Human touch with a(I) helping hand

Health systems continue to seek greater efficiency in care delivery and operations—with AI at the core of many solutions. But as AI-powered offerings become increasingly patient-facing (with regulation trying to keep pace), there’s still room for models that center humans in asynchronous care and communication. For example, one health system reduced burnout among primary care providers by repositioning mid-level specialists as “inboxologists” to triage and reply to patient messages. Another health system’s pilot offers lessons for integrating AI—when piloting a large language model to draft replies to patient messages, nurses viewed the support most favorably, suggesting that tuning to clinical roles may enhance model utility. These findings demonstrate that realized efficiencies will come from multiple sources: care models that position staff for top-of-license practice, and smart application of purpose-built AI.

Headlines

Epic’s AI Tool Didn’t Fare Well In Yale Test Of Early-Warning Algorithms
Rock Health port co AgileMD’s eCART rises above STAT