Abstract
Algorithmic triage tools are increasingly proposed to optimize intensive care unit (ICU) allocation during critical resource shortages. However, their ethical implications and impact on public trust remain poorly understood. This multi-center, scenario-based survey evaluated the attitudes of 1,240 healthcare providers across five European countries (France, Germany, Italy, Spain, and the Netherlands) regarding the use of machine learning algorithms for ICU triage during simulated pandemic-induced bed shortages. Participants evaluated clinical scenarios contrasting algorithmic recommendations with traditional clinician-led decision-making. Our findings indicate that while healthcare providers appreciate the potential of algorithms to reduce cognitive load and bias (68.2%), a significant majority (74.5%) expressed severe concerns regarding accountability, transparency, and the potential to exacerbate systemic health inequities. Furthermore, 62.1% of respondents believed that relying heavily on algorithmic triage would undermine public trust in the healthcare system. Trust was positively correlated with the inclusion of "explainable AI" (XAI) features and human-in-the-loop oversight. We conclude that while algorithmic tools offer valuable decision support, their implementation must be guided by robust ethical frameworks, regulatory oversight, and clinician-led final decision-making to preserve public trust and ensure equitable patient outcomes.
