Background/Purpose: The clinical effectiveness of supervised exercise therapy for individuals with knee osteoarthritis (KOA) is well known. However, it is unclear whether participation in a supervised therapy program impacts downstream utilization of health services. The purpose of this study is to determine the impact of four different supervised physical therapy (PT) programs on KOA-related health service utilization over a 2-year period.

Methods: This is a secondary analysis of data from a 2-year multisite randomized clinical trial of 300 individuals with KOA per ACR criteria. Participants were randomized to one of four PT strategies: (1) 12 visits of exercise therapy alone; (2) 9 visits of exercise therapy + 3 booster sessions spaced across 12 months; (3) 12 visits of exercise + manual therapy; (4) 9 visits of exercise + manual therapy + 3 booster sessions. Participants were queried at baseline, 1 year, and 2 years regarding 12-month utilization of health services commonly accessed by those with KOA (Table 1). Chi-square analyses compared health service utilization at baseline vs. 1 year and baseline vs. 2 years for the full cohort and for subgroups who did/did not receive manual therapy and booster sessions. Logistic regression was used to determine if treatment group allocation predicted health service utilization after adjusting for age, race, and baseline physical function.

Results: Across the full cohort, statistically significant reductions in utilization of many health services were observed following participation in the study (Table 1). Use of non-opioid and opioid pain medications for the knee reduced substantially over the two-year period. Reductions in knee joint injections and visits to physicians (especially primary care physicians) also significantly dropped over the study period. Logistic regression analyses revealed that treatment group allocation was generally not a significant predictor of health service utilization. Individuals who received booster sessions were less likely to have knee imaging at one year (adjusted OR 0.51; 95% CI 0.29-0.92). Those who received manual therapy were less likely to have had recent imaging at baseline (adjusted OR 0.59, 95% CI 0.37-0.94) and less likely to have visited a rheumatologist at 1 year (adjusted OR 0.50, 95% CI 0.26-0.95) and 2 years (adjusted OR 0.49, 95% CI 0.27-0.88). All other regression models comparing booster to non-booster groups and manual therapy to non-manual therapy groups did not achieve statistical significance.

Conclusion: Downstream utilization of many common health services reduced over two years for individuals with KOA participating in a randomized clinical trial of structured PT. The study did not include a control group so the reduction in utilization may not be a direct effect of participating in the PT intervention. Treatment group allocation was not a significant predictor of health service utilization.

Table 1. Utilization of Health Services Over Time: Overall Cohort

*indicates statistically significant difference from baseline utilization (p<.05)