Clinical Validation Of A Risk Prediction Model For Transient Ischemic Attack

Objective: The risk prediction model of transient ischemic attack has been modeled and has passed internal verification.In order to verify the authenticity and reliability of the risk prediction model of transient ischemic attack,the external verification of the model is specially carried out to promote the promotion and use of the model and contribute to the early prevention of TIA.Methods : A total of 80 patients treated and diagnosed with transient cerebral ischemic attack(TIA)from 1 January 2020 to 1 October 2021 were included in the TIA group,including 48 men and 32 women,and 50 patients admitted to our hospital and excluding TIA were randomly included as the non-TIA group,including 24 men and 26 women.Data collected include all enrolled patients hospital number,gender,age,history,smoking history,hypertension,clinical symptoms(dizziness,unilateral limb weakness,unilateral numbness,speech dysfunction,),laboratory data(hematocrit,average red blood cell volume,average hemoglobin,lymphocyte ratio,platelet distribution width,alkaline phosphate,white / ball ratio,cholesterol,triglycerides,LDL cholesterol,apolipoprotein B,homocycystine),carotid plaque formation,stenosis,and sinus inflammation.The 20 predictors(hypertension history,dizziness,unilateral limb weakness,unilateral limb numbness,unilateral limb numbness,speech dysfunction,hematocrit,average red blood cell volume,average hemoglobin amount,lymphocyte ratio,platelet distribution width,alkaline phosphatase,A /G ratio,cholesterol,triglycerides,LDL cholesterol,apolipoprotein B,homocysteic acid,carotid plaque,carotid stenosis,sinus inflammation)into the TIA risk prediction model,The SPSS software was then used for the data analysis and processing,Evaluation of TIA risk prediction model authenticity using subject operating characteristics(ROC)curves,The TIA risk prediction model was also evaluated using the Kappa coefficient.Results:(1)The risk prediction model of transient ischemic attack evaluated the ROC curve of TIA and non-TIA groups.The AUC was 0.799,(95%CI: 0.715,0.882),with a sensitivity of 83.8%,specificity of 76%,and yoden index of 0.598.It shows that the TIA risk prediction model has certain accuracy,application value and slightly lower specificity.(2)The 80 TIA patients were 67 positive by TIA risk prediction model,13 negative,50 non-TIA positive and 38 negative by TIA risk prediction model.The Kappa coefficient calculated a Kappa value of 0.595 and P <0.01,a statistically significant difference,indicating that the TIA risk prediction model has certain accuracy and general consistency.Conclusions:(1)The TIA risk prediction model had good accuracy,with a sensitivity of 83.8% and a specificity of 76%.(2))The TIA risk prediction model predictors of A / G ratio,hypertension history,and sinusitis did not differ between groups,which needs to be verified by a larger sample size.