Personalization via machine learning & outcome-based pricing

Sean Duffy · March 22, 2017 · Short URL:

The future of patient-focused preventive care

Tomorrow, I’ll be joining speakers, panelists, and moderators from across the digital health industry at Vator’s Splash Health, Wellness, and Wearables Conference in San Francisco. My panel will focus on promoting a healthy lifestyle through human-centered preventive care -- as a way to extend lives and reduce healthcare costs.

Earlier this month, I argued this type of prevention was the key to a fiscal sustainability of the nation’s healthcare system. And today, I’ll be focusing on two fundamental elements of Omada’s model I think are critical to support this exact type of successful prevention. One element operates at the patient level, the other at the operational level. Taken together, these dual pillars have rapidly accelerated our product innovation and can continue to drive the next generation of digital preventive medicine.

Personalization via Machine Learning

For decades, healthcare was the industry most resistant to adapting to the needs of consumers and personalizing service to deliver a better user experience. Last year, Omada’s VP of Data Science argued that digital health had learned a great deal from consumer brands like Netflix and Amazon. But in the year since we published that piece, the drive towards health care consumerization and personalization has hurtled forward even faster.

Today, patients expect an element of user experience design in everything from how their hospital care team coordinates to the patient portal they navigate to receive test results. But in order for digital health and preventive interventions to achieve the results necessary to bring down rates of chronic disease and actually bend the cost curve, the next step is to use data science and machine learning to actually personalize the healthcare treatment.

We are seeing the seeds of this action in the “precision medicine” movement, but these conversations are too often limited to targeted genomics. They can, should, and will need to have the same level of personalization when it comes to behavioral interventions.

Think about it. From Omada’s perspective, the behavioral changes and personal journeys we engage our participants in are deeply personal. We are working with them to change habits ingrained over years or decades -- and to maintain those changes over the long-term. A one-size-fits-all approach (like the one that can be accomplished in a standardized, in-person intervention) is ultimately a fool’s errand.

So Omada has invested heavily in both our product user experience team, as well as our robust data science operation. With more than 100,000 participants enrolled, 12 million meals tracked, 15 million weigh-ins, and more than a billion other data points from our users, we have now compiled the largest data set on behavior change in human history.

We’re using that data not simply to build versions of our program geared towards specific demographic groups like the underserved, but increasingly, to sharpen the EQ of our coaches and multiply their efforts through applications of data science and machine learning. There’s no question that the Omada program should look, feel, and be different for a 67-year old African-American woman from rural Georgia than it should be for a 33-year old white man from Minneapolis.

Now, we can predict with a high-level of accuracy which participants may struggle in our version of the Diabetes Prevention Program, due both to demographic factors, and their early activity in the program. So before Mr. Smith even knows that he’s in danger of performing poorly, Omada’s data team has flagged him for his health coach and delivered some critical intel on what type of touches are most likely to steer him towards success. We ask the health coaches to rate this intelligence according to their human instincts and our system learns and gets better over time. Our participants are starting to tell us that “our program feels magical.” They share that it seems like we just know when they need extra support.

This is the future of preventive care -- precision behavioral health, at scale. At Omada, we’ve built a system, and a team, to deliver it.

Outcomes-Based Pricing

Perhaps it was the advantage of blissful naivete that, when my co-founders and I started Omada in 2011, none of us had worked in the benefits space previously. While we had expertise in healthcare, product design, and software infrastructure, we were unbound by the conventions of the industry which we entered. 

So when we devised a revenue structure for our company, we did so from what we believed to be a pretty simple presence: we wanted to make money when, and only when, we delivered the outcomes we promised to our participants and our partners. 

Pretty simple, right? Not so in an industry dominated by a “per employee, per month” revenue structure that prioritized the signing and maintenance of contracts, but not much else. When we first brought this revenue model to our enterprise customers, they were philosophically on-board, but had no idea how to operationalize it.

So the Omada team quickly understood that technological innovation, in all its sleek, shiny, and sexy glory, only mattered if it could be coupled with the operational innovation necessary to make it work.
Today, Omada continues that commitment. We operate on an outcomes-based revenue model that consistently holds our feet to the first to deliver, and maintain, the results we promise from our product. So if we don’t deliver outcomes, we don’t keep the lights on. It’s a commitment we make to our customers when we first pitch them, and it’s written into every contract we sign.

In addition to this enabling a commercial alignment, it actually had an unintended side-effect: further personalization and product creativity. You see, when you know you only get paid when someone succeeds, it allows your product team to get exceptionally creative into what would make THAT specific person succeed. You start hearing questions around the room like:  How might we uniquely invest in this user persona to further their success in the program? It turned out that our pricing model helped accelerate product personalization and innovation as well. 

As healthcare has generally moved towards a more value-oriented model, it’s been gratifying to see others in the disease prevention, wellness, and benefits space follow our lead. We think it’s where healthcare needs to go -- and we think it’s a commitment every enterprise and health plan customer should demand from their vendors.

So, that’s what we’ll be talking about today. It’s an extraordinarily exciting time to be working in healthcare, and we couldn’t be more proud to be one of the leaders of the parade.

Editor's Note: Sean Duffy spoke at the recent Splash Health, Wellness & Wearables 2017 conference.  
Thanks to the Vator Splash Health sponsors: AdvsrAARPAvison Young, Bread and Butter WineSurf AirStratpoint and Scrubbed.


Image Description

Sean Duffy

Sean Duffy is the co-founder and CEO of Omada Health, a digital behavioral medicine company dedicated to inspiring and empowering people everywhere to live free of chronic conditions like heart disease and type 2 diabetes.

All author posts

Support VatorNews by Donating

Read more from our "The Future of Healthcare" series

More episodes