From a Candle to a Light Bulb—a Conversation With Dr Jordan Dale of Houston Methodist

 Jordan Dale, MD, is the chief medical information officer (CMIO) and inaugural chief health AI officer of Houston Methodist, based in Houston.

Q: Let’s start by learning a bit about your background. How did your career path evolve and what led you to Houston Methodist?

I was an engineer by training before I went to medical school, so I think I was programmed at an early stage in my career to be more of a systems-based thinker than a linear thinker. When I went to medical school, I struggled a little bit because medical education is still very focused on cause and effect and memorizing massive amounts of information versus being a systems-based thinker.

Where I was training, I got aligned with the team that was implementing an EHR, and I had the opportunity to support that go-live and to be a clinician who had a very active voice in that arena. I also had early mentors who helped me understand how to balance clinical and technical skills. They convinced me that there's a path where I could use that background of being a systems thinker to help in a more universal way than having isolated interactions with patients.

Regarding my transition to Houston Methodist, I was looking for an organization that had laid a very strong foundation in technology but needed a leader to help with digital transformation at scale across a large system. I joined Houston Methodist about three and a half years ago and it's been a great fit. We're looking at opportunities to disrupt ourselves before we get disrupted using digital transformation and addressing the big problems in our market.

Q: In addition to being a hospitalist and the system’s CMIO, you also took on the role of chief health AI officer in late 2024. How have you adjusted to this dual role?

There's this concept of the “CMIO 1.0” where you basically were just helping all the physicians get across the finish line of using an EHR. That's not what CMIOs want to be doing or are doing anymore. We're much more strategic and trying to make sure that technology is being placed in care delivery where it's going to add value to our clinicians and to our patients. For me, the new role was a natural fit of all the same skills, but now just applying it to the fancier technology that people want to talk about now, which is artificial intelligence.

I’ve always been very comfortable working with a multistakeholder group that may have differing opinions on how we should make progress with digital transformation. I saw toward the end of 2023 that there was an opportunity to bring a group together to focus on AI oversight. I chaired that group and that's really what helped launch this new additional title for me and to formalize that we’re focused on thinking about how we can have constructive conversations about emerging technology. Today it's AI, but in the future, it could be anything else. That multistakeholder group includes our privacy team, legal team, some of our champions in the clinical space, and several other disciplines. It's predominantly focused on making sure that we're building the right structures that will be the foundation of our strategy.

Q: You’ve been piloting ambient listening technology for a bit and recently expanded it into ED and inpatient settings. What are some early results you’re seeing and what are some key differences in using ambient AI across care settings or clinician subgroups?

We have definitely leaned in and have hundreds of users on the ambulatory side using two products from Ambience Healthcare and Nuance. For our primary care group, they don't see a future where they're not using ambient listening. Just the breadth and depth of their conversations with their patients lend really well to this technology. Their usual workload means having 20 15-minute meetings every day and then at the end of the day summarizing each meeting with a level of detail to make sure the note is personalized, meets billing requirements, clinical requirements, quality, etc. The amount of burden that we've been able to reduce there has been significant. In primary care we're removing hours of documentation time every day.

We've had pretty rapid expansion across specialties as well. I think the technology still lends itself very well for those high throughput, acute specialties. Orthopedics is one specialty where we have many users, both because we wanted to challenge maintaining human scribes in those settings, and we wanted to see how much we could push the technology to transform that experience. In some of our other specialties where it's a mix of acute problems and chronic disease that they're managing over multiple visits—that's where I think we still see opportunity and that's where we've pushed those vendors on how they can further mature their products.

I think that's also critical for the ED and inpatient settings, which are very different types of encounters. Usually, a patient who is coming into an ambulatory encounter either gets there themselves or they have someone with them, so there's somebody you can have a conversation with. But while I was working clinically in the hospital last week there were many patients who were not in a state where they could talk with me to navigate their care. There's also a lot of complexity and nonverbal things that we're capturing in those types of encounters. So that's been our challenge to our ambient listening vendors, which is figuring out how we can mature these products to work in fast-paced environments like the ED, but also areas where we're trying to gather more context from the patient's chart if a patient is nonverbal or there's no family in the room.

Based on reactions from our physicians, it's been the most exciting thing that I've rolled out in my career. People don't usually say thank you for transforming their day-to-day, at least not in the first day or two. This is something where they realize the benefit and they're saying, “Never remove this from my life.” It’s great to be able to deliver that to hundreds of our clinicians. I tell my residents and students when I round that I don't see a world in two years where people are not using ambient listening in almost every workflow, and that includes nonproviders. We really need new clinicians to focus on other skills beyond documentation in their first five years of their career. I think it will be a good paradigm shift for those entering the workforce and understanding how this could redirect their time to be excellent clinicians.

Q: What are some other near-term AI applications that you think have strong potential?

We have several other applications. If we look at nursing on the inpatient side, they do 8- to 12-hour shifts and at the end of their shift, they're expected to write a note on all their patients. Most of their notes just summarize what they already did in the EHR over that shift. So, we've seen great results with piloting how we can make an AI-generated draft for them that they can then further personalize. And if they were efficient in doing this before, it’s probably because their notes were not highly personalized to that shift or that patient. When we add AI, we can see it's adding a greater level of detail while summarizing what they've already done, so they're seeing the benefit of how the system's working for them.

