Establishment And Evaluation Of Prediction Models For Predicting Postoperative Delirium After Deep Brain Stimulation Surgery For Parkinson's Disease

Objective:To explore the influencing factors of delirium after deep brain stimulation in Parkinson's disease,to construct a prediction model to predict delirium after deep brain stimulation in Parkinson's disease.Methods:From January 2015 to August 2019,271 patients with Parkinson's disease who underwent deep brain stimulation in the Department of Neurosurgery of hospital in Anhui Province were evaluated with the Confusion Assessment Method for the Intensive Care Unit?CAM-ICU?.According to the occurrence of delirium,they were divided into delirium group and non-delirium group.The general data of patients,perioperative?preoperative,intraoperative and postoperative?clinical data,specialist indicators of Parkinson's disease include disease stages[Hoehn-Yahr grade?H-Y?],motor symptoms/complications of drug therapy[Unified Parkinson's Disease Rating ScaleIII/IV?UPDRSIII/IV?],non-motor symptoms[Non-Motor Symptoms Scale for Parkinson's Disease?NMSS?],cognitive status[Mini-Mental State Examination?MMSE?],sleep quality[Parkinson's Disease Sleep Scale?PDSS?],quality of life[The39-item Parkinson's Disease Questionnaire?PDQ-39?].Univariate analysis of delirium group and non-delirium group data,univariate statistically significant variables?P<0.05?were included in Logistic multivariate regression analysis to screen the independent influencing factors.The independent influencing factors are introduced into R software,and the rms package is used to establish nomogram.The nonparametric repeated sampling method?Bootstrap resampling times=500?is used to verify the Prediction model,and the Hosmer-Lemeshow goodness of fit test is used to evaluate the calibration of the prediction model,and the ROC curve of the prediction model is made.The area under the curve(AUC_ROC)and 95%CI are calculated to judge the prediction efficiency of the model.Clinical decision graph?Decision Curve Analysis,DCA?was used to judge the clinical applicability of the nomogram.Results:In this study,271 patients undergoing deep brain stimulation of Parkinson's disease were investigated,of which 52 patients developed postoperative delirium?19.19%?.Univariate analysis revealed that the age,sex,preoperative length of stay,preoperative Barthel score,the presence of preoperative brain atrophy and the presence of preoperative cerebral ischemia were statistically significant.The results from the Non-Motor Symptoms Scale for Parkinson's Disease?NMSS?,Mini-Mental State Examination?MMSE?,Parkinson's Disease Sleep Scale?PDSS?and Unified Parkinson's Disease Rating Scale III?UPDRS III?were also statistically significant?P<0.05?.Multivariatelogisticregressionanalysisshowedthat Unified Parkinson's Disease Rating Scale III?UPDRS III??Odds Ratio[OR]=1.028,95%Confidence Interval[CI]:1.002¹.055?,Non-Motor Symptoms Scale for Parkinson's Disease?NMSS??OR=1.011,95%CI:1.000¹.022?,Parkinson's Disease Sleep Scale?PDSS??OR=0.980,95%CI:0.963⁰.997?,preoperative length of stay?OR=1.267,95%CI:1.081¹.484?,age of onset of disease?OR=1.048,95%CI:1.007¹.090?and preoperative brain atropy?OR=5.425,95%CI:1.905¹5.445?were independent factors that influenced the occurrence of Postoperative delirium after Deep brain stimulation surgery.A nomogram model was constructed using the six forecasting factors.The nonparametric repeated sampling method?Bootstrap resampling times=500?is used to verify the nomogram.The results show that the area under the ROC curve of the nomogram is?Area Under Curve,before and after internal verification.The AUC?were 0.808?95%CI:0.745⁰.871?and 0.809?95%CI:0.746⁰.872?,and the predicted risk critical values were 0.220 and 0.213,respectively,suggesting that the line diagram model has a good degree of differentiation.The H/L deviation test of the nomogram before and after internal verification is 2⁵.490 and5.220,P<0.704 and 0.734,respectively.Combined with the model calibration curve,it is suggested that the nomogram model has a good calibration degree.According to the clinical decision curve?Decision Curve Analysis,DCA?,when the risk critical value of the nomogram model is 0.220,the standardized net return?Standardized Net Benefit,SNB?is 0.451,that is,when the predicted value is 22%as the cut-off value of the prediction model,For every 100 patients who use this model,45 people can benefit from it without harming the interests of anyone else,suggesting that the nomogram model has better clinical applicability.Conclusions:In this study,the influencing factors of postoperative delirium in patients undergoing deep brain stimulation of Parkinson's disease were studied.Higher UPDRS III,NMSS scores,prolonged preoperative length of stay,lower age of onset of disease and preoperative brain atrophy were independent risk factors for postoperative delirium after deep brain stimulation in patients with Parkinson's disease.The higher PDSS score was an independent protective factor for postoperative delirium after deep brain stimulation in patients with Parkinson's disease.A nomogram for predicting the risk of postoperative delirium after deep brain stimulation of Parkinson's disease was constructed.The nomogram can intuitively and succinctly provide an individualized probability of postoperative delirium risk for patients undergoing deep brain stimulation of Parkinson's disease.In the treatment and nursing decision-making to provide clinical doctors and nurses with predictive tools to reduce the incidence of postoperative delirium after deep brain stimulation of Parkinson's disease.