Construction of a risk prediction model for medical adhesive related skin injury in infusion ports during the treatment period of lung cancer patients
WU Shanshan
LIU Kouying
TANG Ting
Abstract：Objective:To analyze the risk factors for medical adhesive related skin injury(MARSI)in infusion ports during the treatment period of lung cancer patients.And to establish a risk prediction model,so as to provide reference for clinical nursing intervention.Methods:A retrospective collection of 650 patients with implantable chest wall ports who were hospitalized in the respiratory and critical care medicine department of a tertiary grade A general hospital from January 2023 to April 2024 was conducted.Logistic regression model,decision tree classification regression(CART)model,and random forest models were used to establish risk prediction models for medical adhesive related skin injury in infusion ports during the treatment period of lung cancer patients.The accuracy,sensitivity,specificity,positive predictive value,negative predictive value,Kappa coefficient,and area under the receiver operating characteristic(ROC)curve(AUC)of the three models were compared to evaluate their performance.Results:The accuracy of Logistic regression model,decision tree CART model,and random forest models were 84%,86%,and 86%,respectively.The specificity were 97%,98%,and 97%.The sensitivity were 54%,59%,and 61%.The positive predictive values were 54%,59%,and 61%.The negative predictive values were 97%,98%,and 97%.The Kappa values were 0.57,0.63,and 0.64.The AUC were 0.83,0.87,and 0.86.There were statistically significant in the AUC differences among Logistic regression model,decision tree CART model,and random forest(P<0.05).Skin toxicity was a common predictor for three models.Conclusions:Compared with Logistic regression model,decision tree CART model and random forest model had better performance in constructing a risk prediction model for in infusion ports during the treatment period of lung cancer patients,it could provide reference for clinical nurses to predict the risk of medical adhesive related skin injury in infusion ports during the treatment period of lung cancer patients.
Keywords：infusion portmedical adhesive related skin injuriespredictive modelLogistic regressiondecision tree CARTrandom forest method
Publication Date：2025-08-10
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：10( 2525-2534 )