Antimicrobial prescription is a complex decision often influenced by cognitive biases that may contribute to inappropriate use and increased bacterial resistance. In the intensive care environment, where decisions are rapid and high-impact, understanding these factors is essential. The objective of this study was to assess the prevalence of cognitive biases among physicians working in intensive care units and analyze their influence on decisional accuracy in standardized clinical scenarios.

An observational, cross-sectional study was conducted with 150 physicians from different specialties working in intensive care units. Data were collected via an anonymous online form using a non-validated instrument developed by the authors based on cognitive biases established in the literature. The questionnaire contained statements on a five-point Likert scale to assess nine cognitive biases and accuracy in two standardized clinical cases. A descriptive analysis of the data was performed.

The results revealed a paradox between high subjective confidence and low objective performance. Descriptive analysis of bias prevalence (“strongly agree” or “partially agree” responses) showed that the most frequent were the IKEA effect (84.7%), hyperbolic discounting (78.7%), optimism bias (76.0%), and impact bias (71.3%). Other relevant biases included diagnostic momentum (64.0%) and availability bias (54.0%). To a lesser degree, commission bias (32.7%) and negativity bias (28.0%) were observed. Notably, anchoring bias showed the lowest agreement (16.7%). Additionally, 34.0% admitted following the dogma of completing therapies in multiples of seven days. In contrast to the high prevalence of confidence-related biases, the accuracy in choosing the antimicrobial in clinical cases was only 7.3% (11/150) in the first and 19.3% (29/150) in the second.

The high prevalence of confidence-related biases (IKEA, optimism) in contrast to low objective decisional accuracy is the central finding of this study. This suggests that antibiotic therapy is strongly influenced by subjective factors, demanding strategies that transcend conventional education. Future studies should focus on prospectively and multicentrically evaluating the impact of interventions designed to mitigate the most prevalent biases on clinical and microbiological outcomes.