Cherkin 2009.

Study characteristics
Methods	Four‐arm RCT Number analyzed/randomized: 583/641 Statistical analysis: ITT analysis; Chi² and ANOVA test; power analysis Funding source: NIH Cooperative Agreement (U01 AT 001110) Ethical approval and informed consent obtained
Participants	Participant recruitment: mail, advertisements, and newsletters Setting: two research clinics in western Washington and northern California in USA Inclusion criteria: 1) participants planned to continue enrolment in health plan; 2) 18 to 70 years; 3) at least one primary care visit for back pain within the past 3 to 12 months; 4) nonspecific and uncomplicated LBP; 5) physician was willing to have patients included in the study; 6) lived within 45 min travel time from study clinic Exclusion criteria: 1) specific causes of back pain (e.g. cancer, fractures, spinal stenosis, infections); 2) complicated back problems (e.g. sciatica, prior back surgery, medicolegal issues); 3) possible contraindications for acupuncture (e.g. coagulation disorders, cardiac pacemakers, pregnancy, seizure disorder); 4) conditions making treatment difficult (e.g. paralysis, psychoses); 5) conditions that might confound treatment effects or interpretation of results (e.g. severe fibromyalgia, rheumatoid arthritis, concurrent care from other providers) Age (mean): 47 Gender (female): 62% Pain duration (reported ≥ 1 yr): 2/3 participants Pain intensity (mean): NR
Interventions	1) GROUP 1: acupuncture (standardized) Acupuncture points: bilateral BL23, BL40, KI3, Ashi point, and Du3 (8 points in total) Depth: NR De Qi: 20 min treatment with 10 min stimulation by twirling the needles just prior to needle removal to elicit De Qi Sessions: 10 sessions (20 min, 2/week for 3 weeks + 1/week for 4 weeks) Acupuncturist experience: six licensed acupuncturists with experience of 4 to 19 years 2) GROUP 2: individualized acupuncture The treatment was prescribed by an acupuncturist diagnostician at each visit. Average of 10.8 needles (range, 5 to 20) retained for a mean of 18 min (range, 15 to 20 min) were used. Seventy‐four distinct points were used, half on the bladder meridian, which includes points on the back and legs. No constraints on depth or manipulation. Treatment schedule and others were same as Group 1. 3) GROUP 3: sham acupuncture The acupuncturists simulated the insertion and the removal of needles at the eight standardized acupuncture points. A toothpick in a needle guide tube was used, which was found to be a credible acupuncture treatment by acupuncture‐naive participants with back pain. All acupuncture points were stimulated at 10 min with the tip of a toothpick and again at 20 min just before they were 'removed'. Participants were asked to wear eye masks and lie prone, with their heads in a face cradle. Same treatment schedule as Group 1 4) GROUP 4: usual care Participants received usual care that they and their physicians chose (mostly medications, primary care, and physical therapy visits). Co‐intervention: a self‐care book with information on managing flare‐ups, exercise, and lifestyle modifications Duration of treatment: 7 weeks Duration of follow‐up: 52 weeks after beginning of treatment
Outcomes	1) Back‐specific function status: RMDQ (0 to 23), lower values better 2) Quality of life: sub‐scale of SF‐36 in physical health and mental health; higher values better 3) Pain‐related disability: cutting down on activities for more than 7 days in the past month (%); missed work or school for more than a day in the past month (%); no accurate data provided for sham and true acupuncture Assessment times: 8, 26, and 52 weeks after beginning of treatment Costs: NR (the reported total costs excluded acupuncture spending) Adverse effects: 12 of 477 participants in real acupuncture groups reported pain, dizziness, and back spasms.
Notes	Conclusion: "Acupuncture was found effective for chronic LBP, tailoring needling sites to each participant and penetration of the skin appear to be unimportant in eliciting therapeutic benefits." Language: English For results, see comparisons 1, 3, and 5. The data for pain‐related disability were not extracted for meta‐analysis, because the study only reported a range of 5% to 10% in the acupuncture group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A centrally generated variable‐sized block design
Allocation concealment (selection bias)	Low risk	"The random assignments cannot be viewed in advance and cannot be changed after randomisation.”
Blinding of participants (performance bias) All outcomes	Low risk	Participants in the three acupuncture groups were blinded, with credibility assessed. Participants were asked to wear eye masks and lie prone with their heads in a face cradle. Participants rated the acupuncture and simulated acupuncture treatments almost identically with regard to provider skills and caring.
Blinding of personnel /care providers (performance bias) All outcomes	Unclear risk	Although masking of acupuncturist is not possible, this design employed a diagnostician acupuncturist who made the diagnoses and determined which acupuncture points should be applied. An experienced therapist acupuncturist then delivered the assigned treatments, interacting minimally with participants and the diagnostician, who remained masked to treatment. The diagnostician acupuncturists rated the acupuncture and simulated acupuncture groups very similarly with regard to apparent efficacy and likelihood of receiving individualized treatment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Assessor was blinded, and the subjective outcomes were reported by masked participants. Outcomes were measured using computer‐assisted telephone interviews by interviewers masked to treatment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Dropout rates (5% at 8 weeks and 9% at 52 weeks); the reasons were given and acceptable.
Intention‐to‐treat‐analysis (attrition bias)	Low risk	The analyses were conducted according to their randomisation allocation, regardless of treatment compliance.
Selective reporting (reporting bias)	Low risk	All outcomes in the protocol were reported.
Group similarity at baseline (selection bias)	Low risk	Similar baseline characteristics
Co‐interventions (performance bias)	Unclear risk	Participants in the usual care group were visited by a physical therapist twice compared to acupuncture groups. Unclear whether the co‐interventions were different across the groups
Compliance bias (performance bias)	Low risk	Comparable compliance rates were more than 80% across the groups.
Timing of outcome assessments (detection bias)	Low risk	Same assessment time
Other bias	Low risk	Not identified