Pharmacotherapy, Non-pharmaceutical therapy, Other acupuncture type Outcomes Effectiveness evaluation Indispensable NRS, ODI, Effective Rate Safety evaluation Indispensable Adverse events Follow up Indispensable Disease status (Recurrence, Lower back pain related surgery, Seeking other treatments) CT Computed tomography, NA Not applicable, NRS Numeric rating scale, ODI Oswestry disability index, TCM Traditional Chinese Medicine Wei et al. BMC Musculoskeletal Disorders (2019) 20:488 Page 5 of 10 models will be employed to adjust the propensity score and baseline characteristics. Once matched, change in ODI and NRS between-group at different observation points will be calculated by paired t-test. Thirdly, logistic regression models will be used to evaluate the magnitude of impacts contributed by baseline factors on acupuncture effectiveness for CLBP patients. Descriptive analysis will be used for variables such as the recruited research center,s basic characteristics of doctors and patients, initial diagnosis and intervention details. Quantitative data (Height, Weight, NRS, ODI, etc.) will be described by mean, standard deviation, median, maximum and minimum values. The enumeration data (Acupuncture style, Acupoint selection principle, Operation of acupoints, Effective rate etc.) and ranked data (Educational background, Physician title) will be described by numbers, percentages, rate and constituent ratio. If necessary, parametric methods or alternate nonparametric methods will be selected appropriately. We will adopt the propensity score matching of patients at 1:1 to imitate an RCT. Baseline confounder will be equilibrated and matched on variables such as demographic data, medical history, duration, intensity, back function. The comparative analysis will be conducted after the baseline unevenness adjustment. Propensity score will match comparison group, such as Traditional Chinese acupuncture style vs Microacupuncture style (Scalp acupuncture, Ear acupuncture, Wrist and ankle acupuncture, Eye acupuncture, Umbilical acupuncture, Abdomen acupuncture), Local acupoint selection vs Non-local acupoint selection and Single Acupuncture vs Combined therapy etc. Hierarchical models based on the pain grading, medical center, age group, gender, TCM syndromes and classification of diseases will be taken into account to explore the important factors affecting the effectiveness of acupuncture for CLBP [43–45]. Conditional logistic regression models adjusting by different variables to identify predictors of long-term effectiveness presented by ODI and NRS or recurrence at the12-week follow-up. Quality management Management of participants Before acupuncture physicians are confirmed to be enrolled, they need to apply for an individual account on the registration platform. Only after the subject research management unit has reviewed their information and approved can they join the project. All organizational personnel, platform developers and acupuncturists are required to participate in standardized training and must acquaint with an investigator’s manual. Data monitoring Considering the differences in the filling of registry information between different doctors, standardizing and integrating of diagnosis is the main concern. Data Monitoring will conduct online or on-site quality inspection on the registered information every month, and timely feedback the problems. The concern includes the accuracy, integrity, authenticity and timeliness of the registry. Once problems confirmed, verification or traceability will be conducted timely. Meanwhile, problems existing in the research process will be posted and timely solved via network or telephone. For registered contents that Timetable for data collection Data elements Therapeutic Session Follow-up Period Minimum a The number of acupuncture treatments shall be no less than three times, b No less than three follow-up visits Wei et al. BMC Musculoskeletal Disorders (2019) 20:488 Page 6 of 10 are not clearly expressed, such as contradictory information, important data missing. Once verified, online modification will be carried out and recorded for later reference. Data quality management During the data collection process, the platform can proceed with template management, according to the time verification, integrity, validity, and logical verification. Furthermore, the ID number is the unique identification code of the patient on the platform to avoid duplicate entry. The corresponding prompt interface will appear to reduce the wrong data and missing data to facilitate data quality. Preliminary manual verification of data will be conducted quarterly and timely feedback to the research team. A training conference on registry implementation and problem summary will be conducted every year to gradually summarize the experience. Therefore, the accuracy, fidelity and timeliness of data quality will be improved gradually. A data quality and safety oversight committee will be established to conduct regular assessments of data quality and safety, as well as of acupuncturists. Implement management of procedures, policies, procedures to ensure the quality of data within the registry and control the following errors. Entry errors: such errors can be avoided through prior data quality checks (such