Objective To investigate the risk assessment and model construction of intermittent oro-esophageal tube feeding(IOE)for stroke patients with dysphagia.Methods A total of 146 patients with stroke dysphagia who received IOE in the Second Affiliated Hospital of Wenzhou Medical University from January 2020 to December 2022 were selected as study objects.According to whether or not they were aspirated after IOE,they were divided into non-aspiration group(117 cases)and aspiration group(29 cases).Clinical data of patients were collected.Multivariate Logistic regression was used to analyze the risk factors of aspiration in patients,and a nomogram prediction model was established.The C-index was used for differentiation evaluation,and the calibration curve was used to detect the nomogram fit degree.Results There were statistically significant differences in stroke history,aspiration history,stroke site,spontaneous cough,Wada drinking water test,National Institutes of Health stroke scale(NIHSS)score,catheter placement position,and catheter depth between two groups(P<0.05).Multivariate Logistic regression analysis showed that stroke history,aspiration history,brain stem location of stroke,grade 4 to 5 of Wada drinking water test,high NIHSS score,lateral position tube,and low catheter depth were all risk factors for aspiration in IOE patients with dysphagia of stroke(P<0.05),spontaneous cough was a protective factor for aspiration in IOE patients with dysphagia of stroke(P<0.05).The C-index of nomogram prediction model to judge aspiration risk was 0.705(95%CI:0.591-0.857),and Hosmer-Lemeshow test model fit well.The receiver operating characteristic curve results showed that the area under the curve of nomogram prediction model was 0.756,the sensitivity was 75.42%,and the specificity was 77.09%.Conclusion Stroke history,aspiration history,stroke site,spontaneous cough,Wada drinking water test,NIHSS score,placement position,and placement depth were all influential factors for aspiration in IOE patients with stroke dysphagia.The model was successfully constructed,and the constructed model had good differentiation and fitting degree.