Issues in Planning a Placebo-controlled Trial
of Manual Methods:   Results of a Pilot Study

This section was compiled by Frank M. Painter, D.C.
Send all comments or additions to:

FROM:   J Altern Complement Med 2002 (Feb); 8 (1): 21–32 ~ FULL TEXT


Cheryl Hawk, D.C., Ph.D. • Cynthia R. Long, Ph.D. • Robert Reiter, M.D.
Charles S. Davis, Ph.D. • Jerrilyn A. Cambron, D.C., M.P.H. • Roni Evans, D.C., M.S.

Palmer Center for Chiropractic Research,
Davenport, IA 52803, USA.

OBJECTIVE:   There are fundamental differences between the administration of medications and the application of manual procedures, such as those used by chiropractors. The objective of this study was to gather preliminary information on how to address these differences in the design of a multisite, randomized placebo-controlled trial of chiropractic care for women with chronic pelvic pain (CPP).

DESIGN:   Pilot study for a multisite, randomized, placebo-controlled clinical trial.

SETTING:   Three chiropractic research clinics in the midwest United States.

SUBJECTS:   Thirty-nine (39) women with CPP of at least 6 months' duration, diagnosed by board-certified gynecologists.

INTERVENTIONS:   The active intervention consisted of the chiropractic technique, lumbar spine flexion-distraction, combined with manual Trigger Point Therapy. The placebo intervention consisted of a sham chiropractic procedure performed with an instrument combined with effleurage (light massage).

OUTCOME MEASURES:   The primary outcome measure was the change in the Pain Disability Index (PDI) from baseline to the end of treatment (6 weeks), assessed by group and site. If the change score was in the same direction at all sites, the results were to be combined to estimate treatment effect size.

RESULTS:   Patient characteristics were similar to those of patients with CPP in other studies. Recruitment methods, particularly in respect to the eligibility criteria and screening protocols, would require modification in order to recruit an adequate sample for the planned randomized controlled trial. Clinicians followed standardized procedures with apparently minimal deviation, patients in both groups were satisfied with their care and blinding appeared to be successful. PDI change scores were not consistent across sites and so results were not combined and overall treatment effect sizes were not estimated.

CONCLUSION:   The technical and personnel resources required to achieve adequate standardization of procedures at multiple sites may make a placebo-controlled trial unfeasible, given our current lack of knowledge about the active agent in manual chiropractic procedures.

From the FULL TEXT Article:


This was a pilot study; it was not designed to investigate the efficacy of chiropractic care for symptoms of CPP. It was designed to provide the investigators with information on whether it is feasible to move forward with plans for an RCT on that topic.

      Patient characteristics

Comparing this study to our previous singlegroup study (Hawk et al., 1997), baseline PDI scores and symptomatology were similar. Antidepressant use in this study compared to the earlier one was also similar, whereas hormone use was considerably less (15% versus 37%), with prescription analgesic use (10% versus 5%) and nonprescription analgesic use (54% versus 63%) being not too dissimilar. Baseline depression as measured by the BDI was in the mild category for this study and moderate for the previous one.

Compared to characteristics of patients in studies of medical care for women with CPP, our patients were in the same age range and were similar in terms of low-back pain and dysmenorrhea, but there appeared to be fewer with a history of sexual/physical abuse (Peters et al., 1991; Reiter, 1990; Reiter and Gambone, 1990; Reiter & Milburn, 1992). Patients in our study did not report much use of supplements or herbs. It is possible that patients interested in CAM therapies were screened out because of the exclusion of those with recent chiropractic experience. Generally, it appears that our patients were fairly similar to those in our previous study and to medical CPP patients.

      Recruitment issues

It appears that we cannot recruit enough patients for a larger study using this study’s parameters, particularly in respect to the eligibility criteria and screening protocols. Further study is needed on the necessity of excluding patients with recent chiropractic experience. Also, exploration of methods to expedite the screening process and develop referral relationships with local gynecologists is warranted.

      Delivery of standardized procedures and record keeping of procedures used

The chiropractic clinicians followed strict protocols for delivering standardized procedures without a great deal of variation and maintained more detailed records about specific treatment procedures than are often required in clinical studies, facilitating in-depth analysis. Even though no differences were noted in patient satisfaction or perception of group assignment, patients did note differences between clinicians’ treatments, usually in the case where there were large differences in the experience level of the clinicians. Because outcomes were ambiguous, it behooves us to develop better ways for future multisite studies to ensure that clinicians are actually delivering the same procedures in terms of biomechanical forces and techniques.

      Identity of the active agent

One would expect the active group to show greater improvement than the placebo group, if the procedures used were, in fact, “active” and “placebo” agents. However, this pattern was only discernable at site 1, where the primary clinician was the most highly experienced in this technique but also was the only one who varied the active procedure in response to individual patient needs. At site 2, patients in both groups improved approximately equally, and at site 3, it appeared that patients in the placebo group improved more, although the sample sizes were extremely small. These ambiguous results indicate a need to ensure that the active and placebo procedures are applied according to strict biomechanical parameters in a multisite study, in order to eliminate variation introduced by use of different clinicians.

      Success of blinding

From patient report of satisfaction with the intervention and belief about group assignment, the placebo intervention was believable. The question that must be answered before conducting a placebo-controlled trial, however, is related to the issue of the identity of the active agent: Was the placebo intervention, in fact, a placebo? Further investigation of the biomechanical forces producing treatment effects is certainly warranted.


The technical and personnel resources required to achieve adequate standardization of procedures at multiple sites may make a placebo-controlled trial unfeasible, given our current lack of knowledge about the active agent in manual chiropractic procedures. It might be more efficient to reverse the traditional order of experimentation used for pharmaceuticals, which begins with safety, proceeds to efficacy and finally to effectiveness. Because chiropractic—a CAM profession using manual methods for more than 100 years—can scarcely be considered in the same category as a newly developed medication, it might be reasonable to first investigate effectiveness. If chiropractic care that is provided by experienced chiropractors who are allowed to use their best clinical judgment of how to apply the procedures is documented in such studies to improve patient outcomes compared to standard medical care alone, more in-depth and controlled studies would then be warranted to identify specific aspects of that gestalt of care that are most responsible for the outcomes or if there are certain subpopulations of patients who benefit most from them.

Return to PLACEBOS


Since 3-01-2002

                  © 1995–2024 ~ The Chiropractic Resource Organization ~ All Rights Reserved