The inaugural HLTH 2018 conference took place at the Aria in Las Vegas last week. Their tagline:
WE ARE THE HOTTEST, NEWEST, LARGEST AND MOST IMPORTANT HEALTHCARE EVENT CREATING A MUCH NEEDED DIALOGUE FOCUSED ON DISRUPTIVE INNOVATION.
Having been perched on the precipice of self-described disruptive innovation for over a decade, this was clearly a not-to-be-missed event.
HLTH promised to be JP Morgan-like in its cache, brand, and networking potential - finally offering "digital health" the massive cross-disciplinary conference platform it deserved.
And it was an ambitious agenda indeed - five parallel tracks each day ranging from value-based care to genomics to strategic investing by health systems/payers to aging in place to consumer technologies to AI, healthcare policy, and more. Interspersed between track sessions were dedicated general sessions highlighting a who's who of speakers from CMS, Blue Cross/Blue Shield, Nokia, Lyft, Intermountain, Walmart, CVS, Geisinger Health, 23andMe, Arivale, Adobe, Garmin, UnitedHealth, Fitbit, and even the American Medical Association. There were 375 speakers in total representing almost as many companies across the healthcare and technology ecosystem. Absent: Apple and Amazon (though AWS was listed as a sponsor). It was hard to find an industry that wasn't represented. Translation: healthcare clearly impacts and is impacted by ... everything.
So did the conference live up to its own hype? With a standard registration fee of $3,100 (there was a discounted fee for start-ups), not to mention hotel and airfare, expectations were high for other reasons. Even with the $150 discount from A Healthy Dose, I approached this as an expensive experiment. Over the last 10 years, when wireless technology made its first tentative (and disruptive) inroads in healthcare, there have been a run of healthcare-meets-tech conferences: Health 2.0 (acquired by HIMSS), Rock Health Summit, Start Up Health Festival to name a small few. While certain sessions were always relevant, these conferences were largely the same conversation had in different ways:
- Healthcare is expensive & inefficient.
- Data analytics can solve the problem.
- We need interoperability and data exchange.
- Healthcare needs to be personalized.
- We can scale healthcare through automation.
- Except maybe we still need a human touch. Sometimes.
In that sense, HLTH was not so different. Same soundbites: the United States spends close to 20% of GDP on healthcare, yet we have some of the poorest outcomes of any industrialized nation. Our elderly population is growing while the number of physicians is shrinking. The physicians we have are burned out. We need technology.
Except. There was more.
There was an undercurrent of things that you don't traditionally pair with a VC, investor driven world. Social determinants of health (SDoH) for one: the notion that healthcare, more specifically poor health, is regressive: illness disproportionately impacts people without access to quality nutrition, housing, and resources. Anyone who has seen a US map overlay of mean household income with rates of obesity, hypertension, and diabetes understands this. Anyone who has worked with kids and families in urban environments has seen first-hand what lead exposure in low income housing does to early child development and how that impacts outcomes in young adulthood. What was unique here was not that the conversation was being had - frankly that conversation has been well-worn in public health and nursing environments - but it was being had in a conference where not thirty minutes on either end, investors were talking about Unicorns, returns, and multiples. Importantly, these conversations were integrated, not just being had in parallel. This is where disruptive innovation happens. In this space. Where finally healthcare is being understood as the holistic beast that it is - that "tech" in and of itself cannot solve it wholly, but it can have meaningful impact. That in the most expensive areas of healthcare, the unmet need is not an app, but access to food. And that ultimately, your zip code can be more important to your health than your genetic code.
