Nicole Bocskocsky’s Post

A whirlwind of activity last week at the America's Physician Groups conference. Lots of solid insights shared by health system leaders. Some notable thoughts: +++ Artificial intelligence is intelligent in different ways: By now, we all know ai has a diverse set of use cases, incorporated in different applications and tools. But there are fundamental differences in how the intelligence itself is brought to life. Examples include: - Predictive AI, used to forecast future outcomes by leveraging historical data. Shout out to UCLA’s CKD/ESRD early intervention protocol (Joycee Berin) and Diagnostic Robotics’ intelligence care post-dicharge journeys (Kira Radinsky) - Protocolized AI, used to automating a user’s administrative load by using a set of detailed, pre-defined rule sets. Thanks to ARC for sharing their experience using Elaborate 's automated outpatient lab summaries (Manish Naik) - Generative AI, used to create new content. Great overview of ambient documentation at MGB as well as Cedars’ Sinai talking through its medical ontology model powering acute triage (Caroline Goldzweig) Across system leaders and experts, there was clear agreement that the first step to implementing intelligence intelligently is defining the ‘jobs to be done’. That way, internal and external stakeholders are clear on what they're solving for when they architect the solution.  +++ Scaling AI takes time and rigorous observation: Market leaders describe a phased pilot approach that allows clinicians to build confidence and system leaders to measure ROI before deploying systemwide. Loved hearing Rebecca Mishuris outline MGB’s novel approach of piloting two solutions against one another, to pressure test implementation, workflow, and value creation. +++ Proving ROI continues to be challenging: Lots of vendors using lots of different metrics to prove value, but the majority of leaders agreed that estimating the ROI of these new tools today is part science, part art. Demonstrating the ROI of the “soft” metrics around provider burnout and retention seemed to be a big opportunity, especially for investment in more efficient primary care and PCP wellbeing. After all, PCPs drive referrals to specialty-level care (and higher reimbursements) for health systems. Appreciated Anurang Revri's overview of how working team setup can also impact ROI estimates. +++ Value-based care = tools + humanity: VBC will require not only implementing tools to support clinical documentation, risk capture, upcoding, etc, but also humanizing interactions to engage patients. Countless examples, but one that made me smile was a depression focused chat bot responding in the following interaction: - Context: post-discharge check in Bot: Hi Freida, how are you feeling today after your surgery?  Patient: I’m having a tough time, am in a lot of pain.  Bot: Well, it can’t be that bad, I’m sure it will get better with a positive mindset!    🤣 🤣 🤣 Can't wait for next go-around!