Skip to main content

Omada For Employers

Using AI to Strengthen Human Relationships: Q&A with Terry Miller, VP of Machine Learning, AI and Data Science Strategy

AI’s adoption in healthcare continues to grow, with 79% of healthcare organizations reporting they’re currently using AI technology. While it’s heartening that the industry is adopting cutting edge technology, this doesn’t paint a complete picture of how and why they’re using it. Omada’s use of data science and related technology is an evolving story with valuable insights and lessons learned over the course of a decade. The latest chapter in that story is the hiring of Terry Miller, Omada’s new Vice President of Machine Learning, AI and Data Science Strategy. 

Prior to joining Omada, Miller spent time as an analytics service leader in the industrial automation space. He also holds a Master of Science (MS) degree focused in Predictive Analytics from Northwestern University. Under his leadership, Omada Health became one of the first virtual healthcare providers to pledge to the White House-Supported Healthcare AI Commitments Coalition

In his role, Miller hopes to help nurture Omada’s investment in AI with the aim of building out an infrastructure that can meaningfully and safely support the work of compassionate, human-led care teams. He recently took time out of his busy schedule to talk through his approach to core AI principles, technology watchpoints in digital health, and what drew him to Omada.

 

Omada Health: Why did you decide to join Omada? 

Terry Miller: I joined Omada because I felt the opportunity would be both personally and professionally fulfilling. I believe many cultural and policy issues stem from the poor cardiometabolic health outcomes we see today. 

The chance to apply my expertise and help make a difference was too enticing to pass up. I've built data science teams in the manufacturing space, but being able to leverage my technical talent to impact health outcomes and, potentially, how cultural conversations and technology shape the future of healthcare, is an enticing prospect.



 OH: Why has Omada decided to invest in expanding its AI capabilities?

TM: I think what makes Omada special is its human-led virtual care. 

AI can help amplify the best aspects of Omada, specifically the ability of coaches to build rapport and have meaningful connections with members. Our data shows that these factors can help members to succeed in the program. So, AI should be viewed as a tool to help that happen more at scale. 

There's an “AI doomsday” scenario out there in popular culture, but I think, as is the case with any tool, it’s about the intent of the user and its responsible deployment. Our intent in using AI at Omada is to support our coaches and free them up to develop more meaningful relationships. And we’re committed to prioritizing privacy, security, and safety in our use of AI.



 OH: What makes Omada well positioned to innovate with AI?

TM: Omada has a robust data set of billions of data points collected over 13 years. On top of this, is our differentiated care model delivered by proactive, human-led care teams. AI allows us to extract meaningful insights from these data sets to help scale our human-centered care model and, ultimately, enable them to help more people.



OH: How is Omada thinking about AI in terms of chronic condition care? 

TM: We believe that with a private and secure data architecture AI’s ability to parse through insights and surface important health behavior patterns can be very impactful for chronic condition management. AI can help identify behavior patterns that a member may need to change to improve their health. It can also provide real-time feedback on how interventions are getting them closer to their goals, or if different interventions need to be explored. And finally, AI has the power to surface new wellness content to a member that helps them think and create new paradigms about their condition. 

 

 OH: Why should members be excited about AI playing a bigger role in their health care?

TM: Omada members can get insights unique and specific to them based on inputs, like A1C values that members choose to share or glucose values that come directly from the third-party, FDA-approved devices Omada sends to members. This fuels the ability to tailor treatment and the experience of the program for the member that they might not otherwise get without AI. The alternative often is, and has long been for much of healthcare broadly, a de facto one-size-fits-all approach. AI can really unlock how the program is tailored for each member. 



 OH: What excites you most about Omada’s future AI innovations?

TM: It's the ability and it's the promise of the whole field of data science to improve health outcomes and people’s lives that excites me. Healthcare professionals have been looking at data at a population level for years. That’s how researchers originally measured rates of mammogram screening, for example. By looking at those rates, the healthcare system improved them over time. Now, with more sophisticated methods and faster computing, we can look for patterns to improve care in many other ways. We can really learn from the big corpus of data that we have here from our interactions between coaches and members, apply that to make our program more personalized and meaningful for each member, and hopefully help more members reach their health goals. We have the ability to use those as learnings for how to provide better care and how to better interact with members. 



 OH: That’s a lot of data. How do you keep it private and secure?

TM: All of our work occurs within Omada’s existing secure data architecture or with partners with whom we have appropriate security and privacy terms. We are careful to ensure that our use of member data comports with applicable legal requirements, like HIPAA and other applicable health information privacy laws, and we don’t use data files from our employer or health plan customers in our AI work. In many cases, to do this work, we use anonymous data that doesn’t identify specific people. 

This is all very similar to how a typical health system today might use the data in its EHR to track the effectiveness of its programs, for example, across a multi-state health system or to consider whether one medical device causes more issues than another. The learnings that we can develop then become important as our care teams take what we learn and work to apply it for each member.