“Randomised controlled trials of specific therapies…rarelydemonstrate any benefit in terms of quality of life” [1]. Thisis a statement from a recent article in this journal dealingwith complementary therapies for supportive cancer care. Itmade me ponder. Is this really true? I don't think so!Anyone who cares to look up the trial data will findhundredsofstudies—fartoomanytoreviewhereindetail.Solet's just look up (some of) the systematic reviews which areavailable, many from this journal. At least some encouragingevidence has emerged from systematic reviews of thefollowing treatments: (for a fuller discussion of the treat-ments involved, the reader is referred elsewhere [2]).& Acupressure for chemotherapy-induced nausea [3, 4]& Acupuncture for cancer pain [5]& Acupuncture for hot flushes in men with prostate cancer[6]& Acupuncture for hot flushes in women with breastcancer [7]& Aromatherapy for improving well-being of patientswith cancer [8]& Chinese herbal medicine for improving quality of life[9]& Co-enzyme Q10 for reducing the toxicity of cancerdrugs [10]& Exercise for fatigue and nausea [11, 12]& Exercise for improving quality of life [13]& Guided imagery for reducing anxiety [14]& Hypnosis for reducing cancer pain [15]& Massage for a wide range of symptoms and forimproving quality of life [16]& Music therapy for improving quality of life [17]& Qigong for improving quality of life [18]& Tai chi for improving psychological symptoms [19]I am, of course, not claiming that the evidence for all thelisted treatments is compelling, but I insist that there are atleast some encouraging trial data. In fact, some of theapproaches, e.g. acupuncture in the treatment of nausea orexercise in the treatment of asthenia, are today generallyaccepted in routine supportive care. Other treatments, e.g.Co-enzyme Q10 or Qigong, are clearly not well enoughresearched and necessitate confirmatory studies.I believe, we need to build on this evidence. We needmore clinical trials and we need better clinical trials. By“better”, I mean trials that successfully minimise bias.Certain study designs that are popular for pragmatic trialshave zero chance of generating a negative result, even if thetested intervention is devoid of any specific therapeuticeffects [20]. Therefore such investigations have the poten-tial to mislead us.The challenges for future research are numerous andcomplex. On the one hand, we need evidence that applies to“real life” situations in supportive care. On the other hand,we should thrive to establish cause and effect withreasonable certainty. Sailing between the “Scylla andCharybdis:” will not be easy, and a discussion about howthis might be best achieved is clearly beyond the scope ofthis short comment, particularly as this area is the subject ofmuch controversy within CAM research [21].A likely strategy could be to employ a mixed methodapproach of observational studies, pragmatic and fastidioustrials and, at the same time, be cognisant of the limitation ofeach of these approaches.