Automated tagging of intraoperative physiologic events

This study evaluates a novel application of the OR Black Box® for automatically tagging intraoperative physiologic disturbances, specifically hypotension, hypoxia, and hypothermia, and examines whether these real-time events predict postoperative complications. In a retrospective cohort of 2,875 inpatient surgical cases, 20% experienced at least one tagged physiologic event. Incorporating these automated tags into regression models consistently improved prediction performance across outcomes, including surgical site infection, patient safety indicator events, readmission, return to the operating room, length of stay, and mortality. Notably, hypothermia and hypoxia predicted return to the OR, while hypotension was associated with longer length of stay and higher PSI event rates.

The findings highlight that intraoperative physiologic disturbances carry measurable consequences for postoperative outcomes. Automated tagging offers a scalable method for capturing granular intraoperative data and integrating it into risk-stratification models. The study demonstrates an important proof of concept: real-time physiologic metadata can meaningfully enhance predictive analytics and may eventually support more personalized postoperative care and quality improvement efforts within surgical systems.