STAT AI: An Emergency Medicine Podcast · Ethan Abbott and DJ Apakama

EP8: Rishi Khakhkhar MD, MBA

·42 min·3 clips

Rishi explains how emergency medicine misses the first triage decision when patients Google symptoms instead of consulting doctors.

This episode features emergency physicians DJ Abhakama and Ethan Abbott interviewing Dr. Rishi Khakhkhar, a practicing ER doctor and founding medical director of Council Health. The conversation explores AI's role in emergency medicine and asynchronous, text-based care models. Dr. Khakhkhar describes Council Health as a productized version of texting a doctor friend, where patients can message physicians and receive prescriptions, diagnostics, and record reviews. He categorizes healthcare AI into deep learning systems for pattern recognition, like sepsis detection, and LLM-based tools to aid providers, like ambient documentation. Khakhkhar believes AI can best extend the power of generalist physicians, such as by improving diagnostic accuracy with imaging aids. He emphasizes that Council Health maintains a "doctor in the loop" for safety and does not use unsupervised chatbots for acute complaints. A key insight is that 88% of initial health queries go to Google, representing the first triage decision patients make. Khakhkhar argues emergency medicine should aim to own this step in the care pathway. He cites the work of Drs. Steve Smith and Robert Herman with the OMI paradigm as a proud example of EM-driven innovation where AI now excels at subtle EKG pattern recognition. The hosts also discuss a preprint paper titled "GSM-Symbolic," which found LLM performance on math problems dropped up to 65% when irrelevant information was added, suggesting a lack of true logical reasoning. Khakhkhar uses a medical education framework—reporter, interpreter, manager—to conceptualize safely granting AI more autonomy as it proves itself. He states that solving issues like ER boarding requires rearranged financial incentives toward value-based care, not just technology. The tone is conversational, collegial, and forward-looking, blending clinical experience with entrepreneurial perspective. This episode would appeal to emergency medicine clinicians interested in digital health innovation and pragmatic AI integration. Listeners seeking a highly technical dive into AI algorithms or a debate on AI ethics might find it less focused.

As heard by us

A grounded look at emergency medicine, triage, and AI-assisted patient contact.

STAT AI keeps its attention on emergency medicine and AI, with the guest arguing that triage begins earlier than the emergency department, even in the moment patients search symptoms on Google.

Read the full review in PlayNext →

Why you'd press play

You want AI triage talk grounded in real emergency-medicine workflow, not abstract hype.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from STAT AI: An Emergency Medicine Podcast