For those of us who spend our professional lives in healthcare, the last few months have provided more harrowing drama than the last season of Game of Thrones. And it feels like despite the GOP’s initial failure to repeal and replace the Affordable Care Act, we haven’t yet seen the end of this story. Nonetheless, the industry appears to be breathing a sigh of relief as it shifts from “wait and see” mode back to regular go-to-market planning.
Though two of the industry’s leading investment reports differed wildly in their analyses, they agreed that Q1 was a big quarter, especially for companies in the analytics and population health space. A third pointed to consumerism-driven companies leading the pack, a trend that should be old news by now.
Here is a scoop of some recent industry developments, big and small, and what they mean for digital health companies.
Telehealth models at a crossroads
New research emerged suggesting that direct-to-consumer telehealth models can actually wind up driving up costs, in large part because they tap into new, unmet demand for care. While some telehealth CEO’s have pushed back against the study, it’s becoming clearer that the telehealth industry faces a crossroads. As a technology, video visits are quickly becoming a commodity, meaning that the vendor itself doesn’t matter. The licensing model that drove the industry for the last few years has begun to lose its sheen as health systems move past the shiny object phase of innovation into the ROI phase.
This study alone probably won’t be enough to dissuade health plans and employers from their telehealth partnerships, for two (very different) reasons. First, expanding access to care means more people will seek treatment right away, while cost savings of those investments take time to materialize – this holds true in telehealth as it does in insurance expansion. Second, the employer market segment continues to operate in an incredibly dysfunctional way. A quick look at how integrated networks and some provider systems are implementing telehealth technology reveals where the future of this space is headed.
Worth watching: CMS delays bundled payments
It was easy to miss amidst the hoopla of healthcare headlines, but CMS quietly announced the delay of two bundled payment programs in March. First, the final rule for a comprehensive joint replacement (CJR) for hip and knee replacements was pushed back yet again, to May 20th. Second, a new bundle for cardiac care was pushed back from July 1 to October 1. The announcement made it seem like this was part of a routine administrative delay, which may very well be true.
However, there may be cause for concern. Perhaps the biggest question facing digital health industry right now is whether value-based care has had enough time to take root in the industry and change how hospitals and health systems deliver care and invest in technology. The private sector has grabbed the baton in some instances, but other companies are smacking into recent hurdles, particularly around grant programs. Each specific bundled payment program offers startups a clear, tangible way to get a foot in the market. Health Affairs has more information about the impact and opportunity here.
You down with NLP? Yeah, you know me
Natural Language Processing (NLP) is one of the underpinnings of consumer-friendly user interfaces (UI), allowing people to communicate with a technology platform in their own speech. Provider search platform ZocDoc announced a new patient search feature that harnesses NLP to help match patients with more accurate results, improve recognition of slang, shorthand, and typos. While an emoji search tool may seem more gimmick than groundbreaking, these sorts of applications are important for a few reasons.
First, as digital natives, millenials, and other smartphone users begin driving healthcare volume in the years ahead, being able to meet them where they are will become a competitive advantage and eventually, a requirement of doing business.
Second, being able to understand people affords a tremendous opportunity to help steer them towards high-value care, say for instance, recommending a physiologist instead of an orthopedic consultation for lower back pain. Finally, beneath the froth, we’re actually well into the deep learning phase. Whether working on their own, like ZocDoc, or with partners, like WebMD (Amazon) or WellTok (IBM), established companies have been analyzing users’ language for a few years now. Given how valuable ease of access is to healthcare consumers, even an early stage startup can amass thousands of users within a few months. Expect NLP to become a bigger part of industry discussion as the year goes on.
An FDA-approved genetic testing arms race
The US Food and Drug Administration (FDA) just issued a one-of-a-kind press release in which they approved 10 genetic risk screenings from DTC genomics pioneer 23andMe. These tests don’t predict diseases; rather they arm consumers with percentage risk scores. The bigger news is that the FDA also provided them with an exemption to the normal pre-market requirements for subsequent tests.
This means that 23andMe gets a shortcut for the stringent approval process in releasing subsequent screening tests. Importantly, it’s suggested that additional companies will also be able to enjoy this exemption if their first batch of tests clear the FDA’s initial regulatory gauntlet. It’s early – the comment period in the Federal Register is open for a while longer – but we should expect this will trigger a veritable arms race among genomics vendors.
This is great news for the companies and investors that comprise the personalized medicine space – especially given how Health plans and industry are slowly conditioning consumers to consider genetic screening a part of their routine care. The DTC model of selling tests taps right into this zeitgeist to generate one stream of revenue, even while the data that consumers generate and sign over is worth millions to researchers and drug developers.
These developments have happened in the venture backed science laboratories, and as you might expect, there are speedbumps ahead. For one, these tests are not diagnoses; they provide people with probabilities, which consumers are notoriously bad at understanding. Test results are still not consistently accurate, nor workflow-friendly (“[t]he current state of art is a seven- or eight-page PDF”). And at least one company was caught found buying off physicians to pump up profits.
As one Mayo Clinic researcher put it, “we’re starting to see a lot of fumbles,” such as the case of a man who had a defibrillator implanted based on a misinterpreted genetic test results. Healthcare also faces a shortage of genetic counselors, whose role is to help consumers understand, interpret, and act on their test results. So, while the FDA has approved these tests for the mass market, it’s becoming clear the industry has some work to do to turn science and business into healthcare.
The diabetes space is heating up
We’re in a very eventful stretch for companies in the diabetes tech space. New startup Virta generated a huge wave of buzz when it stormed out of the gate in March claiming to “reverse type II diabetes.” Their emphasis on weight loss and lifestyle took a page out of Omada’s playbook, and sure enough, their bold marketing resulted in a puff piece that compared the two.
Virta’s clinical claims are certainly interesting – but they face serious questions around long-term effectiveness and the scalability of their services-driven program as an early stage startup. Beyond those nuts and bolts: If they succeed in reversing diabetes in year one, why should employers renew in year two? If they don’t succeed in reversing diabetes in year one, why should employers renew in year two? I agree with this take – Magical thinking has its limits..
At the other end of the maturity spectrum, WellDoc has shown their staying power (10+ years!): First by expanding their flagship app, Bluestar, into an FDA-cleared consumer product, inking a deal with a trucker’s union, and then partnering with Samsung to make it available to the consumer masses. Their automated, algorithm-driven engine may represent where the future of this space is headed. Let’s not forget that somewhere in Cambridge, Google/Sanofi’s joint venture, OnDuo, and Medtronic’s IBM Watson project, Sugar.IQ are building algorithms to this very end.
In the meter race, Livongo raised yet another round, raising questions about whether a company that builds its own hardware and hires its own coaches is actually a cost-effective solution. Newcomer OneDrop has quickly started earning stripes following their FDA clearance last fall.
Pharma’s build-not-buy app strategy is on display in diabetes, with insulin manufacturers Eli Lilly, Roche and Sanofi) building and getting FDA clearance for their own insulin dosing apps. One-dimensional apps are steadily becoming commoditized as a means to an end – that raises questions for how other companies in this segment decide to proceed, e.g. Livongo’s partner–not-build approach to insulin titration.
