Development and Validation of a Predictive Model for Oral Mucosal Pressure Injury Risk in ICU Patients with Endotracheal Intubation

Limei CaiYijing LiYonggang LiuGuo MaQinfang ZhangXiaoxi LiNa Li^*

• First Affiliated Hospital of Kunming Medical University, Kunming, China

Objective: To identify risk factors for oral mucosal pressure injury (OMPI) in intensive care unit (ICU) patients undergoing orotracheal intubation and to develop and validate a risk prediction nomogram based on logistic regression analysis. Methods: Relevant risk factors for OMPI were identified through a combination of literature review and expert interviews. A total of 426 intubated ICU patients admitted to a tertiary hospital in Yunnan Province between May and December 2024 were included in the model group. Variables with P < 0.10 in univariate analysis were further entered into multivariate logistic regression to identify independent risk factors for OMPI and construct a predictive nomogram. Missing data were addressed using multiple imputation, and potential confounders such as age, BMI, and disease severity were adjusted for in the multivariable analysis. Model performance was evaluated by the area under the ROC curve (AUC), calibration plots, and decision curve analysis (DCA), and internally validated using bootstrap resampling. An external validation cohort of 152 patients from January to March 2025 was used to assess the model's predictive performance. All analyses were performed using SPSS version 27.0 and R version 4.3.2, with a two-tailed P value < 0.05 considered statistically significant. Results: Duration of intubation, use of dental pads, Richmond Agitation-Sedation Scale (RASS) score, Brachycephalic Obstructive Airway Syndrome (BOAS) score, and platelet count were identified as independent risk factors for OMPI (P < 0.01). The model showed good discriminative ability with an AUC of 0.888 [95% CI: 0.849–0.926]. The calibration curve demonstrated strong agreement between predicted and observed outcomes, and the Hosmer–Lemeshow test indicated good calibration (χ² = 3.95, P = 0.861). DCA showed net clinical benefit within a 3%–100% risk threshold. External validation yielded an AUC of 0.854, sensitivity of 86.5%, specificity of 73.0%, and overall predictive accuracy of 83.7%. Conclusion: The validated nomogram demonstrated good discrimination, calibration, and clinical utility, offering a reliable tool for early identification of high-risk ICU patients and for guiding personalized interventions. Nevertheless, as this was a single-center study, further multicenter validation is needed to confirm its generalizability.