The Development and Feasibility of a Triage System for Use in Primary Care i

The Icelandic healthcare system faces challenges common to many Western countries, e.g., staff shortages and high workloads, with negative impacts on patient care. Long waiting times, overcrowded clinics, and overworked healthcare professionals contribute to diminished health care quality. To address these issues, we propose a system for patients and providers to alleviate pressure on primary care clinics. The system consists of an AI model for triage, a Questionnaire User Interface (QUI), and a Consultation User Interface (CUI). The QUI aims to reduce unnecessary in-person consultations by allowing patients to answer questions related to their well-being and medical history, similar to those posed during health center visits. The AI triage model assesses risk levels for specific disease groups by analyzing patient responses to the questions and determines the necessity of an in-person visit. The CUI facilitates the diagnosis and scoring of patients when further examination is required. The obtained risk scores inform treatment advice, optimizing healthcare professionals' decision-making. A feasibility study showed that the system could be integrated with the clinical workflow, such as automatically filling clinical text notes, saving doctors valuable time. By taking an user centered design approach and using AI for symptom scoring and patient classification, the system has the potential to enhance health care provider efficiency and improve consistency in patient care.