Infinitus launches AI agent triage risk detection system - Fierce Healthcare

Infinitus, an AI communications company, has launched its new risk-detection system called Clinical Escalations, aimed at ensuring AI agents correctly triage patients in healthcare settings. The system addresses the problem of under-triage, where a patient's condition might be assigned a lower urgency level than clinically appropriate, especially when patients might not recognize the severity of their own symptoms. Clinical Escalations functions as a background observer, continuously monitoring every conversational turn between patients and AI agents. It uses the industry-standard Schmidt Thompson Call Prioritization Index (CPI) framework to assign a severity score, and can reclassify administrative inquiries into emergent clinical situations in real-time. This dynamic assessment allows for appropriate escalation, such as a warm transfer or a 911 prompt, if a critical threshold is crossed. The system has been uniquely trained on real-world patient conversations, accounting for variations in literacy, emotional state, verbosity, and clarity, to reliably triage situations regardless of how patients describe their symptoms. This robust training helps identify high-acuity cases effectively. Infinitus reports that Clinical Escalations has achieved a 0% under-triage rate in real-world deployments, meaning it did not miss any patient emergencies during testing. Every exchange is also logged for human review, providing transparency and a record of escalation decisions. The solution is available to Infinitus customers, including healthcare providers, health plans, and pharmaceutical patient support programs, to manage large volumes of patient-facing communications safely and efficiently.