TY - JOUR
T1 - When, where, who, what, and why? The five Ws of workflow analysis for implementing an AI decision support tool at the intensive care
AU - de Hond, Anne A.H.
AU - Vosslamber, Suzanne
AU - Lange, Sanne
AU - Engel, Friso
AU - Lindhout, Mette
AU - Noorlag, Puck
AU - Steyerberg, Ewout W.
AU - Cinà, Giovanni
AU - Arbous, M. Sesmu
N1 - Publisher Copyright:
© 2025
PY - 2025/6
Y1 - 2025/6
N2 - Background: There has been a surge in the development of clinical artificial intelligence (AI) decision support tools. Yet, few of these tools have been implemented into clinical practice. A lack of understanding of the clinical workflow may hamper implementation. Objective: This study develops and illustrates a method for the analysis of a clinical decision-making workflow at the ICU to facilitate the safe and efficient implementation of an AI decision support tool. Material & Methods: A method is proposed that uses observations to study the context in which clinical decisions occur (physical workflow), and the key factors influencing decision-making (cognitive workflow). This approach provides a comprehensive understanding of the ‘when’, ‘where’, ‘who’, ‘what’, and ‘why’ of the decision-making process. The method was applied to a use case involving the decision to discharge a patient from the ICU to the regular ward. Results: The results showed that the proposed methodology effectively provided a thorough understanding of the physical workflow, in terms of time, location, actors, and materials for receiving decision support with AI. Additionally, it identified the key factors influencing the cognitive workflow. Conclusion: We presented a method for the workflow analysis of clinical decision-making at the ICU before AI implementation. A thorough understanding of the clinical workflow is essential for implementation of an AI decision support tool. This method forms a blueprint and further validation is needed for other clinical contexts.
AB - Background: There has been a surge in the development of clinical artificial intelligence (AI) decision support tools. Yet, few of these tools have been implemented into clinical practice. A lack of understanding of the clinical workflow may hamper implementation. Objective: This study develops and illustrates a method for the analysis of a clinical decision-making workflow at the ICU to facilitate the safe and efficient implementation of an AI decision support tool. Material & Methods: A method is proposed that uses observations to study the context in which clinical decisions occur (physical workflow), and the key factors influencing decision-making (cognitive workflow). This approach provides a comprehensive understanding of the ‘when’, ‘where’, ‘who’, ‘what’, and ‘why’ of the decision-making process. The method was applied to a use case involving the decision to discharge a patient from the ICU to the regular ward. Results: The results showed that the proposed methodology effectively provided a thorough understanding of the physical workflow, in terms of time, location, actors, and materials for receiving decision support with AI. Additionally, it identified the key factors influencing the cognitive workflow. Conclusion: We presented a method for the workflow analysis of clinical decision-making at the ICU before AI implementation. A thorough understanding of the clinical workflow is essential for implementation of an AI decision support tool. This method forms a blueprint and further validation is needed for other clinical contexts.
KW - artificial intelligence
KW - critical care workflow
KW - decision support tool
KW - implementation
KW - observational study
KW - workflow analyses
UR - http://www.scopus.com/inward/record.url?scp=85218622040&partnerID=8YFLogxK
U2 - 10.1016/j.hfh.2025.100095
DO - 10.1016/j.hfh.2025.100095
M3 - Article
AN - SCOPUS:85218622040
SN - 2772-5014
VL - 7
JO - Human Factors in Healthcare
JF - Human Factors in Healthcare
M1 - 100095
ER -