I want to share my takeaways from the HHBC conference panel "Analyzing the Return on Investment (ROI) of Artificial Intelligence in Healthcare" last Feb 28th.
?
We began with Nestoras Mathioudakis's testimony, which transported everyone in the room to a day in his medical practice prior to using Abridge's AI scribe, and honestly, I felt quite stressed.?
?
Wake up at 5:30am -> prepare for patient visits by reviewing charts,?collecting outside lab and relevant diabetes data ->??eat breakfast -> go to clinic -> see 16 patients/day -> discuss and?document?patient history, exams, lab data, diabetes device data (glucose readings, insulin pump settings,?etc.),?write an assessment, order meds, order lab,?type patient instructions…
?
All in 20 minutes for each patient visit!
?
Go home -> eat dinner -> maybe exercise ->?2.5 additional hours of?note writing?and responding to patient messages?-> go to sleep 11:30pm.
?
Today he uses an AI scribe that records conversations with patients and generates visit notes. He explained the scribe eliminates most of the tedious documentation work, enabling him to remain fully present with the patient. It also encourages him to verbalize his thoughts and observations from the electronic health record (EHR), fundamentally transforming the patient interaction.
?
He has regained his evenings with his family, additional time during each patient visit spent on patient education, and restored his joy of medicine!
?
Reza Alavi, MD, MHS, MBA challenged us to think about context when assessing a specific AI solution. He explained the value of any healthcare innovation, depends on the specific market conditions, workforce availability, and the existing performance baseline of healthcare organizations. What may transform one practice dramatically might only deliver incremental improvements elsewhere. To truly assess ROI, we must ask: Who is the economic buyer, what is their current state, and are we providing affordability, quality and or experience improvements?
?
Maulik Majmudar, M.D. equipped us with a structured framework to analyze the value of AI in healthcare across four dimensions: clinical, operational, financial, and patient experience. He challenged us to frame our thinking in terms of direct and indirect ROI and gave us concrete examples.
Zachary Lipton inspired us with his journey leading product development at Abridge. He openly addressed skepticism around using pre-trained models, describing how real-world clinician/patient interactions feed back into an ongoing process of model refinement. He highlighted critical guardrails around user privacy and rigorous auditing processes—ensuring that innovation never compromises patient safety or clinician confidence.
I want to personally extend my gratitude to all the panelist. The energy in the room was felt! I also want to thank Prof. Tinglong Dai for your thoughtful structure organizing this discussion & enthusiasm moderating it.
Graduate Healthcare Business Association
