There is no management regime for chronic fatigue syndrome (CFS) that has been found to be universally beneficial and no treatment can be considered a "cure". Patients with CFS may use complementary and alternative medicine (CAM). Our aim was to evaluate homeopathic treatment in reducing subjective symptoms of CFS. Using a triple-blind design (patient and homeopath blind to group assignment and data analyst blind to group until after initial analyses to reduce the possibility of bias due to data analyst), we randomly assigned patients to homeopathic medicine or identical placebo. One hundred and three patients meeting the Oxford criteria for CFS were recruited from two specialist hospital out patient departments. Patients had monthly consultations with a professional homeopath for 6 months. Main outcome measures were scores on the subscales of the Multidimensional Fatigue Inventory (MFI) and proportions of each group attaining clinically significant improvements on each subscale. Secondary outcome measures were the Fatigue Impact Scale (FIS) and the Functional Limitations Profile (FLP). Ninety-two patients completed treatment in the trial (47 homeopathic treatment, 45 placebo). Eighty-six patients returned fully or partially completed posttreatment outcome measures (41 homeopathic treatment group who completed treatment, 2 homeopathic treatment group who did not complete treatment, 38 placebo group who completed treatment, and 5 placebo group who did not complete treatment). Seventeen of 103 patients withdrew from treatment or were lost to follow-up. Patients in the homeopathic medicine group showed significantly more improvement on the MFI general fatigue subscale (one of the primary outcome measures) and the FLP physical subscale but not on other subscales. Although group differences were not statistically significant on four out of the five MFI subscales (the primary outcome measures), more people in the homeopathic medicine group showed clinically significant improvement. More people in the homeopathic medicine group showed clinical improvement on all primary outcomes (relative risk=2.75, P=.09). There is weak but equivocal evidence that the effects of homeopathic medicine are superior to placebo. Results also suggest that there may be nonspecific benefits from the homeopathic consultation. Further studies are needed to determine whether these differences hold in larger samples.