Construction And Application Of The Prediction Model For Risks In TIA Patients For Progressing To Ischemic Stroke Within 7 Days

ObjectiveBased on the risk management theory,the objective of this study is to investigate and analyze the risk factors for patients with TIA who progressed to ischemic stroke within 7 days and their correlation with ischemic stroke,to construct a risk prediction model for patients with TIA who progressed to ischemic stroke within 7 days,and to validate the model clinically.Methods1.Research on risk factors associated with the patients with TIA who progressed to ischemic stroke within 7 daysIn this study,a literature review and a retrospective case-control method were used in the risk identification phase of the risk management theory to investigate and analyze the risk factors for patients with TIA who progressed to ischemic stroke within 7 days and their correlation with ischemic stroke.In the course of analyzing the literature,10 studies were finally selected based on the search strategy,inclusion and exclusion criteria.Afterwards,risk factors associated with patients with TIA who progressed to ischemic stroke within 7 days were extracted from these 10 studies by two different investigators.When the two investigators disagreed,a third party should be consulted.Based on the analysis of literature data,the medical records of the final selected 412 inpatients were extracted and analyzed using a retrospective case-control method based on inclusion and exclusion criteria.The included patient sample was obtained from inpatients admitted to the neurology,neurosurgery and geriatrics wards of a level-3 comprehensive hospital in Kaifeng,Henan Province,over the past 10 years（December 2011-December 2021）.Patients who were first diagnosed with TIA and admitted within 48hours were divided into a case group（patients with TIA who progressed to ischemic stroke within 7 days）and a control group（patients with TIA who did not progress to ischemic stroke within 7 days）according to whether they progressed to ischemic stroke within one week.The basic information of patients with TIA and the risk factors associated with patients with TIA who progressed to ischemic stroke within 7 days in the case group were collected and recorded in the medical records.Then univariate analysis and binary logistic regression analysis were used to identify the factors associated with patients with TIA who progressed to ischemic stroke within 7 days.2.The construction of a risk prediction model for patients with TIA who progressed to ischemic stroke within 7 daysIn the risk assessment phase of the risk management theory,a risk prediction model for patients with TIA who progressed to ischemic stroke within 7 days was constructed based on the coefficients in the logistic regression model described above.The risk prediction model was then evaluated by using the Wald test,the Hosmer-Lemeshow test and plotting the ROC curve.3.The clinical validation of a risk prediction model for patients with TIA who progressed to ischemic stroke within 7 daysIn the effect evaluation phase of the risk management theory,medical records of 209 inpatients were selected,extracted and analyzed to test the clinical prediction effect of the above prediction model based on the same inclusion and exclusion criteria as the retrospective case-control method.The patients in the validation group included in the clinical validation and the patients in the modeling group included in the retrospective case-control research were the same inpatients of the neurology,neurosurgery and geriatrics wards of a a level-3 comprehensive hospital in Kaifeng,Henan Province.Unlike the modeling group（December 2011-December 2021）,the patients in the validation group included in the clinical validation were inpatients during January-June 2022.The basic information of patients with TIA and the risk factors associated with patients with TIA who progressed to ischemic stroke within 7 days in the medical record data were collected and recorded.Based on the data obtained,the difference between the predicted and actual incidence of ischemic stroke in TIA patients within 7 days was calculated and compared so as to analyze and evaluate the effectiveness of the prediction model.4.Statistical methodsIn this study,Excel software was used to input the data,and SPSS 23.0 software was applied to statistically process and analyze the obtained data.The count data obtained in this study were expressed as frequency（n）and composition ratio（%）.When comparing the characteristics of the two groups of patients,the differences were analyzed using theχ²test or the independent samples t-test.Results1.Risk factors associated with progression to ischemic stroke within 7 days in patients with TIAIn this study,totally 20 risk factors associated with progression to ischemic stroke within 7 days in patients with TIA were extracted from the final 10 included papers by two different investigators according to the nadir criteria in the risk identification phase of risk management theory:age≥60 years,male,high blood pressure,coronary artery disease,diabetes,atrial fibrillation,family history,number of TIA episodes≥3,duration of symptoms≥1 hour,unilateral limb weakness and/or speech impairment,ABCD2 score,positive diffusion-weighted imaging,unstable carotid plaque,intracranial artery stenosis≥50%,LDL cholesterol,total cholesterol,triacylglycerol,homocysteine,fibrinogen,and ultrasensitive C-reactive protein by adopting literature-based research approach in the risk identification phase of risk management theory.The retrospective case-control study method was adopted accordingly on such basis.The patients with TIA were divided into the case group（group of patients with TIA who progressed to ischemic stroke within7 days）and the control group（group of patients with TIA who did not progress to ischemic stroke within 7days）according to whether they progressed to ischemic stroke within 7 days.The case group included 103patients with a mean age of 65.84±12.16 years old.Among them,there were 54 male patients（52.4%）and49 female patients（47.6%）.The control group included 309 patients with a mean age of 60.33±13.11 years old.Among