Development and Validation of a Nomogram for Predicting Acute Skin Failure in Critically Ill Patients: A Prospective Study.

PURPOSE: Acute skin failure is a common skin injury in critically ill patients. This study aimed to develop a nomogram model to assess the risk of acute skin failure in critically ill patients.

METHODS

This study enrolled 230 ICU patients from a class III hospital in Nanjing for modeling development from August 2024 to February 2025, while 123 ICU patients were enrolled for external validation from March 2025 to May 2025. Lasso regression and logistic regression were used to analyze the independent prediction factors of acute skin failure in critically ill patients. The nomogram model was constructed with R 4.4.1 software, and the web calculator was developed. The model's predictive ability was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC), calibration curves, and clinical decision curves. Finally, the Shapley additive explanation method (SHAP) was employed to improve the model's interpretability.

RESULTS

The results of the study showed that peripheral perfusion index (OR=0.523, p<0.001), skin mottling score (OR=2.239, p=0.007), multiple organ dysfunction syndrome (OR=5.897, p =0.002), lactic acid (OR=1.754, p=0.002) and albumin (OR=0.828, p=0.007) were independent prediction factors for acute skin failure. The area under the curve (AUC) for the modeling set of this model is 0.959 (95% CI: 0.935 ~ 0.983), and for the validation set is 0.930 (95% CI: 0.875 ~ 0.983). The calibration curves and clinical decision curves demonstrate good concordance and the clinical applicability of the model.

CONCLUSION

The nomogram model developed in this study for predicting ASF risk in critically ill patients demonstrates potentially useful predictive performance and assists caregivers in the early identification of high-risk ASF groups.