| Backgrounds In real-world,the influencing factors of acupuncture treatment for chronic low back pain(CLBP)are unclear.There is no individual prediction model suitable for acupuncture treatment of CLBP.Based on the registration platform,we intend to descriptively analyze the diagnosis and acupuncture details in the real-world;To explore the factors affecting the short-term efficacy of acupuncture for CLBP;To establish individualized prediction model of the short-term efficacy of acupuncture for chronic low back pain and evaluate,verify,and visualize it.Design Prospective,Multicenter Registry StudyMethods(1)Descriptive analysis methods are used to describe the diagnosis and treatment characteristics and short-term effects of acupuncture treatment of chronic low back pain in the realworld from different levels;(2)Based on the maximum likelihood ratio stepwise forward binary logistic regression analysis method,explore the influencing factors of short-term efficacy,and draw the receiver operating characteristic curve(ROC);(3)Use systematic random methods in SPSS software to divide patients into developing group and verification group according to 7: 3.Logistic regression analysis and clinical importance are used in the modeling group to screen independent predictors of short-term effective acupuncture treatment of chronic low back pain.And use R language to build a prediction model and draw a nomogram.(4)Use the discrimination and calibration test methods to evaluate the model,use internal data for cross-validation and Bootstrap verification to answer the internal validity of the model.Use the decision curve analysis method(DCA,Decision Curve Analysis)and clinical impact curve(CIC,Clinical Impact Curve)to compare and evaluate the clinical benefit of the model.Research Contents(1)Clean the data in the platform,and analyze descriptively,the diagnostic characteristics and acupuncture treatment rules of acupuncture treatment for CLBP in the real-world;(2)Evaluate the short-term effect of acupuncture on chronic low back pain;(3)Exploring the method of establishing a predictive model of short-term efficacy of chronic low back pain acupuncture treatment,and performing a goodness-of-fit test,discrimination evaluation,calibration evaluation,clinical decision analysisand clinical impact analysis on the established model,and drawing ROC curve Nomograph,calibration curve,decision curve and clinical impact curve are displayed visually.Part I Real-world research on acupuncture treatment of chronic low back pain1.Recruitment(1)Recruitment of physicians: 32 acupuncturists and 563 patients from 12 hospitals across the country(including 10 cases of testing and incomplete baseline).Among them,the top three general hospitals are mostly concentrated in Beijing.The average working life of physicians is 12.68 years.There are 13 senior titles and 19 junior and intermediate titles.(2)Recruitment of patients: Recruit 553 patients with complete baseline,the main patients are from Beijing(320),the age distribution of patients is between 18-89 years old,nearly 65% are female patients,and the Han people account for 96.8% Staff accounted for 35.3%,retirees accounted for 31.6%,university and higher education accounted for 66.0%,and married accounted for 89.9%.2.2.The condition and diagnosis of chronic low back pain(1)Information on the condition of low back pain: The average duration of the disease is 4.70 years,and the main pain is distributed in the middle of the spine and within three inches,with a frequency of 652 times;the nature of pain is mainly sore,with a frequency of 308 times.The causes of pain aggravation are concentrated in factors such as coldness,humidity,and sedentary standing.Nearly half of the patients have low back pain accompanied by lower extremity pain and dyskinesia,and 63.5% of the patients have received acupuncture treatment.(2)Western medicine classification of diseases and chronic low back pain TCM: The Western disease diagnostic categories concentrated in disc disease and soft tissue lesions,accounting for 68.9% and 20.4%,respectively.The syndrome differentiation of traditional Chinese medicine is concentrated in kidney-back pain,cold-wet waist pain and blood stasis back pain,accounting for 40.5%,41.8% and 37.6%,respectively,and some patients have two or more syndromes combined;Shaoyang Meridian,accounting for 88.3%,37.4% and 24.8%,respectively;nearly half of the patients have two or more meridian syndromes,mainly to the foot of the meridian;only 12.4% of patients with other diseases.3.Treatment of patients with chronic low back pain(1)Distribution of acupoint prescriptions the top ten main acupoints are: Shenshu(BL23),Dachangshu(BL25),Jiaji(EX-B2),Weizhong(BL40),Huantiao(GB30),Guanyuan(RN4),Yaoyangguan(DU3),Qihaishu(BL24),Zhibian(BL54)and A-shi.Among them,Shenshu(BL23)and Dachangshu(BL25)accounted for 63.8% and 63.7%,respectively.The top ten matching points are: Weizhong(BL40),Kunlun(BL60),Chengshan(BL57),Yanglingquan(GB34),Huantiao(GB30),Zhibian(BL54),A-shi,Taixi(KI3),Tou Jiaji,Weizhong and Kunlun accounting for 40% and 27.5%.(2)Types of acupuncture and acupoint selection: the main types of acupuncture are traditional Chinese medicine acupuncture(99.8%)and the principle of local acupoint selection(89.9%).Nearly two-thirds of the patients in the principle of acupoint selection used local acupoint selection,distal acupoint selection and acupoint selection.The number of acupuncture points was 168(30.4%),276(49.9%)and 109(19.7%)according to less than 5 points,5-10 points and 10 extra points.