them,170（55.0%）were male patients and 139（45.0%）were female patients.Information related to the progression to ischemic stroke within 7 days in patients with TIA included general information,current medical history,family history,clinical manifestations,as well as the various biochemical and imaging indices.The largest proportion of the 412 patients included were married（93.0%）and employed（67.5%）.The majority of them were non-smokers（70.9%）and non-drinkers（75.7%）.Compared to the control group,there was a higher proportion of those with a family history（65%）and a significantly higher level of ultrasensitive C-reactive protein（10.87±1.33）.It was concluded after univariate analysis that there were significant statistical differences（P<0.05）between patients in the case group and the control group for 16 variables,including family history,unilateral limb weakness,ABCD2 score≥6,cerebral atherosclerosis,ultrasensitive C-reactive protein and total cholesterol.It was revealed by binary logistic regression analysis that the following factors could be regarded as significant risk factors for progression to ischemic stroke within 7 days in patients with TIA:ultrasensitive C-reactive protein（OR,3.990;95%CI,2.853～5.580）、total cholesterol（OR,2.107;95%CI,1.484～2.991）、low-density lipoprotein cholesterol（OR,1.934;95%CI,1.192～3.138）、family history（OR,2.387;95%CI,1.148～4.960）、unilateral limb weakness（OR,2.759;95%CI,1.318～5.775）、ABCD2 score≥6（OR,3.144;95%CI,1.449～6.822）、hyperlipidemia（OR,2.173;95%CI,1.062～4.447）、intracranial artery stenosis≥50%（OR,4.598;95%CI,1.650～12.812）、number of episodes≥3（OR,3.052;95%CI,1.385～6.725）.2.Construction and validation of a risk prediction model for progression to ischemic stroke within 7days in patients with TIAIn this study,the risk probability equation for predicting progression to ischemic stroke within 7 days in patients with TIA was derived from theβvalues in binary logistic regression analysis during the risk assessment phase of risk management theory as:Logit（P）=-21.994+1.384×ultrasensitive C-reactive protein+0.745×total cholesterol+0.660×LDL cholesterol+0.870×family history+1.015×unilateral limb weakness+1.145×ABCD2 score≥6+0.776×hyperlipidemia+1.526×intracranial artery stenosis≥50%+1.116×number of episodes≥3The final model for predicting the risk of progression to ischemic stroke within 7 days in patients with TIA was then established according to P=e^Logit（p）/1+e^Logit（p）i.e:P=1/{1+exp（1.384×hypersensitive C-reactive protein+0.745×total cholesterol+0.660×LDL cholesterol+0.870×family history+1.015×unilateral limb weakness+1.145×ABCD2 score≥6+0.776×hyperlipidemia+1.526×intracranial artery stenosis≥50%+1.116×number of episodes≥3 episodes-21.994）}Theχ²value of the prediction model for the risk of progression to ischemic stroke within 7 days in patients with TIA was 93.237,P<0.001,which further indicated that the regression equation was significant;the Hosmer-Lemeshow goodness-of-fit test statistic was calculated（χ²=7.573,P=0.476）,which further indicated that the logistic model fitted the observed data well;it was revealed by the interaction classification table results that the overall accuracy of the model was 88.8%.When the Jorden index（sensitivity+specificity-1）took the maximum value（0.767）,the prediction probability value（0.165）was the optimal threshold for the model and the sensitivity and specificity reached the highest values which were 93.2%and 83.5%respectively.The prediction ability of the model was the highest by this time.The probability that the patients with TIA had the highest probability of progression to ischemic stroke within 7days was the highest when the prediction probability was≥0.165.The area under the ROC curve of this model was 0.947（P<0.001;95%CI,0.924～0.969）.In this study,the risk prediction model was clinically validated during the effect evaluation phase of risk management theory.Totally 209 patients with TIA and a mean age of 63.00±12.31 years,were included in the clinical validation according to the nadir criteria.Of these 209 patients,120（57.4%）were male and 89（42.6%）were female.Among them,the highest percentage was among married（95.7%）and employed（73.7%）.Comparing the prediction of the risk prediction model constructed above with the actual situation,the area under the ROC curve value was 0.828（P<0.001;95%CI,0.765～0.891）,the sensitivity was 62.1%（41/66）and the specificity was 92.3%（132/143）,that is,the true positive rate was slightly lower than the true negative rate,which indicated that the model has slightly lower ability to predict TIA patients to progress to ischemic stroke within 7 days and the higher screening ability for patients without ischemic stroke.It was possibly related to the size of the sample included in the clinical validation.However,the overall prediction accuracy of the present risk prediction model was 82.8%（173/209）,which also indicated the high discrimination of the model.Based on the above findings,the study combined the risk factors for progression to ischemic stroke within 7 days in patients with TIA,the risk prediction model and clinical validation of the risk prediction model with risk management theory for constructing a preliminary theoretical model of risk prediction of progression to ischemic stroke within 7 days in patients with TIA.Conclusion1.The independent risk factors affecting TIA patients’progression to ischemic stroke within 7 days include ultrasensitive C-reactive protein,total cholesterol,low-density lipoprotein cholesterol,family history,unilateral limb weakness,ABCD2 score≥6,hyperlipidemia,intracranial artery stenosis≥50%and number of episodes≥3.2.A risk prediction model for TIA patients progressing to ischemic stroke within 7 days was finally constructed.After that,it was found out that the model has good predictive risk ability and application value,and it can achieve screening and assessment for high-risk groups of TIA patients progressing to ischemic stroke within 7 days.3.The incidence of TIA patients progressing to ischemic stroke within 7 days was 19.6%,which suggested that risk assessment and prevention of short-term progression to ischemic stroke in TIA patients should be given high priority by health care professionals.