So where is the opportunity for "tech" or "digital health" or whatever name you want to give the magic fairy dust that, when sprinkled on healthcare, force amplifies scarce human resources and/or strips billions of cost out of the titanic system? It's there. In the "Investing in Health Innovations" track, the panelists (which included Nina Achadjian) were asked by moderator Bill Geary where the big opportunities were. In response, Krishna Yeshwant noted the following:
- "Machine Learning" in healthcare is an overused term
- Google has - over the last 10 years - immersed itself in a gradual process of understanding healthcare by spending time with patients, physicians, and providers
- When applied to the EHR, analytics/AI/ML may have a significant impact on the "unsexy stuff" like operations
- There are billions of inefficiencies in healthcare that can unlock tens of billions of dollars
(Note the latest $30M round of funding for Qventus announced yesterday - AI to improve hospital efficiencies.) One might see that with those billions saved, new investments can be made by hospitals/systems to expand community- and home-based programs, triggering strategic investments in companies like City Block Health, ChenMed, and Iora Health. And dedicated VC funding for those companies is coming - former CMS Acting Administrator & Goldman Sachs IB, Andy Slavitt announced the launch of Town Hall Ventures, a fund dedicated to innovations that will support the underserved. As Dan Buettner, Founder of Blue Zones, articulated: "Healthcare is complex because people are complex. Environments are complex." There is no quick fix, and no shortage of opportunities.
Which leads me to another area of HLTH2018 dedicated pundit space - AI & machine learning. It kicked off somewhat grandly during Sunday's general session with Vinod Khosla predicting that 80% of what doctors do could be replaced by technology. The tension here is not that providers/physicians don't like technology, but that there needs to be some basis for understanding how and that it works. And at its very core, AI/ML depend on ... data. Quality data. Data that is representative of populations - of people, of diseases, of socioeconomic conditions, of environments. Sourcing that data in the first place is complicated (hint: it's not Epic or Cerner) and cleaning it is more complicated still. Healthcare doesn't just happen in the hospital. It happens at home, school, in the community, and in physician offices. And it happens over time. Depending on the answer one is trying to find, some or all of these data might be needed. Over a short, medium, or long period of time. Phenotypic expression of the underlying genome is one area where eventually all of these data will be needed. If a company is trying to understand drug discovery, then a smaller subset suffices. This is discipline. Pharma gets this - they are the consumers and end-users of their own algorithms, and they depend on this for their revenue. And so they have been mindful. Amy Abernathy (CMO, Flatiron Health) and Chris Gibson (Co-Founder/CEO, Recursion Pharmaceuticals) discussed the often under-highlighted realities of AI:
- In order for AI to work, there needs to be manual curation of data (from the EHR)
- Even in structured datasets, there is a lot of clinical interpretation
- Biased data begets a biased answer
- There is no magic button
These themes recurred across several of the AI and genomics sessions. And certainly David Feinberg, MD is betting on Geisinger Health's ability to integrate their longitundal phenotype datasets with genomic testing - he announced that Geisinger will provide whole exome testing for ALL patients because he sees it as (1) cost effective, and (2) decreasing healthcare cost overall. As a forward-thinking integrated healthcare system that is both payer and provider, is providing access to affordable, healthy nutrition through its Fresh Food Farmacy and affordable housing to opioid addicted pregnant women, Geisinger will also have the opportunity to understand the impact of each of these separately and as their aggregate sum on the health outcomes and cost of their population. Ultimately, this is what it is all about.
In the spirit of full disclosure, I had to leave HLTH2018 one day early; I'm a Big Sister and promised my Little that I'd be back for his after school program on Wednesday. I also had some real work to do. My interest in attending HLTH2018 was largely curiosity-driven: as a clinician for ten years (in both urban and wealthy environments) followed by ten years in strategy/business development across medical device, wireless health, data analytics, and genomics industries, HLTH was an unusual conference of convergence of ALL of these things. It is atypical to find one conference that promises the breadth and attempted depth as the HLTH agenda. I assumed the challenge would be to find those sessions that promised something innovative & provocative, and wasn't simply reinforcing an existing belief system. As Alex Drane highlighted in her general session talk where she described becoming a Walmart cashier - you need to immerse yourself in the world of others to engender understanding.
HLTH certainly presented that opportunity - it was IMPOSSIBLE to attend every session or even a few sessions from every track. In many ways, it reinforced my understanding of the challenges and opportunities and healthcare. Most importantly, it was invigorating to see the passion and energy from a diverse group of entrepreneurs, providers, investors, and policy-makers who were listening to one another outside of their respective echo chambers, who could be honest about the hard problems to be solved (and how long we've been trying to solve them), and who remain confident that the solutions could be brought to bear.