(3)The law of acupuncture treatment: the acupuncture tools mainly consist of milli-needle.Among the body reactions caused by acupuncture,Deqi and muscle twitch response account for 98% and 23.3%,respectively.The auxiliary operation of milli-needle is mainly TDP irradiation(70.7%),electro-acupuncture stimulation(48.3%)and manual acupuncture(23.7%).Most patients will use TDP irradiation and electro-acupuncture as auxiliary therapy.Special acupuncture was used in 26.6% of patients,including unconventional needle specifications,acupuncture site,acupuncture angle,acupuncture method and acupuncture depth,acupuncture technique and acupuncture posture,such as sitting acupuncture,beating needle Law etc.For needle retention,needle retention for almost all patients and the needle retention time parameter is selected to be 20-30 minutes.The cumulative number of acupuncture treatments more than 3 times was 428,accounting for 77.4%.Of the 125 people with less than 3 treatments,only 39 were cured,and the rest were all shed.In the cumulative acupuncture cycle,the main acupuncture cycle is 2 weeks to 1 month,of which less than 2 weeks,2 weeks to January,and more than January account for 34.0%,53.2%,and 6.5%,respectively.Among the acupuncture frequency of millineedle(times / week),more than 3 treatments per week are more common.Among them,1 times/week,2 times/week,and more than 3 times / week accounted for 32.9%,21.0%,46.1%.(4)The combined treatment of acupuncture most patients did not combine other acupuncture and non-acupuncture treatments.Among the 225 combined acupuncture treatments,cupping therapy was the main method,accounting for 32.4% of the total number of patients treated;Among the 198 non-acupuncture patients,drug therapy was the mainstay,accounting for 22.8% of the total number of patients treated.4.The effect of acupuncture on patients with chronic low back pain(1)Effectiveness of registered patients: Among the registered 553 patients,the number of censors was 77,and the censoring rate reached 13.9%.According to the principle of intention analysis,the patient was still analyzed according to 553,and the final scale score was subtracted the baseline score was judged to be effective by MICD as the standard.Among them,427 people had an NRS score of 2 or less,and the treatment effective rate was 77.2%.ODI index fell by more than 10 points to 399.The treatment effective rate was 72.1%.NRS At the same time as ODI,it was determined that 386 people were effective,the treatment effective rate was 69.8%,the total number of effective treatments was 440,and the total effective rate was 82.5%.(2)The short-term effectiveness of the patients included in the study: Of the 553 registered patients,428 cases were judged as eligible cases after more than three treatments.The scale score after treatment minus the baseline NRS or ODI index,with MICD as the standard 312 patients were judged to be effective,of which 303 were NRS scores greater than or equal to 2 points,the treatment effective rate was 70.8%,139 were reduced by an ODI index of more than 10 points,the treatment effective rate was 25.1%,NRS and ODI At the same time,it was determined that 130 were effective,the effective rate of treatment was 30.4%,the total effective number of treatment was 312,the invalid number was 116,and the total effective rate was 72.9%.5.Safety evaluation: The incidence of adverse events was 2.4%,a total of 13 person-times,all of which were not serious adverse events,and most of them were not related to acupuncture.6.Comparative study of the effective group and the ineffective group included in the study(1)Comparison of baseline data and single-valued variables between the two groups by Mann-Whitney U test: Comparison of baseline low back pain duration,NRS score,and ODI index variables between the two groups,the P value was less than 0.05,and there was a statistical difference between the two groups;After chi-square test or Fisher exact test,the ratio of the two groups of patients in age,gender,region,ethnicity,marital status and other variables,the P value is greater than 0.05,there is no statistically significant difference between the two groups.(2)Comparison of the frequency of multiple-selected variables between the two groups: The variables such as the patient's pain location,type of pain,exacerbating factors,accompanying symptoms,causes of pain,etiological dialectics,acupuncture response,adjuvant therapy with milli-needle,and other acupuncture treatments combined For multi-valued variables,we count the values corresponding to each variable and generate new variables.The number of dialectics of the cause and the number of body reactions caused by acupuncture were compared by chisquare test,and the difference between the two groups was statistically significant with P value <0.05.(3)Comparison between multiple selection variables between two groups: Because the multiple selection variables between the two groups are multi-dimensional categorical variables,the difference comparison cannot be performed for the time being,each group of variables is defined as multiple response variable levels,and the cross-tab method is used.Statistical description.The distribution of pain groups,nature,induced remission factors,concomitant symptoms,diagnosis types,TCM syndrome types,meridian dialectical types,patients' response to treatment and acupuncture adjuvant treatment,combined treatment and other levels were basically the same in both groups.The main pain parts are distributed in the middle of the spine and within three inches on both sides of the spine;the pain distribution is concentrated in soreness and bloating;the aggravating causes are concentrated in sedentary stations and cold and humid factors;the symptoms are concentrated in the lower limbs;The etiological dialectical focus is kidney back pain and cold and damp back pain;the meridian dialectic aggravates the bladder meridian and the Shaoyang meridian;the most common adjuvant therapy and combined treatment of micro acupuncture are TDP irradiation and cupping,respectively.7.Univariate and multivariate logistic regression(1)Baseline univariate logistic regression and screening: Univariate binary logistics regression analysis results,NRS score,ODI index,course of disease,number of causes of pain exacerbation,number of causes of pain relief,number of acupoint selection principles,number of causes dialectic,millimeter needle Variables such as types of adjuvant therapy are significant(P <0.20).