Abstract
<title>Abstract</title> <p> <bold>Background</bold> : Identifying patients at high risk for mechanical ventilation weaning failure remains a persistent challenge in critical care. The Braden Scale, a bedside functional assessment tool already routinely administered in ICUs, may capture systemic functional status relevant to weaning capacity, yet its predictive value for weaning outcomes and its incremental contribution beyond disease severity scores remain unknown. This study evaluated the association between Braden score and weaning failure in severe pneumonia patients and systematically assessed its incremental predictive value beyond APACHE II. <bold>Methods</bold> : A retrospective cohort study was conducted on 184 mechanically ventilated patients with severe pneumonia admitted to the ICU of a tertiary hospital from January 2021 to December 2025. Patients were divided into weaning success group (n = 85) and weaning failure group (n = 99). Multivariate logistic regression was used to analyze the association between Braden score and weaning failure. Four models were constructed upon a base model (age, sex, tracheostomy, vasopressors): Model A (base+Braden), Model B (fully adjusted), Model E (base+APACHE II), and Model F (base+APACHE II+Braden). Incremental value was assessed using ROC curves, decision curve analysis (DCA), net reclassification improvement (NRI), and integrated discrimination improvement (IDI). Bootstrap 1,000 iterations were used for internal validation. Subgroup analysis was performed with Benjamini-Hochberg FDR correction. <bold>Results</bold> : Braden score was independently associated with weaning failure (OR = 0.77, 95% CI 0.65–0.92, P = 0.004), with a significant dose-response relationship across risk strata (P-trend = 0.041). Braden provided significant incremental value beyond the base model (ΔAUC = 0.029, NRI = 0.389, IDI = 0.044; all P < 0.05). However, its incremental value beyond APACHE II was limited (ΔAUC = 0.009, P = 0.224), indicating substantial information overlap. Bootstrap validation confirmed model stability (optimism-corrected AUC = 0.754). Subgroup analysis with FDR correction identified tracheostomy as a significant effect modifier (OR = 0.60, FDR P = 0.034), suggesting Braden may be particularly informative in tracheostomized patients. <bold>Conclusions</bold> : The Braden Scale is independently associated with weaning failure and offers a practical bedside tool for functional status assessment. Its incremental value beyond APACHE II is limited, supporting its use as a complement rather than a replacement for disease severity scores. The tracheostomy subgroup finding warrants further investigation. <bold>Trial registration</bold> : Clinical trial number: not applicable. This was a retrospective observational cohort study, not a clinical trial. </p>