Establishment Of A Visual Forewarning Model Of Enteral Feeding Intolerance For Individual Prediction In Patients With Severe Stroke

ObjectiveTo investigate the status of enteral nutrition feeding intolerance(FI)in severe stroke patients,including the incidence,duration and common types of FI in patients with severe stroke,to analysis of the influencing factors that may lead to the occurrence of FI in patients with severe stroke,to establish a line graph model to predict the FI influence in patients with severe stroke,and to carry out internal validation and clinical evaluation of the early warning model,the aim of this study is to provide an intuitive,concise and effective individual assessment tool for clinical staff to assess the risk of FI in patients with severe stroke.Methods1.The first part adopts a forward-looking Case-control study approach,according to inclusion and exclusion criteria,from September 2020 to May 2021,120 stroke patients were selected from three grade III,Class A hospital intensive care units in Shandong province.The general data and observation data of the patients were collected with the self-designed questionnaire on influencing factors of enteral feeding intolerance in patients with severe stroke.Exploring the Current Situation of FI in Severe Stroke Patients by Statistical Description.Univariate and binary logistic regression analysis methods were used to explore the influencing factors of FI in severe stroke patients.2.In the second part,using R software,the independent variables and regression coefficients were obtained by multi-factor binary Logistic regression analysis,and a line graph model was established to predict the risk of FI in patients with severe stroke,C-index,correction curve and decision-making curve were used to verify the distinction,consistency and clinical usefulness.Refer to the inclusion and exclusion criteria in Part 1,100 patients with severe stroke were collected from June 2021 to January 2022 as an early warning model to evaluate the population.Data collection is the same as the first part,through sensitivity,specificity,accuracy to evaluate the clinical predictive effect of the line chart.Results1.The incidence of enteral nutrition FI in severe stroke patients was 43.22%.The peak time of FI was 1~2 days after enteral nutrition.2.Analysis of the related factors affecting the occurrence of enteral nutrition FI in patients with severe stroke,the results showed that age(OR=1.081,95%CI:1.015-3.435,P=0.003),APACHE II score(OR=2.101,95%CI: 1.016-13.620,P=0.013),duration of bed rest(OR=1.471,95%CI:0.084-0.648,P=0.018),albumin(OR=1.088,95%CI:1.018-1.422,P=0.002),vasoactive agents(OR=3.092,95%CI:0.019-0.450,P=0.003),and bedside angle ≥30 °(OR=0.058,95%CI:0.183-0.429,P=0.001)were independent influence factors for the occurrence of enteral nutrition FI in patients with severe stroke.3.The total score is 223,the probability of FI is 50%,and the higher the score,the greater the risk of FI occurrence.The non-parametric repeated sampling method Bootstrap,C-index 0.879,is used to verify the model,and the consistency test of the calibration curve shows that the predicted probability is consistent with the actual occurrence probability.At the same time,the decision-making curve shows that the net benefit of this model is higher than that of any single index.4.This map can predict the occurrence of enteral nutrition FI in 100 patients with severe stroke,and its clinical prediction accuracy was 90.0%;sensitivity was 92.5% and specificity was 87.2%,respectively.It shows that the nomogram early warning model has high clinical predictive ability,and it can effectively predict the risk of early FI in patients with severe stroke.Conclusion1.The incidence of FI was higher in patients with severe stroke.Age,critical condition,long-term bedridden,low albumin,vasoactive drugs and the angle of head of bed < 30 ° were independent risk factors for FI.2.A preliminary line graph model was constructed to predict the risk of FI in patients with severe stroke,including age,Apache II score,duration of bed rest,albumin,vasoactive agents,and head-of-bed angle ≥30 °,through internal validation and clinical evaluation,the line graph model can accurately predict the risk of FI in patients with severe stroke,and help to accurately screen the high-risk population of FI in patients with severe stroke.