Tai Chi versus physical therapy for knee osteoarthritis: a cost-effectiveness analysis

Abstract

Purpose: Although few interventions effectively treat long-term pain and disability from knee osteoarthritis, some studies demonstrate that Tai Chi reduces symptoms. A single-blind randomized comparative effectiveness trial showed that Tai Chi yielded beneficial effects similar to those of a standard course of physical therapy in the treatment of knee osteoarthritis, so we aimed to examine the cost-effectiveness of Tai Chi versus physical therapy for knee osteoarthritis. Methods: Based on between group differences of Tai Chi versus physical therapy in longitudinal models of an intention-to-treat analysis of all randomly assigned patients at baseline, 12, 24 and 52 weeks with a treatment interaction, we sampled from the confidence interval distributions of the differences at each time point. Using standard microeconomic methods, we developed a simulation model to estimate healthcare resource utilization (hospitalizations, inpatient, outpatient, testing, medications and nutraceutical use) and cumulative effectiveness outcomes (WOMAC pain, WOMAC function, Global Visual Analog Scale and SF36 pain component score) over 12 weeks, 24 weeks, and a year. To account for uncertainty in the results and estimates, we performed Monte Carlo simulations incorporating the uncertainty surrounding all estimates. We estimated physical therapy costs at $100 per session for 6 weeks twice a week and group Tai Chi classes at $25 per session at twice a week for 12 weeks and accounted for the adherence observed in the study. Results: In the 1000 simulations for 12, 24 and 52 weeks, Tai Chi was always less expensive than Physical Therapy, with mean (SD) savings of $289 (47) at 12 weeks, $648 (60) at 24 weeks and $1668 (112) after 52 weeks. The analysis was most sensitive to the cost of PT or Tai Chi. At $25 per session for Physical Therapy, Physical Therapy was cost-saving at 12 and 24 weeks, but Tai Chi was cost saving at 52 weeks. At $60 per session for Tai Chi, Physical Therapy was cost-saving at 12 weeks, but Tai Chi became cost-effective at 24 and 52 weeks. In terms of effectiveness for 100 patients with knee osteoarthritis, Tai Chi improved WOMAC pain (0-500) by 50,925 points over 12 weeks, 220,159 points over 24 weeks and 368,642 points over 52 weeks. Similarly, Tai Chi improved WOMAC function (0–1700) by 93,339 points over 12 weeks, 279,460 points over 24 weeks and 439,979 points over 52 weeks. Tai Chi improved Global Visual Analog Scale (0-10) by 1899 over 12 weeks, 5770 over 24 weeks and 12,724 over 52 weeks. Lastly, Tai Chi improved SF36 Physical Component Score (0-100) by 13,589 over 12 weeks, 44,197 over 24 weeks and 84,294 over 52 weeks. Conclusions: Our analysis suggests that Tai Chi is cost-saving relative to physical therapy for knee osteoarthritis because it on average both reduces costs and improves outcomes. The results are relatively robust (consistent) when varying the cost of physical therapy and Tai Chi until extreme values were considered. The results suggest that reimbursement for Tai Chi as a treatment for knee osteoarthritis should be considered by health payers.