Dr. Yumi Taylor is reinventing Sutter Health's primary care practice

Taylor is a board-certified internal medicine and clinical informatics physician, and a founder of Tera. She received her medical degree from the University of Pennsylvania School of Medicine and worked for six years in rural Pennsylvania, where her practice was recognized as a five-star quality clinic by Highmark BCBS and CMS (Center for Medicare Services). She recently completed a clinical informatics fellowship at Stanford, where she learned the ways data and technology can deepen relationships and improve patient health.

Taylor is a voracious reader (and audiobook listener) who enjoys running, swimming and playing board games with her two children in her spare time. In addition to English, Taylor speaks conversational Japanese.

VatorNews: Tell me a bit about yourself, your career and your role at Sutter.

Yumi Taylor: I have an incredible position that landed on my lap 2 years ago. I am the founder of Tera Practice, a new primary care delivery model within Sutter. We run like a startup within a large organization.

My background is as an internal medicine primary care physician and fellowship trained clinical informaticist. Sutter was looking for someone who could redesign primary care from a blank slate. They had many existing efforts to redesign current practices. I realized it was an incredible, yet risky opportunity. I mean, it could be a great opportunity to have a meaningful impact in a large organization or it could be one of those opportunities where you can’t move quickly enough to accomplish much. We’re fortunate that leadership were thoughtful about the right stakeholders sponsoring the project. The Design and Innovation team (Chris Waugh, Linette Fung), Digital Patient Experience (Dr. Albert Chan, Guy Tennyson), Palo Alto Medical Group and Foundation (Doctors Rob Nordgren, Elizabeth Vilardo, Mat Hernandez, Jenni Gillett, and Ron Sinha), and the Sutter Health Plan (Phil Jackson, Peter Bergamo, Rob Carnaroli) have been key champions for our work. Our project has evolved to become an initiative at the C suite level. It’s been a really wonderful learning opportunity for me and a dream to fix some fundamental things that are broken in healthcare.

A little bit about what informs my perspective for the model: I started out in private practice doing traditional primary care and eventually became involved in an innovations project with a large payer to transform our fee for service rural practices into value-based payment models. Care team visibility to data, investment in panel management, and the challenges of timely workflow changes were key lessons I learned from this experience. Wanting to learn more about healthcare IS operations and data governance, I completed an accredited fellowship in clinical informatics at Stanford. During this time, I was exposed to the national landscape of innovations in care delivery and tech, federal regulation around healthcare data, and AI. A 6 month internship with the Brain team, Google’s healthcare AI shop, helped me better understand the potential and limitations of AI to transform healthcare. 

I’m grateful to have a position where I take my clinical care experiences and health systems learnings to, as the overused adage says, “transform and modernize healthcare."

VN: When you said that there are things that are broken in healthcare, what were you referring to?

YT: Oh, this is a big question and many things are broken. I’ve tended to focus on the things broken along the quadruple aim for TeraPractice. The quadruple aim pillars are patient experience, total costs of care, quality of care, and care team experience. Primary care plays an important role in a health system accomplishing these pillars. We’ve taken an ambitious approach to tackle all 4 pillars. We are bringing care to patients, instead of expecting them to come to us through a virtual first approach. We have redesigned the care team and roles, upshifting responsibilities to allow ownership of panel management and patient engagement. We give data to the care team to drive actionable interventions from a team that understands how to personalize an effective care approach and prevent more costly care utilization.

VN: Tera is a virtual-first model, which is something relatively new in healthcare. How has technology evolved to allow that to happen?

YT: Electronic health records like Epic now have integrated video, mobile health data and text messaging capabilities that integrate with the patient’s secure patient portal. The technology integration makes for a seamless patient experience and one central place where communication about their health occurs. The technology and EHR integration continue to rapidly evolve, but I’d say the pace of clinical care team adoption of these tools is much slower. To increase the pace of these technology adoptions, there must be an evolution in payment for care delivered using these modalities, federal statute clarification on privacy rules for SMS text messaging, and care team member roles and responsibilities. 

VN: What works and what doesn’t with telemedicine when it comes to demographics, including age and socioeconomic issues?

