Automating Aspiration Precautions to Reduce Latent Safety Risk in the Emergency Department

Aspiration events in hospitals are preventable yet difficult to systematically identify and mitigate. In high-risk environments such as the emergency department (ED), fragmented documentation and high clinician cognitive load create latent safety risks. Traditional strategies such as education or interruptive alerts have shown limited effectiveness.

A multisite quality improvement initiative across nine EDs within a large northeastern health system from July 2022 through May 2025 was conducted on timely implementation of aspiration precaution and "nothing by mouth" (NPO) orders. The intervention progressed through three phases: an initial interruptive pop-up alert, a revised interruptive alert, and an automated standing order. Primary outcomes were the proportion of ED encounters in which aspiration precautions and NPO orders were placed and time from ED arrival to order placement.

During the preintervention period, aspiration precautions and NPO orders were placed in 37 of 550,621 ED encounters (0.007%), with median time to order placement of 597 minutes (interquartile range [IQR] 346-1,973) for NPO orders and 704 minutes (IQR 479-2,115) for aspiration precautions. During the initial and revised interruptive alert phases, orders were placed in 438 of 54,407 encounters (0.81%) and 918 of 182,434 encounters (0.50%), respectively. After implementation of the automated standing order system, orders were placed in 16,393 of 559,744 encounters (2.93%), and median time decreased to 24 minutes (IQR 11-53) for both NPO and aspiration precautions. No aspiration-related ED safety events were reported during the one-year post-automation period.

An automated, standing order workflow was associated with decreased duration of latent aspiration risk. This approach transformed a low-reliability, human-dependent process into a high-reliability system-level safeguard and offers a scalable model for addressing safety risks in emergency care.