A Path Towards Human-AI Decision-Making in Sepsis Care through Human-Centered Systems-Based Design Approach

Sepsis is a critical condition when the body's response to infection triggers widespread inflammation, leading to severe health complications. Despite advancements in medical care through AI-based prediction models and clinical decision support systems, sepsis remains a leading cause of death in hospitals, highlighting the urgent need for improved diagnostic methods, treatment protocols, and public awareness to prevent this severe condition effectively. Further, the ethical implications and risks associated with AI implementation necessitate careful consideration of diversity, equity, and inclusion in designing and deploying human-AI decision support. To consider such challenges, we propose a human-centered systems-based (HCSB) design approach to AI implementation that prioritizes user needs systematically and ensures fair and equitable treatment across diverse patient populations, addressing disparities that may arise from socio-economic factors, geographic location, or cultural differences. Key considerations include the development of inclusive datasets that reflect diverse patient demographics, designing transparent and interpretable algorithms, and establishing protocols for continuous monitoring and evaluation to detect and mitigate biases. Moreover, collaboration among interdisciplinary teams, including healthcare professionals, AI engineers, ethicists, and community representatives, is essential to embedding an HCSB design approach throughout the decision support lifecycle.