YT: When we first launched, we did our initial patient meetings as a video visit and purposely enrolled patients who had at least one chronic disease and seniors. We wanted to make sure we tested outside of a young, healthy demographic. We learned very quickly that telemedicine is welcome across any age range, disease state, and socioeconomic. The key factor for adoption instead was individual. Was this person open to try something unfamiliar to them or did the current way of doing traditional care work just fine for them. Anecdotally, I noticed for patients born outside the US, more often a discomfort with a virtual relationship. To counter this, we did our first visit in-person instead to establish the relationship. We found once they understood the value of increased access to their primary care team by virtual means they welcomed a transition into a virtual relationship. 

Another key factor for adoption was price point. In our practice where 30 percent of the population has an income under $50,000, if a virtual touchpoint saves a co-pay, this becomes a preferred choice. Another key reason for adoption is convenience, no surprise here, who wouldn’t want to save travel and wait times if an in-person visit isn’t needed to resolve an issue.

VN: Being a startup inside an organization of Sutter is an interesting and unique way to run a company. Basically it’s like you have the best of both worlds: the nimbleness of a startup with the resources of a big organization. Tell me what you’ve learned from that experience?

YT: We’ve certainly been learning and figuring it out on the fly. A few key learnings: number one, a large organization has lots of internal talent. Build a team of individuals with diverse skill sets that understand the org culture and decision making structure. Number two, executive sponsorship is needed to remove barriers. Number three, Champions build excitement for the model to raise tailwind for the startup, Number four, in parallel to number three, develop some things under the radar/in a less visible way to accomplish more quickly. Number five, build upon existing functional infrastructure, particularly IS systems and operational assets, so that you can use this energy instead on fixing or developing new areas. 

I have been surprised, knock on wood, how little resistance we’ve gotten. We just needed to have a concrete plan and do. I think people are generally excited about innovation and if an idea makes sense, solves/addresses key problems in their realm, has a business case, and the team has credibility, then full steam ahead. 

VN: What benefit do patients receive from being a part of Sutter? 

YT: Sutter is an integrated health network. Being part of Tera is access to a modern, personalized care team that knows you, your life. No repeating your story. We fit into a patient’s life, not the other way around. This team can resolve most issues virtually, provide coaching support for lifestyle changes, and help you navigate the larger resources within Sutter when needed - specialty, ancillary services, home health, palliative services, hospital, and retail clinics. We call it, “primary care, in your pocket” with the backing of a large health system. Tera has a 4.9 out of 5 star patient rating. 

VN: Are there any hard numbers you’ve been able to calculate, like cost savings or improved patient outcomes?

YT: Yes. Data is a large part of what we use in our day to day care as well as to measure impact on cost, quality, patient and care team outcomes. Does that sound familiar? Those are the quadruple aim pillars. We are excited about our early findings of significant savings in total costs of care. 

VN: What are some of the things you see coming next in healthtech? What’s exciting for you?

YT: Lots of exciting things are happening in AI, mental health, and mobile monitoring. Sutter uses Ada, a symptom assessment software. Patients input symptoms into a decision tree which then reports an initial assessment and triage suggestion. Not quite AI but the potential for learning exists. Exciting work in mental health solves problems with access. Virtual access to licensed therapists are needed but also less human heavy approaches like self-guided modules and chatbots need evaluation. Mobile monitoring still holds promise, although, we still need to figure out engagement for its adoption in at risk populations, not just those trying to optimize their health. Finally, not a sexy healthtech tool now, but critical in managing total costs of care are population health management software. This software interface should integrate data sources between patients, care teams, employers, and payers to support the scale of value based care.

VN: When you talk about data, AI and machine learning are often a part of that. Do you use those technologies?

YT: As I mentioned above, Sutter uses Ada as a patient facing tool. Ultimately machine learning will need very large datasets to train their algorithms and I expect to see more partnerships between large health delivery systems and AI invested companies. A challenge to current healthcare data are the missing social and behavioral risk factors that do impact healthcare costs. At Tera, we are capturing these meaningful data points which will add dimensionality to risk.

VN: So you’re saying that other companies can use your data to facilitate their own AI and machine learning?

YT: Yes, I think those partnerships would be interesting.

VN: Is there anything else I should know about you or Sutter?

YT: These are my own views and does not represent what Sutter might endorse.