Evaluation And Prediction Of Clinical Effect Of Acupuncture On Chronic Low Back Pain

Chronic low back pain is one of the common clinical diseases,frequent diseases and high disability rate diseases.Most people will experience low back pain.Low back pain affects the work and life of patients.At present,there is a lack of safe and effective treatment.Acupuncture as a possible alternative therapy,its effectiveness lacks evidence-based medical evidence.Based on the registration platform of acupuncture and moxibustion hospital of Chinese Academy of Traditional Chinese Medicine for the treatment of chronic low back pain,according to the clinical case data,including the relevant information of the doctor,the demographic characteristics of the included cases,basic condition,diagnosis of traditional Chinese and western medicine and acupuncture treatment,this dissertation is focused on the basic analysis to the effect of acupuncture treatment of chronic low back pain,prognostic factors,and predicts the curative effect.First of all,using exploratory data analysis to analyze the basic features of observed variables,to understand the doctor,patient,condition and treatment,and to evaluate its short-term effect of acupuncture treatment of chronic low back pain.It is concluded that the short-term effect of acupuncture treatment of chronic low back pain is quite good,and the total effective rate is 85.1%.Secondly,according to the evaluation criteria of curative effect,the patients can be divided into effective group and invalid group.Then whether there are significant differences in the performance of these two groups in various aspects is inspected.The Wilcoxon rank sum tests are carried out for the quantitative variables and the independent tests for contingency tables are employed for qualitative variables,and the predictor variables are selected.Then the validity is taken as the the response variable and several prediction models are considered.Finally,the classification prediction models of double logistic regression based on SCAD penalized extended quasi-likelihood function,random forest and BP neural network are constructed.The double logistic regression model based on the extended quasi-likelihood function with SCAD penalty is concluded to perform well in terms of AUC value（0.774）,F₁ value（0.925）and accuracy（87.1%）.Of the influencing factors of this model,the first NRS,doctor’s gender,level II hospitals at the hospital level and millic-acupuncture-assisted therapy combined with TDP irradiation show positive correlation.Chief physician’s title,doctor’s master degree,non-drug therapy in other traditional Chinese and western medicine,meridian syndrome differentiation（Shaoyin meridian）and induced body reaction（muscle twitching reaction）showed negative correlation.In addition to doctor’s title and doctor’s degree,other influencing factors of this model have practical clinical significance.This prediction model provides reliable evidence for the short-term efficacy evaluation of acupuncture in the treatment of chronic low back pain.It can make personalized and short-term efficacy prediction,find out its influencing factors,which is beneficial to classify patients with chronic low back pain and optimize the structure of acupuncture diagnosis and treatment,so as to improve the effectiveness of acupuncture treatment and optimize the allocation of medical resources.