As executive vice president and chief digital officer at Providence St. Joseph Health (PSJH), a 51-hospital health system spanning across the western U.S., Aaron Martin’s job quickly shifted once his health system—and the broader Pacific Northwest area—started to see COVID-19 cases trickle into the region.
The health system, known for its rapid innovation efforts thanks to its venture capital arm, called Providence Ventures, was able to leverage digital solutions it had already created and pivot them to meet the growing needs caused by the COVID-19 crisis.
Martin, previously an executive at Amazon, recently spoke to Healthcare Innovation about the different ways PSJH has been able to use digital health to respond to the health pandemic. Below are excerpts of that interview.
As chief digital officer, how do your team’s responsibilities change when a public health crisis such as COVID-19 hits?
We have gone through several phases. The first phase was about ensuring that non-COVID patients and the “worried well” didn’t show up to our clinics—when they were still open—and mix with people who might have [the virus]. So, how do we make sure that the health system does not become a vector for disease transmission?
The way we have dealt with that is by launching a self-assessment chatbot—called Grace—using the framework we fortunately had already built for communicating with patients in different situations, such as for FAQ and directing them toward the right care venue based on their chief complaints. So we pivoted that technology, launched a microsite, and then effectively helped patients self-assess on whether they needed to seek care at all based on certain risk factors.
We partnered with Microsoft using their health bot, and it works by inputting different “brains” into Grace based on the [specific] situation that calls for it. In this case we needed something that was quickly going to focus on self-assessment, and then later on we launched an FAQ if you had questions about COVID. That first piece of it was to help comfort people where they could feel like they were having a dialogue without actually having one, and this technology helped take patients through the self-assessment process. We reached somewhere closed to 15 percent of Seattle’s population with this.
Secondly, we scaled up virtual visits. We had built a technology we had been working on for a little more than three years ago called DexCare, a platform that supports our ExpressCare [urgent care] brand. This is a same-day service that serves folks who have a wide variety of concerns and illnesses through either telehealth, in-person clinics, or home visits. After we shut down the in-clinic and home visit pieces, we pivoted everything toward virtual visits, and we [have seen] a 20x increase from what we are used to seeing in terms of volume. We’re on a run rate of 400,000 video visits this year, and last year we had 15,000.
We have a team of about 120 product and software engineering folks, and most of them hail from places like Amazon and Microsoft. I used to get on our engineers about possibly over engineering our solutions for scale. You always have a trade-off between three things [when you’re building solutions]: technical debt, how much you are building for scale and how much you are building for features and functions. Thankfully, they built [DexCare] for scale, so it behaved much more like an e-commerce platform than other telehealth platforms in the market.
The third thing we did was work with our clinical teams to find them a solution for at-home monitoring. We asked ourselves, for some patients who were coming in or were visually assessed through telehealth, how do we make sure they are being well cared for so they don’t need to come into the hospital? But we are concerned enough about them—either they are in a high-risk category or are exhibiting symptoms—that we need to check in on them frequently. So, how do you do that at scale?
We used two portfolio companies of Providence Ventures to help with that—one which was Xealth, a platform that allows you to prescribe anything that’s not a pharmaceutical directly from the EHR. So we can very quickly identify the patient and prescribe him or her via the second part of the equation—a company called Twistle that allows you to do check-ins effectively through SMS [messaging]. And in this case we were collecting two pieces of information—pulse oximetry and temperature. So they were able to quickly scale that up where they are now monitoring over 700 patients on a panel, and that will expand into the thousands soon.
So, there have been three core pieces: how to help people self-assess and avoid showing up at the clinic unnecessarily; how to replace the clinic with the ultimate PPE, which is telehealth; and how to monitor at-risk patients at home, and do it at scale. This all happened within 15 days, so we shifted a lot of technology very quickly.
How tough would it be to deploy some of these innovations in just a matter of days or weeks if you didn’t already have the foundation in place?
Part of our model [at PSJH] is that when we build something new—such as DexCare—we see if we already have a solution to the problem. If not, we go out and look for one in the market, and that’s where Providence Ventures comes in. If we don’t have a solution and can’t find it, then we actually build it; and if we build it we create a company out of it. We just recently hired the CEO for DexCare, and even before we spin these things out we typically pine up other health systems for the technology we have built. With Xealth, UPMC was the second customer to us; they became an investor and are now on the board.
[I can’t publicly name it yet], but we launched an entire health system on DexCare in the Midwest while all this was going on. Healthcare IT is not known for alacrity in terms of how quickly health systems and plans typically get things done. In the case of [DexCare], we have seen four five other health systems [show] interest. It’s happening quickly since this platform serves telehealth, home visits, and in-person clinics. The real pivot is towards how soon you can get the telehealth platform up and running. There has been a sense of urgency with climbing rates of telehealth adoption [in most places].
We have it pretty much knocked it out to the point in which we could get a new health system deployed with the next DexCare platform within 30 days, including an integration with Epic. And that’s assuming it’s a priority for the health system. What typically hurts you in these deployment timeframes is that you get put into a queue, so it depends on how big a priority the [project] is for the health system. The Xealth/Twistle solution for home monitoring we stood up in four days. The configuration work that’s needed to be done is not much; it’s where it sits on the priority list.
What lessons learned can you share from an implementation perspective?
You don’t get extra points for originality or doing it yourself. For every one of these solutions, we have 10 or 20 health systems looking into deploying them. Hopefully if you are [located] in a state not getting hit as hard, so you will deploy it and it never becomes an issue. You don’t want to be in a position like New York is in where you have to skip to the end and are focused on the patient surge instead. But I would suggest to start making these investments as soon as possible to get ahead of it, and be able to assess and communicate with your patients.
The key questions are: How can you get patients to do self-service and self-assessment to determine if they have a problem so when they come in for virtual visits you can more accurately and more efficiently serve them? And then how do you deploy home monitoring at scale? Our answer was that you need high connectivity to your EHR for continuity of care, and then you also need some way of communicating with the patients.
How are clinicians taking to some of these newer tools for COVID?
It’s the ultimate PPE, so it’s protecting them. The more work you can do remotely and safely where you can take care of patients even though you’re nowhere near them, the better. That’s the whole game [behind] how digital is impacting COVID.
