What's in a Name? AI & Digital Health

Words are important. Specificity is important. It can be an interesting (and entertaining) exercise to ask individuals to define what they mean when they use a word or term that the room assumes is generally accepted and understood. In the latest Tech Tonics podcast, Lisa Suennen laments the term "digital health" as ripe for obsolescence (actually, murder was her preferred word): 

"The truth is that it's not a real concept. There's healthcare and the application of technology to healthcare. But by making it a thing, "Digital Health", it's sort of illusory and silly. We don't call it digital banking, we don't call it digital manufacturing, why are we calling it digital health? We should be finding ways that apply technology to healthcare that makes sense and stop calling it a thing."

In some sense - this is correct. When a term is overused, it begins to lose its meaning - as a consultancy, we've met precision oncology companies, sensor-based companies, analytics based companies, and telehealth companies - all of whom self-identify as "digital health". One look at the 2017 Rock Health funding reports over the last year and the ambiguity of what defines a digital health company becomes manifest. Because we find comfort in familiar jargon and sound bites, someone says "We are a digital health company", and the room nods in apparent understanding. But what does that mean, exactly?

Last week's Rock Health Summit got to the heart of Lisa Suennen's thesis - clearly demonstrating "digital health" as the application of technology to healthcare. Whether that technology is a continuously worn MRI (OpenWater), a mental health chat bot (WoeBot), a solution to compare physician services (Amino), or a community based health application addressing social determinants of health (Cityblock) - each is using technology applied to a specific healthcare problem. And though each of these technologies differ tremendously, there is a common foundation that enables and potentially amplifies their potential: Data. A LOT of data. The only way to make sense of these terabytes upon terabytes of data (ie to translate it into a diagnostic or a therapeutic or some sort of decision support) is to apply some sort of analytics to the massive dataset. We used to call this predictive analytics. This notion seems to have become quickly antiquated, and evolved into artificial or augmented intelligence (AI) encompassing far more than just predictive analytics, often creating an aura of something too complex for the common man or woman to ever understand. Every industry seems to be powered by AI, or promises to be "revolutionized" by it. Including, and especially healthcare. With six terabytes of health data generated at each encounter, we certainly need something - because it's not just the point in time data that matters, but the change in data over time. The Framingham Heart Study proved it old-school style in cardiovascular health. Now, Verily's Baseline aims to do this more broadly ... powered by Google. Others, like Flatiron Health, are disease-focused - collecting diagnostic and treatment data over time together with clinical response and outcomes. Our most promising solutions anchor at the convergence of healthcare, technology, and data science. 

That potential and the excitement with which that potential is accompanied leads to the inevitable "AI-powered Digital Health" solutions. And what does that mean? What does it mean to investors, patients, payers, providers ... and regulators? There is real power there, and it is a signal of entrepreneurship, progress, and speed. There is a sentiment that if "AI" isn't in your pitch deck, you may as well not play. It's a little like discovering Sriracha for the first time - it's so good, you want to put it on everything. Too much of it, and you can't taste anything else. AI - like digital health - means different things to different people. It is the specificity of the technology, and how it impacts the end solution/product that becomes important.

Our healthcare problems in the US are large and real, and seem to be increasingly pressured with every news cycle. We absolutely need technology to solve these problems. We need a regulatory infrastructure that reflects the changing face of healthcare innovation and encourages entrepreneurs to think outside the box, so long as a culture of quality and excellence continue beyond the borders of that box. But we also need to get specific about the problems we are trying to solve, for whom we are solving those problems, and the tools by which those solutions will be delivered. Let's start with clarity about what the solutions really are.


Click on the links for highlighted session notes from both the Digital Health Regulation and Turning Data into Action Sessions from the Rock Health Summit.