Evidence for placebo effects on physical but not on biochemical outcome parameters: a review of clinical trials

BACKGROUND Recent reviews on placebo effects in clinical trials suggest that objective changes following placebo treatments may not exist or, at least, have been considerably overestimated. However, the possibility that yet unidentified subsets of parameters are responsive to placebo treatments has not been taken into account. Therefore, the aim of the present study is to examine the effects of placebo treatments on objectively measured outcome parameters by specifically focusing on peripheral disease processes. METHODS An initial dataset was collected from a MEDLINE search for placebo-controlled, randomized clinical trials. Trials with stable disease conditions were identified, and the effects of placebo treatments on peripheral outcome parameters were estimated by the changes from baseline in the placebo groups. An explorative data analysis was conducted in order to identify parameter classes with differential responsiveness to placebo treatments. A subgroup meta-analysis of a second dataset was performed to test whether the preliminary classification would also apply to placebo effects derived from the comparison of placebo groups with untreated control groups. RESULTS The explorative analysis of outcome parameters and strength of placebo effects yielded a classification into responsive "physical" versus non-responsive "biochemical" parameters. In total, 50% of trials measuring physical parameters showed significant placebo effects, compared with 6% of trials measuring biochemical parameters. A subgroup meta-analysis substantiated the differential response (physical parameters: n = 14, Hedges' pooled effect size g = 0.34, 95% CI 0.22 to 0.46; biochemical parameters: n = 15, g = 0.03, 95% CI -0.04 to 0.10). The subanalysis of the second dataset supported the classification and revealed a significant improvement for physical parameters (n = 20, g = 0.22, 95% CI 0.07 to 0.36) and a deterioration for biochemical parameters (n = 6, g = -0.17, 95% CI -0.31 to -0.02). CONCLUSION The results suggest that placebo interventions can improve physical disease processes of peripheral organs more easily and effectively than biochemical processes. This differential response offers a good starting point for theoretical considerations on possible mediating mechanisms, and for future investigations in this field.