Boston Children's Uses Etiometry's Continuous Risk Analytics to Guide Extubation and Vasoactive Weaning in ICUs

Boston Children's Hospital has made Etiometry's AI-driven clinical intelligence platform, which uses large language models, a decision support tool across its critical care team, Healthcare IT News reported in March 2026. Bedside clinicians use its continuous risk indicators to cross-check extubation readiness and time the weaning of vasoactive drugs, and charge nurses use unit-level screens to see which patients need attention. "Our experience, along with broader reports from other adopters, shows that targeted, data-driven decision making is associated with reductions in ICU length of stay," said Dr. Peter Laussen, executive vice president of health affairs. The hospital's critical care teams had no efficient way to store, curate and analyze high-frequency physiologic signals and bring them to the bedside. Decisions relied on varied heuristics, overnight vital signs were recorded only hourly, and in one case a delayed decision led to patient harm. Etiometry captures continuous, high-fidelity physiologic data so teams can replay events down to a five-second window, and aggregates signals into trajectories rather than dictating care. In ventilation weaning its findings serve as a secondary measure alongside clinical judgment, and for de-escalation the team tracks coronary perfusion pressure continuously to decide when to wean vasoactive infusions. The rollout also changed how the team runs rounds and trains staff. A multi-center study led by Boston Children's found that elevated Etiometry indices, IDO2 and IVCO2, before extubation were associated with higher odds of extubation failure after congenital cardiac surgery, and the hospital says incorporating these continuous risk signals has lowered the odds of failure nearly twofold in its published experience. The hospital notes it still struggles to link every bedside decision systematically to the underlying data; return-on-investment figures cited in the article come from Etiometry's modeling rather than measured hospital results.