In patients with cancer, the demand for complementary and alternative medicine (CAM) is considerable. Unfortunately, however, for many of these interventions there is a lack of evidence for efficacy, effectiveness and safety in patients with cancer. This review focuses on the prospective, randomised, controlled trial (RCT) as a tool for evaluating CAM. Although a number of difficulties and limitations are acknowledged, the RCT will continue to be the gold standard for evaluating the efficacy, effectiveness and safety of CAM. Developments in clinical trial methodology and in psychosocial oncology have made it more appropriate and feasible to evaluate CAM using RCT methodology. Two different kinds of RCTs are now accepted as valid, namely explanatory and pragmatic trials. The latter does not necessarily require that the patient or the therapist is 'blind' to the treatment being given. Furthermore, pragmatic trials can be designed to take patient preferences into account. A number of practical issues are discussed, including the choice of comparator or control interventions, ways of assessing the effects of individual differences, minimising therapist variability, the problem of finding acceptable inclusion-exclusion criteria and the assessment of treatment outcome. A number of randomised, controlled trials have demonstrated the efficacy, effectiveness and safety of various complementary and alternative interventions (the Cochrane Data Base has now established a CAM field). The publication of positive results from randomised trials of complementary interventions that have not yet been studied using this methodology would do a great deal to alleviate the scepticism of conventional practitioners towards these types of CAM and would facilitate further the integration of complementary and conventional interventions.