The number of health systems that choose to work with ABRIDGE, based in Pittsburgh, in deployment of the entire clinical documentation tools based on generative the AI continue to grow. Just last week, Abridge announced agreements with Mayic Clinic, Duke Health and Johns Hopkins. Other customers include Yale New Haven Health System, Emory Healthcare, Kaiser Permanent, health network at Vermont University and Uchicago Medicine. Eric Poon, MD, Duke Health’s Health Information Director, spoke this week with Health innovation about the implementation that begins there.
The ABRIDGE AI AI platform will be available for 5,000 Duke Health Doctors in more than 150 primary and specialized clinics. The platform will be used in Duke Health Integrated Practice and Duke Primary Care Clinics, which serve communities in the triangle and other places in North Carolina.
Health innovation: There are several market solutions to capture clinical notes, but it seems that Aberidge is winning many customers in the academic space of the medical center. Is there anything particular in your approach that was attractive?
Poon: We began to try two products in October, and although our suppliers were very happy with both .
HCI: Duke Health has been put to the test with a specific user group before considering expanding it?
Poon: We implement the two products in 160 of our suppliers, after making a security pilot on a smaller scale. We wanted to make sure that the products we were implementing were ready for clinical use, but after that, we displayed it to another 160 suppliers, and that is where we really collect many very useful comments from suppliers of several specialties.
HCI: Have you heard initial comments from doctors that it helps them to feel more present and concentrate on the patient in the interaction instead of having to think about documenting during that time?
Poon: Oh absolutely. And you are talking to one of them. Before using this technology, I did not realize how much of my brain during a clinical visit I was dedicating to being a central transcriptor and writing some cryptic notes in the electronic medical record while talking to patients. Once I could disconnect from the keyboard, I could have much more natural conversations with patients. The encounters really were faster for me. I could finish my clinical visits more or less on time, which had never happened before. I started using this technology and, of course, the notes were much easier to clean later.
HCI: So, has you started the generalized launch and is it completely optional for doctors to use or not?
Poon: We will certainly not do it mandatory. We are still in the first week of mass launch. Our doctors are very excited about this. We receive many positive comments during the 160 suppliers test. I would say that we have not heard this feedback very positive about technology for a long time, so that really made the decision to display the solution much easier for our leadership.
HCI: Does this also have the potential to work in an environment as an emergency department?
Poon: Yes, he does, and that is one of the things that we are going to try very soon. They have not yet done so.
HCI: I was intrigued to read that, in addition to implementing this clinical notes platform, Duke Health was exploring opportunities with Open to jointly develop other clinical applications that use environmental, and asked me if I could have some examples of areas where those found. The type of technology could have an impact of the way it has clinical documentation.
Poon: I think we are certainly very interested in implementing this technology in other clinical environments, and working with Aberidge to think about how I could use it in other upstream and downstream aspects of the patient’s encounter.
HCI: With regard to how AI is impacting clinical attention in general, do you see opportunities for things as better support for the clinical decision or better communication with patients through the portal?
Poon: Yes. I think we are beginning to see the introduction of generative AI in clinical practice. I think environmental technology has entered the scene very quickly. Also, separated from Abridge, we are analyzing how we can use AI to write MYCHART messages, something that other organizations have also been doing. We are doing a very reflective evaluation to think where it adds value and where it will be widely used. I think there are opportunities to use for some administrative functions, to help with several aspects of the income cycle, for coding, for graphics reviews. I think technology could summarize complex information in the graph. It can really help closer doctors to bed, so that doctors return to what led them to the profession first.
HCI: I understand that you are at the front of the broader deployment.
But, are there some obvious research questions that opens the use of this technology that would interest you to follow more or less?
Poon: Yes, we are certainly evaluating in all areas what is the impact of the introduction of this technology on the self -informed exhaustion of the clinician, whether there is satisfaction with the documentation practices, if the satisfaction of the clinic in general will improve and if we will make improvements in Our recruitment and retention and productivity. We are also actively exploring what would be the correct way to present this technology to doctors in training. That is still an unknown territory. I certainly believe that doctors in training should be exposed to this technology before they are performed with training and start practicing independently, because this technology will become increasingly ubiquitous. But when that should start, it is still an open question.
HCI: Is there a government framework in Duke Health around AI implementations? Do things have to go through a research process before implementation?
Poon: Yes. In fact, we have thought a lot in recent years. Even before the generative AI became the warm of continuing, we really stood a governance process. We call it ABCDS as what the supervision of the «algorithm based on clinical decisions support» represents. It is a multidisciplinary group. We work with the clinical groups that are interested in implementing AI in several cases of use. We ask them to present the evidence of the evidence of exploring technology, and we ask them to explain how they will evaluate it.
Once they have tried it, they will present the evidence that this is really having an impact before giving them the green light to continue deploying the technology. So, somehow, it is a fair approach in time to give some guidance. We do not ask people to demonstrate in the nth grade that something is successful before they try it. But this is a way of promoting innovation while we maintain security, efficacy and equity in our minds while we use AI; It is really part of our responsible strategy.
