Emergency departments (EDs) are often the first point of contact for patients in acute distress, yet they remain among the most overburdened areas in healthcare. Overcrowding, long wait times and clinician burnout are endemic challenges that compromise patient outcomes and staff morale. With artificial intelligence technologies gaining traction across clinical settings, UMass Memorial Health is demonstrating how AI-based decision support tools can help ease these pressures, particularly by assisting nurses during the critical triage process.
Streamlining Triage with AI Support
UMass Memorial Health has deployed a machine-learning platform, KATE AI, across its emergency departments to support nurses responsible for triage. The triage process is foundational to emergency care, as it determines which patients require immediate attention and which can safely wait. KATE AI analyses intake form data and electronic health records (EHRs) to generate an Emergency Severity Index (ESI) score, offering a more nuanced picture of a patient’s condition. The tool enhances the triage process by identifying clinical subtleties, such as underlying chronic conditions that might otherwise go unnoticed during the initial patient interview. For example, a fever in a patient with sickle cell disease carries significantly more clinical risk than in someone without such a history. By surfacing such correlations, the tool improves both safety and efficiency. Additionally, the AI platform offers early sepsis detection capabilities, alerting nurses to potential red flags and supporting timely interventions.
Must Read: Adapting to Digital Triage: Nurses’ Experiences in Primary Healthcare
Implementing the System and Addressing Resistance
The rollout of KATE AI began in early 2023 with pilot implementations at the University and Memorial campuses of UMass Memorial Medical Center. The urgency created by COVID-19-related capacity challenges prompted the health system to seek technological solutions. One major advantage of the KATE AI system was that it integrated with existing workflows, avoiding the need for additional training or infrastructure, which facilitated adoption. Nonetheless, the introduction of AI in clinical environments raised concerns among some experienced nurses, particularly around liability and decision-making authority. These concerns were proactively addressed through engagement with nursing associations and union representatives. Nurses were reassured that they retained full responsibility for patient care decisions and would not be penalised for disagreeing with AI suggestions. Over time, confidence in the tool grew, supported by real-world examples in which AI insights proved prescient. Experiences like these helped foster greater trust in the system and demonstrated its potential to complement, rather than replace, clinical judgment.
Measurable Improvements in ED Operations
The AI tool’s implementation has yielded a range of operational benefits at UMass Memorial Health. Nurses reported that clinical documentation became more accurate and less burdensome, as the tool helped identify and correct data entry errors prior to physician evaluation. Improved accuracy in ESI scoring has also been noted, with a marked increase from initial levels around 55% to current rates approaching 70%. This improvement has direct implications for patient prioritisation and outcomes. Additionally, the tool contributed to better performance on key ED metrics, such as reduced left-without-being-seen rates and fewer risk reports tied to triage. Beyond immediate clinical gains, the system also supports broader analytical efforts. By aggregating structured and unstructured data from various sources, including triage notes and discharge summaries, KATE AI provides deeper insight into complex patient populations, such as those affected by drug overdoses or cardiac events. This ability to unify and interpret disparate data sets enhances the health system’s capacity for targeted interventions and long-term planning.
UMass Memorial Health’s use of AI-powered clinical decision support in its emergency departments illustrates the technology’s growing role in alleviating systemic stress. By assisting triage nurses in identifying high-risk patients, improving documentation accuracy and generating valuable operational insights, KATE AI offers a practical response to the increasing demands on emergency services. The project’s success also underscores the importance of thoughtful implementation, including workforce engagement and system integration. With emergency departments facing mounting pressures, AI-based tools like this may become essential to maintaining quality, safety and efficiency in urgent care delivery.
Source: TechTarget
Image Credit: iStock