At our main campus, over 70% of our discharge summaries are now AI generated. For long stays where there are multiple teams involved, a human doing that work today is going to struggle reading basically a Moby Dick novel worth of content in your chart. To summarize all of that is a challenge and so having them start from a draft to verify key details and make sure that the final plan is incorporated—that's reducing the time that those notes are entered into the chart by almost a full day. There is just lots of opportunity in reducing administrative burdens and that’s where our early focus is.

Q: And you’re also piloting Apella’s platform for ambient support in the OR?

Yes. In the same vein, event tracking in an OR is not a staff member’s primary task. Usually, they have a lot of other things to do, and so just the validity of that data that they're tracking manually in a system is not going to be high quality. And so, when we incorporate Apella’s cameras, we're really trying to do process mapping in real time of every case to understand where there are opportunities. For example, we've detected where a scrub tech was always trained, maybe years ago, that a certain tray should always be opened even though it's only used in 5% of cases. We can now go back and see that and say don't open that tray unless it’s for the 5% of cases that you need it. And that can save minutes per case to better utilize that individual OR. That's been a huge opportunity of how we can use some of that machine vision type of technology to do process improvement work.

Q: The newest buzzword is agentic AI. Is it fair to say this is essentially a chatbot 2.0 opportunity or is there something more here?

I'm typically the healthy skeptic in the room, but I do think there is some unique potential for it. Similar to how [large language models] have been applied, their broad flexibility opens up opportunities in conversational use cases that couldn't be done before. Everybody knows how to break a chatbot 1.0, right? I think with agents there's a lot more potential to make it feel like a true experience that meets the user's expectation. I think there's a lot of potential to use them in administrative areas first, like with appointment scheduling or rescheduling; all the things that a patient is calling in for that don’t require a clinical license. We can have an agent that collects a lot of information and then pulls somebody into the encounter if needed.

A challenging thing with AI is how it has been riding this wave of hype, and I think the cost that's represented to organizations follows that hype. I think we're starting to ground that a little bit in the generative AI realm, and we need to do that in the agentic AI realm as well. These should be calculated explorations with shared risk between the organization exploring AI and the vendor to fully mature their products—then we'll see the true value. In health care we shouldn't have tolerance for hype because that's impacting dollars that we could be directing immediately to care delivery or recruiting clinicians, etc. So that's where my mindset is at—making sure that we're having grounded conversations in its potential and making smart bets for exploration.

Q: In late 2023, your organization opened the Houston Methodist Tech Hub at Ion, which is described as a living laboratory and space for community collaboration. Then in late 2024 came the launch of the Houston Methodist-Rice Digital Health Institute in partnership with Rice University. How do these partnerships and collaboration spaces influence your team’s approach to innovation?

I think the key theme is we never want to be doing this transformation alone. We see value in sharing and coworking through these problems and understanding where the true limitations are and having transparent conversations about it. Regarding our space in the Ion, we’re the only health care organization there, so it was a nice jump for us to get connected with other organizations that aren't just in our market.

These are coworking spaces where we can do rapid prototyping as we have a lot of our test networks and test systems there. We can also have a vendor come in to use the space for a project—so we don't just take their word on their desire to partner, we can say ok show up for a few weeks to evaluate if this partnership is going to mature into something that we want to do long term. How our partners show up is as important to us as what they can actually deliver. That's just the reality of how we do innovation.

And then the Rice opportunity gives us areas to evaluate and develop things premarket in a way that we haven't had a dedicated space to do that before. We can leverage one of the premier educational organizations in Houston to understand how we can educate the next generation of people who should have these methodologies and understanding and be successful in whatever market they go into. But we also want an arm where we can do some of that rapid prototyping with dedicated teams that are eager to enter and solve big problems. It just gives us a longer runway for our innovation journey.

Q: I saw an article you wrote where you said the difference between a clinician with or without AI is like the difference between a candle and an electric light bulb—can you speak more to that?

The point of that statement is I do not think any clinician in the future is going to say, “I want to go back to practicing without any AI in my workflow at all.” The reality is, even for skeptics, they're using AI in their workflows today and they may not even know it. One of the examples I give to consumers is how spell check on your phone is technically AI, and I can almost guarantee everybody uses that nearly every day. I think that opens up some understanding about how although it feels like things are rapidly changing, it has actually been going on for a long time. There are early use cases that you know you cannot live without in your work today, so how do you get to that next step where you're embracing some of that change?

Our unique approach at Houston Methodist is that we do not want to just deploy a technology, pat ourselves on the back and move on to the next. We're hyper focused on adding value and engaging stakeholders the right way. We see high adoption rates in how we focus and guide our users throughout the entire process. So, our people who have been sprinting on ambient listening for six months are much more open to trying the next thing. They've had that “aha” moment. I think the potential with AI is that clinicians can actually be the ones that drive us to say, I want that light bulb, I want the latest technology every time I'm taking care of this patient because I know it works for me and with me to provide higher quality care.

This blog post is part of an ongoing Q&A series with digital health leaders at organizations across the US. These in-depth interviews aim to provide real-world perspective and insights in this rapidly evolving space. Don’t miss the other blogs in the series, found here.

Sg2 members are encouraged to check out the full library of digital health and AI resources in our online resource kit. Not a member? Email learnmore@sg2.com for information on the expert intelligence, data-driven insights and strategic perspective we offer to health systems nationwide.