Designing for Nurse-AI Collaboration in Triage
MetadataShow full item record
- Master theses 
The Local Emergency Medical Communication Centers (LEMCs) play a crucial role in the Norwegian healthcare system by receiving calls for immediate medical assistance. Registered nurses operate the phone calls, and their task is to assess the situation and triage the caller into appropriate triage levels indicating when and how help should be provided. Telephone triage poses challenges due to the limitations of audio communication, time sensitivity, and complex decision-making. Additionally, nurses often face the burden of managing clinical tools across multiple interfaces. This thesis explored how to design a system to support nurses in telephone triage and how we can facilitate nurse-AI collaboration in the process. A Research through Design (RtD) methodology was employed, and an iterative design approach was utilized. The research investigated the design aspects of AI-based suggestions and the use of natural language when creating semi-structured documentation. Four prototype iterations were developed throughout the study, and researchers from RE-AIMED and telephone operators conducted evaluations of the prototypes. Designing a tool for telephone triage requires understanding the user's needs and workflow. It is, therefore, crucial to involve telephone operators in the design process. The prototype demonstrated how we could design for incorporating AI in the triage process, and this thesis explores the various considerations when designing for nurse-AI collaboration. One notable finding was the importance of enabling documentation in natural language, as relying solely on structured documentation may fail to capture the caller's specific situation. Additionally, it is important to design a system that facilitates documentation of patient-initiated information and questions initiated by the nurses or the system.