Development and validation of a nomogram to predict peristomal dermatitis for patients who underwent ileostomy

Abstract

Abstract Background The incidence of ostomy procedures in China is on the rise, with an estimated annual count of around 100,000. Peristomal dermatitis, a prevalent and consequential complication, manifests as inflammation and discomfort surrounding the stoma. This condition significantly impacts surgical recuperation, overall well-being, and may give rise to psychological and physical complications. Timely intervention is imperative, necessitating the identification of risk factors and the development of predictive models. Objectives We aimed to establish a predictive nomogram specifically for peristomal dermatitis after ileostomy. Methods This study analyzed 459 ileostomy patients to identify the risk factors for peristomal dermatitis. A predictive nomogram was constructed based on these factors using a training cohort, which was then validated in an independent validation cohort. The performance of the nomogram was assessed through ROC curves and calibration curves, while its clinical utility was evaluated through decision curve analysis (DCA). Results Six parameters were determined as the risk factors of peristomal dermatitis. Considering the residence may not be universally applicable, we further excluded residence to build the prediction model without reducing discrimination ability. Eventually, diabetes, ileostomy positioning, stoma care lecture, sex, and depressed skin at the stoma were selected for nomogram construction. The calibration curve demonstrated excellent calibration (Hosmer-Lemeshow Goodness-of-fit (GOF) test, P = 0.976), and the nomogram achieved an AUC of 0.672 in the training cohort. The decision curve analysis showed potential clinical utility. These results were consistent with the testing cohort. Conclusion Our study developed a nomogram based on four parameters that demonstrated excellent discriminating and calibration for predicting peristomal dermatitis. These findings suggest the potential clinical application of the nomogram.