a whitepaper by Breyden Taylor - 01/14/2025 - in association with Skyward Prompted
special thanks to Reuven Cohen for giving me confidence enough to write this
Automated or semi-automated healthcare triage systems are playing an increasingly critical role in remote patient management, telemedicine, and emergency call centers. Their core objectives—ensuring patient safety, optimal resource allocation, and standardized decision-making—are intrinsically tied to the underlying logic. Traditional triage protocols often employ heuristic or flowchart-style decision trees that can be difficult to maintain and lack flexibility in handling complex, evolving medical knowledge.
This whitepaper presents a predicate logic–based framework, enriched by symbolic mathematics, to provide a more rigorous, transparent, and adaptable approach to triage. By formalizing symptoms, conditions, and recommended dispositions in well-defined predicates and rules, our system overcomes many limitations of typical rule-based tools. The paper will describe core features such as negative-constraint overrule (red-flag handling), mandatory escalation for critical conditions, edge-case logic for unusual or outlier presentations, and a priority-based mechanism to differentiate levels of care. We also discuss extensions for fuzzy or probabilistic modeling to address real-world uncertainty.
The outcome is a triage model that is more interpretable, scalable, and clinically aligned. It can help reduce under-triage errors, provide clear traceability for audits, and integrate seamlessly into existing electronic health record (EHR) infrastructures.
Healthcare triage typically involves an initial assessment of a patient’s condition to determine urgency and appropriate care pathways. While many organizations still rely on human-driven phone triage or in-person screening, these manual processes can be time-consuming, subject to individual variance, and prone to misinterpretation of symptoms. Automated or AI-enhanced triage protocols aim to standardize care while ensuring patients with severe conditions are identified and managed promptly.
Most current triage systems use simple decision trees or if-then-else statements that can quickly become unwieldy as the number of conditions and exceptions grows. By contrast, a predicate logic–based approach:
This document details the conceptual design and practical implications of a logic-driven triage system. It provides technical underpinnings of predicate logic, symbolic maths, and how they fit into an end-to-end clinical workflow. We also cover extensions for uncertainty, describe implementation considerations, and discuss benefits and limitations of our approach.