(2)Logistic regression of multiple response variable values: The variable variables of multiple response variables were subjected to single factor logistic regression.The low back pain was within and beyond three inches of the spine on both sides of the waist.The pain was reduced by warming,rubbing,muscle spasms,coughing,sneezing,kidney deficiency The P values of low back pain,blood stasis and low back pain,five-element acupuncture,exercise acupuncture,cupping and fire acupuncture are all less than 0.2,which has a higher significance.(3)The establishment of the equation of the binary logistic regression and the regression of the ROC curve by stepwise maximum likelihood estimation regression: soreness,tingling,three inches beyond the spine on both sides of the waist,coughing and sneezing,Shaoyang Meridian,symptomatic point selection,NRS Seven independent variables scored were included in the regression equation.The obtained OR and P values were: soreness(OR = 0.397,95% CI 0.218-0.723),tingling(OR = 0.384,95% CI 0.155-0.951),and three inches beyond the spine on both sides of the waist(OR = 3.035,95% CI 1.150-8.010),cough and sneeze(OR = 0.374,95% CI 0.129-1.085),Shaoyang Meridian(OR = 0.403,95% CI 0.204-0.797),symptomatic point selection(OR = 0.404,95% CI 0.225-0.725 and baseline NRS(OR = 1.588,95% CI 1.330-0.895).Construct a short-term acupuncture effective regression equation based on the OR value of the variable,the equation is Logit(P)= 0.615 + 0.397 * ache + 0;384 * No stinging pain + 3.035 * Pain is more than three inches from the spine on both sides of the waist + 0.387 *(coughing and sneezing)+ 0.403 * dialectic of the Shaoyang Meridian + 0.404 * Selected points for symptomatic + 1.55 * NRS score.8 Evaluation of regression model and drawing of ROC curve According to the Hosmer-Lemmershaw test of the model,P = 0.256> 0.05,which indicates that the model has good calibration and good fitting.According to Negorko's R2 evaluation,the model can explain 26.6% of the variation in the effective factors affecting acupuncture in the short term.The negative predictive value is 45.2,the positive predictive value is 94.9%,and the comprehensive predictive value of the model is 81.1%.Drawing the ROC curve according to the predicted probability value indicates that the area under the curve of the model AUC is 0.759,95% confidence interval(0.694-0.824),and has a good fit;the maximum Youden index calculated based on the coordinates under the curve is 0.44(Yordon index = sensitivity-(1-specificity)),the Cutoff value at this time is 0.614,the sensitivity is 0.839,and the specificity is 0.393.9.Collinearity diagnosis of regression variables the matching tolerance of the seven variables included in the model and the VIF method(variance expansion factor)are stably distributed around 1,indicating that the factors in the model have good independence from each other.Part II Establishment,verification,and evaluation of shortterm curative effect prediction model of acupuncture and moxibustion for CLBP.1.Population characteristics of the modeling group and the verification group: After the system randomization,the modeling group and the verification group are age,gender,disease course,baseline NRS,ODI,patient regional distribution,education level,acupuncture treatment history,marital status.Compared with baseline variables such as ethnic,Western medical diagnosis categories,P >0.05,and there is no statistical difference in baseline between the modeling group and the verification group;2.In the regression analysis of the modeling group based on univariate multivariate,a total of 9 variables with a P value of less than 0.05 were obtained: baseline NRS score,baseline ODI index,age,pain location_three inches outside the waist and spine,pain nature_Soreness,symptomatic treatment,Shaoyang Meridian,causes of pain_cough,sneeze,diagnostic classification_spondyloarthropathy,included in the regression model obtained AUC = 304,indicating that the model fits well.We conducted a collinear diagnostic analysis of the selected variables.The results suggest that the variance expansion factor VIF is less than 10,indicating that the nine variables are independent of each other and there is no collinearity.3.The calibration and differentiation of the prediction model are good.The probabilistic prediction probability of acupuncture in patients with chronic low back pain in the short term is consistent with the actual clinical probability.Conclusions 1.In the real-world,the 7 influencing factors that affect the short-term curative effect of acupuncture on chronic low back pain are soreness,tingling,pain distributed three inches away from the spine on both sides of the waist,cough and sneezinginduced pain,dialectical Shaoyang Meridian,symptom selection,NRS score.2.The prediction model and nomogram established based on the registration research have good discrimination,calibration,and clinical impact,and can be used as an effective tool for shortterm prognosis assessment of acupuncture treatment for patients with chronic low back pain.The construction,verification and evaluation methods of acupuncture clinical prediction models will help to promote the development and visualization of individualized acupuncture prediction models. |
PDF Full Text Request