Conclusions

This evidence report assessed the distribution of published studies of mind-body therapies in general and performed a more detailed review of mind-body therapies for gastrointestinal disorders.

With regard to mind-body therapies in general, we identified a large body of literature.

• The most common conditions for which studies of mind-body therapies have been published are:
  > Neuropsychiatric
  > Head/ear, nose, and throat
  > Gastrointestinal
  > Circulatory
  > Musculoskeletal
  The mind-body therapies that have been most commonly the subject of published studies are:
  > Biofeedback
  > Hypnosis
  > Relaxation
  > Behavioral
  > Cognitive

Our review supports the following conclusions regarding mind-body therapies for gastrointestinal disorders:

• The controlled trials of mind-body therapies have substantial methodologic shortcomings that affect the internal validity of the results.
• About 75 percent of the controlled trials of mind-body therapies are not powered sufficiently to detect even large therapeutic benefits.
• There is no evidence to support the efficacy of biofeedback therapy for children. Seven controlled trials have all failed to demonstrate statistically significant benefits of biofeedback compared to other therapies.
• There is limited evidence (at least one trial whose quality score characterized it as “good” that reported statistically significant benefits and the majority of other studies also report statistically significant benefits) to support the efficacy of the following mind-body therapies:
  > Relaxation
  > Behavioral
  > Cognitive
  > Guided imagery
  This level of evidence falls short of conclusive proof of efficacy, but does suggest that these therapies are the most promising for further, high quality studies assessing efficacy and effectiveness.
• The methods shortcomings of studies reporting beneficial effects of hypnosis preclude drawing conclusions about its efficacy. Although all but one of the studies of hypnosis reported statistically significant benefits, none of the studies had a “good” quality score and therefore we can only conclude that more research is needed.
• There are mixed results regarding the use of biofeedback in adults. There was one study of biofeedback in adults whose quality score characterized it as “good,” and this study did not report any benefits of biofeedback. Three other studies of lesser quality were additionally found, one also reporting no benefit and two others reporting some benefit of biofeedback. These disparate results will need to be clarified by additional research.
• There were no reports of the mind-body therapies being associated with any harms.

Limitations of the Review

Four factors prevent any stronger conclusions being drawn from the data other than the ones presented here.

(1) The low number of small (in terms of sample size) studies for some conditions and some therapies supports the hypothesis that publication bias occurred. This research situation, in which there exists a small number of relatively small and low-powered published studies, might occur because the small sample size and nonsignificant results deter publication of other studies. Because of the heterogeneity of the studies, and the lack of a common outcome measure and sufficient data on means and variance measures of some sort, we are not able to do a test for publication bias that requires a common statistic across studies. Therefore, we cannot address this issue statistically.

(2) The poor quality of the controlled trials may be exaggerating the estimate of the effect of the mind-body therapies. Empiric evidence has shown that studies whose quality is characterized as “poor” report substantially increased beneficial effects. When combined with the concern about publication bias, above, this means our review may overestimate the efficacy of mind-body therapies.

(3) The third factor also stems from our inability to calculate a common statistic across studies that would allow a more quantitative comparison among and synthesis of the studies. We present our results in the form of a qualitative and narrative review that discusses each individual study similarly and includes a systematic presentation of study characteristics in our evidence table. Some of our discussion focuses on the number of studies that showed beneficial results versus the number that did not in certain settings. This might be considered to be an application of the meta-analytic method of “vote counting” in which each study is given equal weight in the consideration of the results. This method is often all that is possible in a synthesis when the individual studies do not have comparable outcomes and are heterogeneous. However, this method has been criticized on several dimensions (Bushman, 1994). Under a strict vote counting method, all studies are counted equally, regardless of sample size or quality. We tried to minimize this by placing more emphasis in our summary on those studies characterized as “good” quality. The method also does not produce a clinically interpretable result, and the strength of the conclusion is not apparent, unlike in a quantitative synthesis, which produces a pooled effect size and confidence interval. The last problem is that vote counting has been shown to have low statistical power to detect small effect sizes (Hedges and Olkin, 1980). Given these drawbacks of the method, we have attempted in our discussion of the results to consider each study individually rather than just to present the overall number of studies in each category, and we have presented our caveats about the method as well.

(4) The conclusions of this review are based on the limitation that arises from focusing on controlled trials only. Although this is methodologically a good decision, it may be that review of studies that do not achieve this status would result in the identification of promising areas for further research. However, as the controlled trials identified do not establish the efficacy of any of the mind-body therapies, sufficient areas of promising research are already identified without having to assess studies of lesser interval validity with respect to efficacy.

Future Research

Our review of the literature has identified several areas where future research into mind-body therapies could benefit from our findings. We have noted recurring shortcomings in the design and execution of published studies that weaken the conclusions that can be drawn from the studies. Future research needs to be better designed and implemented. Studies need to enroll adequate numbers of well-defined, clinically homogeneous populations. They also need to compare the mind-body interventions both to other potentially effective therapies and to a convincing control, if possible. Furthermore, future studies should employ randomization, use blinding where feasible, and measure outcomes that are meaningful to patients and that can be reliably assessed. Ultimately, only those studies with a control group for comparative analysis can address the question of the efficacy of mind-body therapies.

In addition to needing improved design, future trials of mind-body therapies need better reporting. This would aid interpretation and the application of the research results. Two types of information are essential: a clear description of the research design, particularly of the control and comparison groups, and a detailed description of the patient sample. It is frequently difficult to tell from published studies how comparable the patient populations are, not only demographically but also clinically, in order to interpret the diagnosis and prognosis.