The Oakland, Calif.-based health system offered Abridge, an ambient AI listening tool that drafts clinical notes for the EHR, to tens of thousands of providers.
Becker’s caught up with Brian Hoberman, MD, executive vice president of IT and CIO at the Permanente Federation, at the HIMSS conference in Las Vegas to find out how the implementation is going.
Question: How has the rollout of ambient AI at the Permanente Medical Group been going so far? What have been the benefits and challenges?
Brian Hoberman: People love it. Whenever I talk about it with people, whether they know what I do for a living or if I happen to be in the exam room with a relative and the doc doesn’t have any idea that I work for KP, what I hear is: “This is great.”
The ultimate goal of healthcare IT was always: Make me a better doctor. We did a lot of things that made the healthcare delivery system much more effective in delivering quality and a lot of other great things. But people didn’t feel it, and then you come along with ambient, and the clinicians really feel it, and their patients feel it.
As you and I are talking right now, I’m looking at you and you’re looking at me. That’s what ambient does in the exam room. So it makes the doc and the patient, or the practitioner and the patient, connect more. It actually is a better listener than human memory. It captures more details. Docs will sometimes read the note — because, of course, they always read the note that ambient creates because they have to edit it — and they’ll say, “Oh, it picked up on some things that I hadn’t remembered that are actually important.”
So it truly does meet that make-me-a-better-doctor criteria. Docs always want the patient to have a good experience. Patients have a better experience. It’s a win-win.
Q: How many clinicians are using it currently and what is your goal? What will it take to get to that goal?
BH: I’m not a numbers guy. What I can tell you is that we’ve deployed it across the whole enterprise, and every doc and advanced practitioner has access to it.
I can give you an example that’s really local to where we are right now. If you’re not staying here, do you want to walk to your hotel or take an Uber? It’s a choice.
So if I’m a gynecologist and I’m going to see a patient who’s got a complicated complaint, I’m going to use it. It’s going to help me take a better history. But if the patient is coming in for a procedure, I have a template. So it just depends on what I’m doing and does the tool work for what I’m doing. Particularly for visits where history-taking is important, it’s indispensably useful.
So the question then is, how many people are using it? I don’t know the answer to that, but what I can tell you is that, because the tool is so well-liked by people, it spreads virally. People talk about how much it helps them. Somebody who might even be a little bit reluctant to try technology discovers it’s actually very easy to use because it’s integrated into the EMR and, lo and behold, you have a new user.
Q: Where do you see ambient AI technology going next?
BH: I see it evolving to mature for different settings, different roles — like nursing, of course, the ambient vendors are talking about — and for different specialties.
The question is: What else can it do? It’s a very interesting situation ambient finds itself in. It’s listening to the doc and the patient. The EMRs aren’t listening; the ambient vendors are listening. So is there a place within the listening where it can nudge? Can it be a real-time decision support before the doc actually looks at the computer screen where the EMR would typically do it?
Of course, when they’re integrated in the EMR, some of that is actually in the EMR itself, but it’s in the mobile version of the EMR. A frontier in the industry is going to be how do we take advantage of the ability for ambient to understand what’s being said and work with the EMR to provide useful nudges in whatever type of thing might be useful to the clinician.
Again, it falls into that same category of: Make me a better doctor. What do I need to ask a follow-up question about? What is it about this particular patient that I should ask about?