von Brian Berman, Complementary Medicine Program, University of Maryland School of Medicine, Baltimore, MD

By definition, the successful practice of evidence-based medicine requires a strong foundation of quality research. However, it is important that this foundation encompass a wide variety of evidence types and study designs. Totality of evidence is essential in the field of acupuncture research because different study designs will often address different research questions or approach a question from different angles.

Due to the large quantity of research on acupuncture and painful conditions, the results of acupuncture research are summarized by painful and other conditions. While a number of reviews show positive results, many more report inconclusive findings. Most often, this is due to widespread methodological problems found in acupuncture trials.

The Federal Drug Administration has developed explicit guidelines for clinical research that are relevant to the development of clinical trials in acupuncture; following these guidelines can help to minimize the potential for methodological problems and therefore maximize the effectiveness of the research. Phase I involves the establishment of dose, outcomes, follow-up, and safety. Phase II tests hypotheses on a small scale and allows the researcher to develop appropriate control groups. Currently, acupuncture trials use: 1) delayed treatment, 2) non-acupuncture inert controls, 3) placebo acupuncture, 4) sham acupuncture, 5) active controls, and 6) combined controls. The choice of control group(s) is particularly important, for each type of control has advantages and disadvantages; moreover, it directly affects the research question being asked. Phase III is the final phase in clinical trials. These trials are large, definitive efficacy trials that are often multi-site.

The RCT method is advantageous because it allows the researcher to separate treatment effects from:

1. non-specific effects,
2. natural history of the disease,
3. regression to the mean, and
4. effects due to heterogeneous patient characteristics. However, the process is also limited because:

1. it does not test acupuncture in the context in which it is actually practiced,
2. it is difficult to randomize patients, and
3. a large number of subjects is often needed to detect treatment effects. Thus, while the RCT has a definite place in acupuncture research, it must be recognized that a totality of research is needed to obtain a broad understanding of the relevant issues in acupuncture research and treatment.