BALTIMORE, July 28, 2026 — Emergency departments are searching for ways to shorten patient waits without adding beds, staff, or physical space. New work in Management Science points to a counterintuitive lever: route the right patients to the right part of the care process, faster and more consistently.
The study, “Vertical Patient Streaming in Emergency Departments,” reports that a structured, data-driven triage protocol can cut total emergency department length of stay by about 11 minutes (4.2%) while maintaining safety. The approach targets “vertical processing,” a workflow that routes eligible patients to a seated treatment pathway instead of occupying scarce bed capacity.
Many hospitals already have seated areas, but who gets sent there is often decided informally, creating variation. The researchers replaced guesswork with a standardized decision tool that clinicians can apply during triage—before beds are assigned and bottlenecks begin to form.
To build the model, the team analyzed nearly 50,000 emergency visits from Mayo Clinic Arizona. They trained a machine learning model using only information available at triage—such as Emergency Severity Index (ESI) level and presenting complaint—to predict whether a patient would ultimately require a traditional emergency department bed.
Next, the team coupled predictions with mathematical patient-flow optimization to determine where routing decisions should be applied to maximize system efficiency. Importantly, the resulting policy was converted into an implementable decision tree, avoiding the need for new software or IT integration.
The protocol was tested in a 13-week prospective field trial with 11,015 patients at Mayo Clinic Arizona’s new emergency department. Compared with usual routing, the intervention reduced time from arrival to clinical disposition by roughly 8 minutes (4.5%).
Quality signals remained stable: there was no increase in 72-hour return visits, suggesting that shifting patients to a seated pathway did not degrade care outcomes. The researchers emphasize that the logic relies only on routinely collected triage inputs, including whether the department is operating over capacity.
The authors estimate that a medium-sized emergency department handling around 40,000 visits annually could recover thousands of “bed-hours” each year—potentially enabling care for thousands of additional patients—without expanding facilities or staffing. Their message is clear: analytics-led operational redesign can unlock capacity inside existing walls.
Read the study via the journal link: https://pubsonline.informs.org/doi/epdf/10.1287/mnsc.2024.07517.
Subject of Research: Vertical patient streaming / emergency department patient routing and triage optimization
Article Title: Vertical Patient Streaming in Emergency Departments
News Publication Date: 28-Jul-2026
Web References: https://pubsonline.informs.org/doi/epdf/10.1287/mnsc.2024.07517
References: doi:10.1287/mnsc.2024.07517
Image Credits: Not provided
Keywords: emergency department overcrowding; patient triage; vertical processing; machine learning; patient flow optimization; operational efficiency; healthcare analytics
Tags: data-driven triage protocolsemergency department flow managementemergency department patient triagehospital operational efficiency improvementsimpact of structured triage on patient safetymachine learning in emergency careoptimizing hospital resource utilizationpatient routing to minimize bed occupancyreducing ER wait times with workflow optimizationseated treatment pathways in hospitalsstandardized decision tools for triagevertical patient streaming
