The authors’ experience of conducting clinical trials in homeopathy and analysing data from these has drawn attention to a fundamental problem with the interpretation of results from placebo controlled trials in homeopathy: It is not reasonable to assume that the specific effects of homeopathic medicine and the non-specific effects of consultations are independent of each other—specific effects of the medicine (as manifested by patients’ reactions) may influence the nature of subsequent consultations and the non-specific effects of the consultation may enhance or diminish the effects of the medicine.For clinical trials of homeopathy to be accurate representations of practice, we need modified designs that take into account the complexity of the homeopathic intervention. Only with such trials will the results be generalisable to homeopathic practice in the real world. The authors propose that comparative trials are a meaningful way of evaluating the effectiveness of homeopathic treatment.