Epidemiological Characteristics Of Children With Cp In Xinjiang And Evaluation Of Muscle Function After Operation

Objective: Objective: To retrospectively analyze the data and information of cerebral palsy children in Xinjiang for 7 years,to analyze its epidemiological characteristics and distribution regularity,and to provide information reference for the prevention and control of cerebral palsy;to screen,extract and simplify the clinical diagnosis and classification of cerebral palsy children and the related clinical indicators of the choice of surgical methods by using data mining method,and to provide methodological reference for clinical decision-making of the disease.Methods: Data of children with cerebral palsy treated in Xinjiang Cerebral Palsy Rehabilitation Center from 2011 to 2017 and case information of hospitalized children from 2017 to 2018 were collected.Descriptive epidemiological method was used to analyze the sex,age,nationality and regional distribution characteristics of 1926 children with cerebral palsy from 2011 to 2017,and to compare the differences of epidemiological characteristics.In view of the main clinical diagnosis and classification of children with cerebral palsy and the surgical treatment plan,the decision tree method was used to extract and screen important decision-making indicators,and the significance of clinical decision-making was judged by classical logistic regression model.And feasibility.Results: The descriptive epidemiological method of 1926 children with cerebral palsy showed that there were more males than females in Xinjiang,and the age distribution was mainly from 3 to 14 years old.The regional distribution in southern Xinjiang was much higher than that in northern Xinjiang and Eastern Xinjiang.There was no significant difference in gender distribution between different regions(P > 0.05),but there were regional differences in age distribution and ethnic distribution(P <0.05).Decision tree model showed that 10 important decision-making indicators fordiagnosis and classification of cerebral palsy were muscle tension assessment,involuntary movement,muscle strength assessment,sitting position,supine position,Gordon sign,age of walking alone,age of ankle clonus,age of climbing and inability to speak.The prediction accuracy of decision tree model is 81.53%.The discriminant accuracy of Logistic regression model is 71.4%,which is slightly lower than that of decision tree.However,compared with decision tree,the effect of each decision index and whether it has statistical significance can be shown.Decision tree model was used to screen the clinical signs related to surgical methods in children with cerebral palsy.The order of importance was ankle clonus,diagnostic typing,sign,muscle strength evaluation,age of climbing,age of not raising head,control of stool and urine,age of sitting alone,muscle tension evaluation and lateral position.The accuracy of decision tree classification is87.11%.Logistic regression model can be used to further screen clinical features with statistical significance,and the accuracy rate is 81.30%.It can be seen that the combination of data mining method and classical logistic regression model is feasible in clinical decision-making.Conclusion: In view of the epidemiological characteristics of children with cerebral palsy in different areas of Xinjiang,it is necessary to carry out targeted prevention and control studies of cerebral palsy.It is feasible to apply data mining method and classical regression model to the diagnosis and operation discrimination of children with cerebral palsy.It can provide methodological research ideas for clinical decision-making.