The objective of our study was to assess the efficacy of external qigong therapy (EQT), a traditional Chinese medicine practice, in reducing pain and improving functionality of patients with knee osteoarthritis (OA). One hundred twelve adults with knee OA were randomized to EQT or sham treatment (control); 106 completed treatment and were analyzed. Two therapists performed EQT individually, five to six sessions in 3 weeks. The sham healer mimicked EQT for the same number of sessions and duration. Patients and examining physician were blinded. Primary outcomes were Western Ontario MacMaster (WOMAC) pain and function; other outcomes included McGill Pain Questionnaire, time to walk 15 m, and range of motion squatting. Results of patients treated by the two healers were analyzed separately. Both treatment groups reported significant reduction in WOMAC scores after intervention. Patients treated by healer 2 reported greater reduction in pain (mean improvement -25.7 +/- 6.6 vs. -13.1 +/- 3.0; p < 0.01) and more improvement in functionality (-28.1 +/- 9.7 vs. -13.2 +/- 3.4; p < 0.01) than those in sham control and reduction in negative mood but not in anxiety or depression. Patients treated by healer 1 experienced improvement similar to control. The results of therapy persisted at 3 months follow-up for all groups. Mixed-effect models confirmed these findings with controlling for possible confounders. EQT might have a role in the treatment of OA, but our data indicate that all EQT healers are not equivalent. The apparent efficacy of EQT appears to be dependent on some quality of the healer. Further study on a larger scale with multiple EQT healers is necessary to determine the role (if any) of EQT in the treatment of OA and to identify differences in EQT techniques.