why:
This is a high-growth, high-margin business which would be very attractive to companies like Medtronic (MDT), whose cardiac and vascular segment did over $11 billion last year. It’s often a decision of “build vs. buy.” Just last year, Medtronic announced that it “will discontinue marketing for its SEEQ Mobile Cardiac Telemetry (MCT) System to concentrate support on its Reveal LINQ Insertable Cardiac Monitor (ICM) System.” The former was a device like iRhythm’s, while the latter is an implantable device which seems to address an entirely different use case as seen below.
While there appears to be room for more than one type of mobile cardiac telemetry device, for companies like General Electric Healthcare which sells Holter monitors, the future may not be so bright.
The Swedes recently published a study which looked at +280,000 heart attacks over 16 years and concluded that the chance of having a heart attack increases by nearly 40% on Christmas Eve. Holiday stress causes more heart attacks, and so does putting all your eggs in one basket and placing a single bet on iRhythm stock. While companies like iRhythm offer exciting stories with lots of potential growth, tenured investors will make sure their core holdings are well diversified before putting some side bets on exciting companies like this one which are successfully leveraging the power of artificial intelligence to add value.