Diaphragm ultrasound beyond numbers: A call for Bayesian reasoning in critical care.

Diaphragm ultrasound (DUS) has become a widely available tool in critical care for assessing respiratory muscle function, monitoring ventilator-induced diaphragm dysfunction (VIDD), and supporting weaning prediction. However, its expanding use as a routine or screening examination, including in clinical scenarios where the pre-test probability of clinically relevant diaphragm dysfunction is low, raises concerns about diagnostic overuse. This narrative opinion article argues that DUS should be understood through the lens of Bayesian reasoning applied to clinical decision-making. The framework is intended to strengthen clinical reasoning rather than to assert superiority over alternative diagnostic strategies. Without explicit estimation of pre-test probability and prior definition of decision thresholds, the positive predictive value of DUS decreases, favoring false-positive findings, physiological overinterpretation, and potentially inappropriate clinical decisions. The available evidence indicates substantial heterogeneity in protocols, populations, and cut-off values, limiting the generalizability of findings and reinforcing the need for clinical contextualization. We propose a conceptual repositioning of DUS as a confirmatory and integrative tool, grounded in genuine clinical uncertainty, explicit Bayesian reasoning, and value-based care.