Cleveland Clinic’s Rohit Chandra Sees Clinical Scribe as First Step in AI Journey
Cleveland Clinic recently announced plans to work with Ambience Healthcare on AI clinical documentation and point-of-care coding. In a recent interview with Healthcare Innovation, Rohit Chandra, Ph.D., chief digital officer at Cleveland Clinic, discussed the health system’s assessment of multiple vendors based on product quality and long-term potential. Cleveland Clinic aims to roll out the technology to over 4,000 physicians, with over 2,000 already trained.
Founded in 2020, Ambience is headquartered in San Francisco and has raised $100 million in total funding. Its solution has been deployed at health systems such as UCSF, Memorial Hermann Health System, and John Muir Health.
During the interview, Chandra said the AI technology has shown significant improvements in provider and patient experiences, reducing burnout and administrative burdens. He also highlighted ongoing AI projects in risk prediction, sepsis detection, and readmission prediction, emphasizing the transformative potential of AI in healthcare despite its complexity and costs.
Healthcare Innovation: Before we talk about ambient AI for clinical documentation, could you talk a little bit about your role there at Cleveland Clinic, and perhaps how your experience in other fields besides healthcare helps you bring a fresh eye to rethinking the digital experience, both for patients and for employees?
Chandra: Of course. I joined the clinic about three years ago as chief digital officer. I’m responsible for leveraging technology and AI in our clinic. My background is as a software engineer, and this is my first experience in healthcare. One of the aspects that I try to bring is technology expertise, as opposed to healthcare expertise. Obviously I have to put in the time and the effort to understand the domain, but then every once in a while, ignorance of healthcare can be an asset as well.
HCI: I’ve interviewed several health system chief medical information officers about deployments of ambient AI for clinical documentation, and it seems that regardless of which vendor they chose to work with, most of them expressed pretty deep satisfaction with the uptake by the clinicians and the potential impact on physician burnout. Have you been surprised at all by how widely these solutions are being adopted, and this positive feedback we’re hearing from clinicians?
Chandra: The answer is a resounding yes. I’ve been very pleasantly surprised. To use a phrase that is used in the tech industry, the scribe capability has almost perfect product-market fit in that it’s a relatively easy technology to use. You don’t have to do drastic changes to your workflows and processes, and it tackles head on a significant pain point for both patients and providers. I wish every technology innovation that you brought into healthcare had those properties.
HCI: I read that Cleveland Clinic conducted a pilot program of AI documentation solutions last year. Many of the large academic medical centers we’ve spoken to have chosen to work with Abridge. Did Cleveland Clinic do kind of a “bake-off” and let the clinicians try a bunch of different solutions and see which is the best fit?
Chandra: Our belief going into this journey was that the technology has the potential to be transformative with the practice of care, and should be a significant lever that can improve both patient and provider experience. So we wanted to think about the decision from a longer-term perspective.
We know that we’re in the early stages of AI, so the tools that we buy and deploy today are likely to evolve in the future. We’re not just assessing the tool as it exists today; we’re also assessing the company, its ambition, its vision and its roadmap on how it wants to engage with healthcare transformation. We’re trying to make a multi-year decision here.
HCI: Can you talk about some other potential use cases as these platforms evolve? Could they extend into revenue cycle or integrating more deeply into clinical workflows for things like visit preparation or referrals and orders?
Chandra: The entry point that everybody, including us, is starting with is using these technologies to document an outpatient encounter, and they’re remarkably good at doing that. But there also are tasks that you do pre-encounter. Can you summarize the chart before you go in? Then there are follow-up instructions. How do we expand from outpatient settings to ED settings, maybe even to inpatient settings, maybe to nursing use cases? For inpatient encounters, can you generate discharge papers for patients? All of these have their own variations in terms of workflow for a nurse or a provider in those settings.
HCI: Do you think the EHR vendors such as Epic and Oracle will try to do this themselves?
Chandra: I think the answer is yes, and I think in some ways it’s a good thing. These are early stages of technology, and it’s actually helpful for the industry at large to have multiple players bring their particular expertise, their particular perspective on the problem. That’s good for innovation — to have multiple people go to the same problem.
HCI: How widespread is the use of Ambience right now across Cleveland Clinic, and are there plans for scaling it across the whole organization in 2025?
Chandra: We started rolling out in late February, and we have a plan that’s targeting a little over 4,000 physicians out of about 6,500 physicians that we have. We’re trying to go after outpatient settings. I will say adoption is going faster than we expected. I think we have, at this point, more than 2,000 physicians that have gone through some basic training. I think the adoption has been amazing.
HCI: A couple of weeks ago there was a report published by the Peterson Health Technology Institute that said while ambient scribe technology is likely to improve clinician burnout, the financial impact on health systems was still unclear. Do you feel confident that it’s going to have a strong financial return on investment impact for your organization, as well as alleviating the administrative burden on the clinicians?
Chandra: Obviously, this is not providing direct financial benefit to the organization, but there are areas in which we hope it will help. First and foremost, it improves the provider experience and the patient experience. The quality of the provider/patient encounter is markedly better, and let’s face it, when all is said and done, that is priceless. We do hope that the technology will help reduce burnout, will help reduce turnover, will make it easier for us to scale our services. Also, downstream coding is closely tied to documentation. So better documentation should lead to better quality coding. We hope to get some benefits in all of those are areas.
HCI: Is there a whole plethora of AI projects impacting different areas of the clinical world at Cleveland Clinic that you are involved with?
Chandra: David, you’re describing my day job. That is what I do all day long. It is impacting both clinical and non-clinical areas. So just to give examples, there are plenty of opportunities of using AI in the back office. We look at areas like coding, automating managing hospital operations, where predictive analytics and forecasting are important tools to help us do our job more efficiently. The scribe is an example where it is really trying to alleviate the documentation burden on physicians. So it’s not direct clinical care, but it is trying to help streamline our ability to deliver care.
We have a few areas where we are trying to use AI carefully in clinical settings as well, and that’s through risk prediction, so that we can find ways to have smarter and more precise alerts for a nurse or a doctor, so that they can pay attention to the most pressing patients.
We have a project where we are rolling out AI-powered sepsis detection and alerting for doctors. We have a project where we’re rolling out using AI to do readmission risk prediction. The better alerts will enable us to help take better care of our patients in the most appropriate setting of care. So we’re pushing on AI on a variety of fronts. The technology is always is not just evolving; it is also expensive, and it’s also not easy to deploy it in a complex organization like ours, so we have to be mindful of all of those factors. But there’s no question that in the medium to long term, I expect us to use AI in a variety of more areas to drive transformation.
