Study On The Clinical Features Of Systemic Lupus Erythematosus Based On New Big Data Platform

?? Objective It is challengeable to accurately and personalized treat systemic lupus erythematosus(SLE),due to its high heterogeneity in clinical manifestation and outcome prognosis.This study attempts to take advantage of a new clinical data analysis platform to efficiently organize,collect of clinical manifestations,infection information,and treatment in SLE.Through developing rules and algorithms for lupus patients to efficiently and precisely evaluate,analyze disease activity and lupus' characters in east China so as to provide accurate?? Material and methods Establishing SLE database,through OCR,NLP,AI and other technologies based on Hitales platform,efficiently evaluate 01/06/1997 to 30/06/2015 in Shanghai a Hospital SLE patients' 791 medical files,in the form of electronic version,picture through scanning,recognition,proofreading and other process input into the system.According to the research content,develop appropriate interpretation of the rules,write a computer program interpretation,automatic extraction of variables,and outputs the formatted data,checks the datas to ensure the accuracy,and use SPSS statistical analysis.?? Results The accuracies between Hitales platform and manual tagging have no significant difference(P <0.05),especially in SLEDAI score.Male and female ratio is about 1:10,aged44 ± 15(19-86)years,age of first symptoms 34 ± 15(8-78)years,delayed diagnosis(413-0)days,the number of days per admission 13 ± 8(2-66)days.Using this platform,also find some clinical features that are difficult to find in previous research,such as,the ratio of combined Sjogren's syndrome in SLE patient,whose age of first symptoms is 41-60 years,is higher than other group(P <0.05);The ratio of combined respiratory infection in SLE patient,whose age of first symptoms is 61-80,is higher than other group(P <0.05);Men are more prone to kidney performance(P <0.05);Meanwhile preliminary analyze the main treatment.Patients mainly use corticosteroids and immunosuppressive agents,only two patients were prescribed with biological agents.Patients use corticosteroids to treat,including methylprednisolone(57.35%),prednisone(73.04%)and medrol(25%).Immunosuppressant mainly refers to hydroxychloroquine(62.75%).41 patients used shock therapy,the type of shock therapy for methylprednisolone(4.41%),cyclophosphamide(13.73%)and combination therapy(1.96%).Those lay a foundation for accurate treatment.?? Conclusion Extracting data by using Hitales platform is of convenience,fastness,efficiency and accuracy.On this basis,the accuracies of Hitales platform and manual tagging of SLEDAI have no significant difference(P <0.05).Lupus patient's condition is a symptom of the real world and diverse,highly heterogeneous,only the use of Big Data technologies,efficient use of each patient's information before they can grasp the whole picture of the disease,to be truly personalized treatment.Systemic lupus erythematosus is a disease with high heterogeneity.Only by using of big data technologies to use of each patient's information efficient,can we grasp the whole picture of the disease so as to be truly individualized treatment.