Acupuncture for chronic pain and depression in primary care: a programme of research

This comprehensive research program represented a decisive milestone in the scientific evidence on acupuncture, using innovative methods to resolve decades of uncertainty about its true efficacy. The study combined multiple methodological approaches, including the first individual patient data (IPD) meta-analysis in the field of acupuncture, network meta-analyses, and a pragmatic randomized clinical trial.

The IPD meta-analysis, conducted by the Acupuncture Trialists' Collaboration (ATC), analyzed raw data from 29 high-quality clinical trials involving approximately 18,000 patients with chronic pain. This method, superior to traditional meta-analysis, allowed for standardization across different analytical approaches, greater statistical power, and detailed investigation of patient characteristics that influence outcomes. The investigators included only studies with adequate allocation concealment, the most important methodological criterion for preventing bias.

The results unequivocally demonstrated that acupuncture is significantly more effective than both sham acupuncture and usual care for all chronic pain conditions studied (p < 0.001). For low back and neck pain, the effect size versus sham was SMD 0.23 (95% CI 0.13-0.33), while versus usual care it was 0.55 (95% CI 0.51-0.58). These findings definitively refute the claim that acupuncture is merely a 'theatrical placebo.'

The network meta-analyses compared acupuncture with other physical therapies for knee osteoarthritis, including 114 trials with 9,709 patients. Acupuncture emerged as one of the most effective therapies, significantly outperforming muscle-strengthening exercise and sham acupuncture when only higher-quality studies were analyzed. The cost-effectiveness analysis revealed that acupuncture is cost-effective when based on high-quality trials, with incremental cost-effectiveness ratios ranging from £7,000 to £14,000 per QALY.

The depression trial (ACUDep) enrolled 755 patients across 27 primary care practices in northern England, comparing acupuncture and counseling with usual care. Both interventions showed statistically significant reductions in PHQ-9 depression scores at 3 months: -2.46 points for acupuncture and -1.73 for counseling versus control. Importantly, approximately half of the participants also had comorbid pain, reflecting the common clinical overlap between these conditions. Acupuncture demonstrated cost-effectiveness with an incremental ratio of £4,560 per QALY.

The innovative methods included the development of new Bayesian techniques for the analysis of multiple individual patient data sets and the mapping of various quality-of-life instruments to the EQ-5D scale, enabling robust synthesis for economic analyses. The program also explored specific features of acupuncture (points used, frequency, duration) and different types of sham controls, providing valuable guidance for future clinical trials.

The clinical implications are substantial: the program provides the most robust evidence to date that acupuncture produces clinically relevant benefits beyond placebo effects for common chronic pain conditions. For managers and policymakers, the economic data support broader provision of acupuncture within the NHS, particularly given the limitations of conventional pharmacological treatments and the growing patient demand for nondrug options.

Limitations include heterogeneity in acupuncture protocols across studies, lack of long-term data for many nonpharmacological interventions for osteoarthritis, and, in the depression study, the absence of a control for nonspecific effects versus usual care (although time and attention were controlled for in the acupuncture versus counseling comparison). Future research should focus on the durability of effects, identification of predictors of response, and implementation of acupuncture in different health care settings.