Healthcare AI Daily Briefing
New to OpenRounds? Get the free daily email or see how we work.
Clinical Practice
Interview: Staffing shortage spurred Baptist Health AI bot, halting help‑desk hires.
Podcast from Lifers (Second Opinion)
The interview describes Baptist Health’s front‑facing AI bot that automates password resets and routine account‑access calls, deflecting enough volume that the system, which had struggled to hire help‑desk staff, has not brought on a new agent in over eighteen months. [4]
Diagnostics & Pathology
AI echocardiography detects epicardial fat, AUROC up to 0.93.
Preprint
PanAdipo, a deep‑learning model trained on 28,797 echocardiograms from Yale‑New Haven, detected prominent epicardial fat in a held‑out set of 5,167 patients. Its scores correlated more strongly with CT‑derived epicardial fat volume (ρ=0.76) than with BMI (ρ=0.41), indicating residual cardiometabolic risk. [2]
Diagnostics & Pathology
EMR model predicts BSI mortality (AUC 0.83); microbial data add no benefit.
Preprint
In a retrospective cohort of 35,746 bloodstream infections from Calgary, Canada, an EMR‑based machine‑learning model achieved an AUC of 0.83 for predicting 30‑day mortality; adding more than 70,000 microbial genes and proteins did not improve this performance. Microbial factors co‑segregate with specific patient cohorts, which is why they are not independent predictors of mortality. [1]
Clinical Practice
AI-generated podcasts achieve non‑inferior student satisfaction versus clinician recordings.
Journal article
One hundred and seventy‑eight medical students were randomised to receive either clinician‑produced or AI‑generated podcasts on type 1 and type 2 diabetes; mean satisfaction scores did not differ significantly between formats, indicating that AI‑generated podcasts achieved non‑inferior satisfaction to clinician‑recorded podcasts. [3]
Sources
- Microbial traits have minimal power for predicting death from bloodstream infection · medRxiv Original source
- AI-Enabled Echocardiography Identifies Prominent Epicardial Adipose Tissue Associated With an Adverse Cardiometabolic Profile · medRxiv Original source
- Teaching on demand: Exploring the role of AI in medical education podcasts. · Medical teacher Original source
- Healthcare runs a 1980s model in 2026 | Aaron Miri, Baptist Health · Lifers (Second Opinion) Original source
Daily healthcare AI brief
Get the free healthcare AI briefing
One concise briefing built from reporting, research, policy, blogs, videos, and full podcast transcripts.
Free. Sign in to subscribe. Unsubscribe anytime. Already subscribed? Sign in to save and personalize.