Construction And Validation Of A Risk Prediction Model For Orthostatic Intolerance After Total Hip Arthroplasty

Objective1.To construct the risk prediction model of orthostatic intolerance（OI）in patients after total hip arthroplasty.2.To validat the efficacy of the OI risk prediction model.Methods1.After literature review and discussion by the research group,the draft of Orthostatic Intolerance Risk Factor Questionnaire for patients after total hip arthroplasty was developed,and the final draft was revised through expert meeting method.A total of295 patients after total hip arthroplasty,from June 2022 to September 2022,who were admitted to the Department of Joint Surgery of a tertiary Class A general hospital in Qingdao city,Shandong Province,were selected as research subjects by convenience sampling method.The status quo of orthostatic intolerance after total hip arthroplasty in joint surgery was investigated by using patient general data questionnaire,orthostatic intolerance risk factor questionnaire and OI symptom questionnaire.2.Using SPSS 25.0 statistical software,the relevant data of the 295 patients after total hip arthroplasty were collated for univariate analysis on the influencing factors of orthostatic intolerance in the modeling group,and the significant influencing factors were analyzed by backward Wald stepwise selection method.Then,multivariate Logistic regression analysis was used.The risk prediction model of orthostatic intolerance after total hip arthroplasty was constructed according to the weight of influencing factors.The R 4.2.2 language was applied to draw column line plots to visualize the model results for clinical application.3.The Bootstrap self-sampling method was applied to internally validate the predictive model.Following the same exclusion and inclusion criteria,127 patients admitted to the joint surgery department of the same hospital from October to November2022 were used as the verification group.Hosmer-Lemeshow test could verify the goodness of fit of the model,and AUC value（area under the curve）was obtained by drawing ROC curve.The specificity,sensitivity and accuracy of the risk prediction model for orthostatic intolerance in patients after total hip arthroplasty were further determined.The prediction probability（P=0.5）was taken as the decision point,and a value（P>0.5）was considered as orthostatic intolerance.The closer the AUC value is to 1,the better the prediction effect of the model.The fitting effect of the prediction model was verified by the Hosmer-Lemeshow test,and P>0.05 was determined to be a statistically insignificant difference.According to calibration curves and Brier score,the model calibration was checked.Results1.According to the expert meeting and literature review method,the related factor questionnaire of patients after total hip arthroplasty was prepared.The questionnaire includes basic patient data,surgical data and postoperative data.2.A total of 422 patients were included in this study.The incidence of orthostatic intolerance was 42.37%（125/295）in 295 patients after total hip arthroplasty in the modeling group.Univariate analysis of orthostatic intolerance after total hip arthroplasty showed that:age（P<0.001,t=37.29）,sex（P=0.025,χ²=5.35）,body mass index（BMI）（P<0.001,t=24.01）,postural adaptation training（P<0.001,χ²=21.91）,hypertension（P<=0.009,χ²=6.74）,fear of falling（P<0.001,t=41.29）,decreased hemoglobin level（P<0.001,t=29.8）and pain score（P<0.001,t=4.72）were significantly different.3.The multi-factor Logistic regression analysis showed:age X₁（P=0.006,OR=1.042,95%CI:1.012-1.072）,body mass index X₃（P=0.002,OR=1.191,95%CI:1.068-1.328）,postural adaptation training X₄（P<0.001,OR=0.874,95%CI:0.852-0.905）,fear of falling X₆（P<0.001,OR=0.956,95%CI:0.935-0.978）,decreased hemoglobin level X₇（P<0.001,OR=1.096,95%CI:1.049-1.145）,pain score X₈（P=0.001,OR=1.461,95%CI:1.178-1.812）were included in the logistic regression model.The logistic regression model was constructed to predict the risk of orthostatic intolerance in patients after total hip arthroplasty:Logit（Y）=ln（p/1-p）=-10.041+0.041*X₁+0.175*X₃+0.092*X₇+0.379*X₈-0.042*X4-0.045*X_6。The probability formula of orthostatic intolerance is as follows:P=Exp（-10.041+0.041*age+0.175*body mass index+0.092*decreased hemoglobin level+0.379*pain score-0.042*postural adaptation training-0.045*fear of falling）/[1+Exp（-10.041+0.041*age+0.175*body mass index+0.092*decreased hemoglobin level+0.379*pain score-0.042*postural adaptation training-0.045*fear of falling）].Draw the nomogram of the orthostatic intolerance risk prediction model for patients after total hip arthroplasty.4.The risk prediction model of orthostatic intolerance in patients after total hip arthroplasty was verified by Hosmer-Lemeshow test,χ²=2.553,df=8,P=0.959.The calibrated C-statistic obtained by Bootstrap self-sampling method was 0.812,indicating that the model distinguishment was good.The area under ROC curve of the modeling group was 0.843,and that of the validation group was 0.852.There were 39 overlapped patients of orthostatic intolerance between two groups by the prediction model.There were 64 overlapped patients without orthostatic intolerance.The sensitivity,specificity,and accuracy of the model for predicting orthostatic intolerance after total hip arthroplasty were 68.42%（39/57）,91.43%（64/70）,and 81.10%（103/127）.The brier score was 0.162（<0.25）.According to the calibration curve,the predicted and measured values agreed well.Conclusions1.Fear of falling,age,decreased hemoglobin level,body mass index,pain score（NRS）and postural adaptation training were independent predictors of orthostatic intolerance after total hip arthroplasty,and age,decreased hemoglobin level,body mass index and pain score（NRS）were risk factors of orthostatic intolerance.Fear of falling and Postural adaptation training were protective factors for orthostatict intolerance.2.The established risk prediction model of orthostatic intolerance in patients after total hip arthroplasty has been verified to have good prediction value.The prediction model is convenient and feasible,with stable prediction performance.It can provide guidance for medical staff to effectively predict and intervene the patients of orthostatic intolerance after total hip arthroplasty.