Complementary medicine comprises many different disciplines and some, for instance, acupuncture, chiropractic, herbal medicine, and homoeopathy, are widely used. Patients with chronic pain show particular high rates of use. Efficacy and safety evaluations are warranted to sort out helpful interventions from unsafe practices. Thousands of randomized trials have evaluated the efficacy of the most popular complementary medicines but there is less information on their safety. However, important methodological problems in design and conduct, disagreement on the practice, variability of comparisons, and discordant interpretation of the trial results of studies have precluded reaching solid consensual conclusions. The challenge is to design studies in a way that fully satisfies methodological standards and that is consistent with everyday practice of complementary medicines including individualization of treatment and holistic aspects. The specific methodological problems of trials evaluating complementary medicines are discussed, with an emphasis on the necessity to use the full spectrum of trial designs, from very explanatory to completely pragmatic, in order to achieve a complete evaluation.