A Systematic Review of the Effect of Expectancy on Treatment Responses to Acupuncture

This systematic review, conducted by Colagiuri and Smith, investigated a fundamental question in acupuncture research: how patient expectancies influence treatment outcomes. The study arises from an intriguing observation — many randomized clinical trials show that both real and placebo acupuncture produce better results than no treatment, but there are often no significant differences between them. The investigators conducted a systematic search across four main databases (Medline, PsycInfo, PubMed, and the Cochrane Clinical Trials Register) from inception through December 2010, initially identifying 392 articles. After a rigorous selection process, nine independent studies met the inclusion criteria, which required assessment or manipulation of patient expectancies about acupuncture and its relationship with at least one health outcome.

The methodology of the included studies was quite heterogeneous, precluding meta-analysis. Five studies assessed expectancies through questionnaires, while four manipulated expectancies by providing different information to participants about the expected efficacy of the treatment. Most focused on pain-related conditions, both clinical and experimentally induced, with some investigating angina pectoris. The results were mixed but revealing.

Five of the nine studies found statistically significant effects of expectancies on at least one outcome, with three also finding suggestive evidence of interaction between expectancies and type of acupuncture (real versus placebo). Interesting patterns emerged: all three studies with experimentally induced pain found significant relationships, while only two of the six clinical studies did. Three of the four studies that manipulated expectancies found significant effects, compared with only one of the five that only assessed expectancies. Four of the five studies with electroacupuncture showed significant relationships, versus only one of the four with manual acupuncture.

The risk-of-bias assessment revealed methodological limitations in almost all studies, including inadequate sequence generation, unclear allocation concealment, and problematic dichotomization of expectancy variables. The clinical implications are substantial. The findings suggest that patient expectancies may be an important mechanism through which acupuncture exerts its effects, contributing to the placebo effect. This does not diminish the validity of acupuncture as a treatment, but indicates that the therapeutic relationship and communication about expectancies are important components of care.

The results also raise important methodological questions for future research, particularly about when and how to assess expectancies, whether to manipulate them experimentally, and how to avoid having the expectancy assessment itself influence the outcomes. The heterogeneity of expectancy assessment methods across studies — varying in wording, response scales, and timing — highlights the need for standardization. The limitations include the impossibility of meta-analysis due to methodological heterogeneity, potential publication bias favoring studies with positive results, and restriction to English-language publications. The authors emphasize that the three mechanisms proposed to explain the equivalent effects between real and placebo acupuncture are not mutually exclusive: acupuncture may involve multiple active components, placebo controls may not be inert, and expectancies may influence both real and placebo treatments.