Background: In critical care settings, especially during surges such as the COVID-19 pandemic, physicians are often required to make rapid triage decisions under resource constraints. While clinical indicators should ideally guide these decisions, emerging research indicates that non-clinical factors—such as a patient’s race or gender—may inadvertently affect judgment. Aim: This study aims to validate clinical profiles using a decision-making tool and examine the predictive role of clinical and social factors in intensive care unit (ICU) admission decisions under contingency conditions, such as during the COVID-19 pandemic. Methods: A total of 432 ICU admission decisions (trials) were collected from a simulated task in which nine ICU physicians evaluated 48 fictional patient profiles under conditions of limited bed availability. Each participant reviewed all profiles and selected half of the profiles for admission. Each profile included six clinical criteria (e.g., prognosis, comorbidities, and respiratory failure severity) and two non-clinical features (i.e., gender and race). The trial served as the unit of analysis. Multilevel logistic regressions assessed the predictive power of clinical and social variables on acceptance decisions, omission errors (rejecting qualified candidates), and false alarms (accepting less qualified candidates). Results: The findings demonstrate that participants relied primarily on clinical information: high-scoring profiles were admitted more often (67.4% and 52.5%) than lower-scoring ones (37.5% and 13.5%) (p<0.001). However, social factors also shaped decisions. Male candidates were more likely to be admitted than females (b = −0.51; t = −4.35; p<0.001; 95% confidence interval [CI] = [−0.75, −0.27]), and Black candidates were admitted more than White candidates (b = −0.52; t = −3.34; p<0.001; 95% CI = [−0.85–−0.19]), even when less qualified, thereby suggesting possible overcorrection. Conclusion: Although clinical criteria primarily guided ICU admission decisions, social characteristics also subtly influenced outcomes. Together, these findings validate the novel decision-making paradigm as a valuable tool for assessing both clinical accuracy and the presence of social bias in triage contexts. They also provide empirical evidence that, under pressure and uncertainty, healthcare professionals may be susceptible to the influence of social cues. Relevance for patient: This study explores how critical care physicians decide who receives life-saving treatment in ICUs, especially during times of crisis when medical resources are limited. By simulating real-world triage situations, the research shows that even experienced clinicians may be influenced by non-clinical factors—such as a patient’s race or gender—despite their intention to prioritize clinical indicators. These findings highlight the need for increased awareness and targeted training to reduce unconscious bias in clinical decision-making.