| Objective This study aims to clarify the risk factors of VTE in critically ill patients,compare the predictive value of commonly used VTE risk assessment models for the formation of VTE in critically ill patients,and establish an individualized VTE prediction model for critically ill patients.Methods(1)The risk factors of VTE in critical patients were preliminarily summarized by literature review and study group discussion to form a questionnaire.(2)Through the retrospective case-control study method,a total of 163 patients with VTE were included as the case group by reviewing the patients admitted to Central ICU and EICU from The Second Affiliated Hospital of Dalian Medical University from September 2020 to September 2022 who met the inclusion and exclusion criteria,with an incidence of 9%.At the same time,163 patients without VTE who met the inclusion and exclusion criteria in the same department at the same time(±15 days)were selected as the control group.Univariate and multivariate Logistic regression were used to analyze the risk factors of VTE and independent risk factors of VTE in critically ill patients.(3)Hosmer-Lemeshow test was used to judge the goodness of fit of the model,and the risk factors were evaluated by single-factor analysis.Logit(P)value was obtained to predict the evaluation efficiency of the model.Finally,the other prediction models included in the study were compared through the area,sensitivity,and specificity under the ROC curve to verify the prediction value of this prediction model.Results(1)Compared with the non-VTE group,the length of hospital stay,history of with the largest AUC value was 0.640,its sensitivity was 76.7%,and its specificity was4 4.6 %,indicating that the international venous thrombosis evaluation model had the highest predictive value for the study object.(3)In univariate analysis,P values less than 0.05 were used as independent variables(length of stay: ≥ 13 days was 1,< 13 days was 0;Cardiac insufficiency: Yes = 1,no =0;Hypoproteinemia: yes = 1,no =0;Sedative:yes = 1,no =0;Analgesic: yes = 1,no =0;Dehydrating agent: Yes = 1,no =0;Vasoactive drugs: Yes = 1,no =0;Hormone: Yes = 1,no =0;D-dimer at admission: ≥2.26μg/ml was1,< 2.26μg/ml was 0;Percentage of neutrophils at admission: ≥ 89.9% was 1,< 89.9%was 0;International venous Thrombosis Score Model: ≥3 was 1,less than 3 was 0),the occurrence of VTE in critically ill patients was used as the dependent variable(yes = 1,no = 0).According to the above binary logistic analysis,the joint prediction variable PRE was obtained,and the prediction model of venous thromboembolism after critical care was constructed.Hosmer-Lemeshow test was used to analyze the fitting degree of the model,and the results showed that P=0.670,> 0.05,indicating that the model had a good fitting degree,and there was no significant difference between the predicted value and the actual value,which could be used to predict the occurrence of VTE after critically ill surgery.The prediction efficiency was evaluated based on the critical value of Logit(P)0.203.When Logit(P)≥0.632,the risk of VTE was high.(4)There were 163 patients with actual thrombosis,145 patients predicted by the model,and the sensitivity was 75.1%.There were 163 patients without thrombosis,115 patients predicted by the model,and the specificity was 86.5%.The total accuracy of the model prediction was79.8%.VTE,cardiac insufficiency,hypoproteinemia,sedatives,analgesics,dehydrators,hormones,braking,percentage of neutrophil granulocyte and D-dimer on admission in the VTE group were all < 0.05,indicating statistical significance.(2)Independent risk factors for VTE in critically ill patients were determined by multivariate regression analysis,including length of stay,cardiac insufficiency,and dehydrating drugs,all with P values < 0.05.ROC curve analysis was carried out on the present 9 evaluation models,and it was concluded that the international venous thrombosis evaluation modelConclusion Although the prediction model established in this study has high accuracy in predicting the occurrence of VTE in critically ill patients and is simple and practical,which can be applied to clinical work,there are still many shortcomings in this study,especially it is only a single-center retrospective study with small samples.Therefore,in order to enable medical staff to timely discover high-risk patients with VTE and take preventive measures,It is suggested to conduct multi-center and large sample experiments to further improve the prediction model of VTE in critically ill patients. |
PDF Full Text Request