Background/Purpose: Physical therapy (PT) is a key component of treatment for knee osteoarthritis (OA). There is a high demand for PT services in many healthcare systems, resulting in a need for evidence-based models for delivering PT in an efficient manner.  A group-based approach to PT can extend services to more patients with lower staffing requirements than typical individual PT.  The objective of this trial was to compare the effectiveness of a group-based PT program (GPT) with usual individual PT (IPT) for knee OA. 

Methods: 320 patients with knee OA at the VA Medical Center in Durham, NC (mean age = 60, SD=10; 88% male; 58% non-white) were randomized to either GPT or IPT.  GPT included six 1-hour sessions, every other week, co-led by a physical therapist and PT assistant, with 8 participants per group. IPT, modeled after typical outpatient PT care for knee OA at the Durham VAMC, included two 1-hour visits with a physical therapist, 2-3 weeks apart.  Both PT interventions included a home exercise program, as well as individual evaluations of functional limitations and needs for braces or assistive devices. GPT sessions also included exercise sessions supervised by the PT assistant. The primary outcome was the Western Ontario and McMasters Universities Osteoarthritis Index (WOMAC; range 0-96, higher scores indicate worse symptoms), and the secondary outcome was objective physical function (Short Physical Performance Battery; SPPB, range 0-20, higher scores indicate better function); both were assessed at baseline and 12-weeks, and WOMAC was also assessed at 24-weeks.  Linear mixed models were used to assess the difference in improvement in outcomes between arms, adjusting for clustering of group sessions within the GPT arm.

Results: The median numbers of sessions attended for GPT and IPT were 5 (out of 6 possible) and 2 (out of 2 possible), respectively. At 12-week follow-up, WOMAC scores were 2.7 points lower in the GPT group vs. IPT [95% confidence interval (CI) = -5.9, 0.5; p=0.10], indicating no meaningful difference in improvement between arms. However, mean total WOMAC scores declined -4.5 points from baseline across both arms combined [95% CI =-6.8,-2.2; p=0.0001], indicating meaningful improvement.  Similarly, for the WOMAC pain and function subscales and SPPB scores there was improvement across both arms at 12-weeks (p<0.0001, p=0.001, and p=0.02) but no difference in improvement between arms (p=0.19, p=0.12, and p=0.37). At 24-week follow-up, WOMAC scores across both arms were 3.1 points lower compared to baseline [95% CI = -5.4, -0.7; p=0.01], indicating some sustained improvement in both groups, with no difference between groups (p=0.45).

Conclusion: Results of this study confirm that PT improves pain and functional outcomes in patients with knee OA.  Outcomes did not differ substantially between GPT and IPT arms, suggesting that either is an effective means of delivering PT services for knee OA.  The GPT approach in this study required less overall staff time per patient to deliver, and it could provide services efficiently to larger numbers of patients.  Therefore it should be considered as a viable model for health systems to provide this service to patients with knee OA.

---

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/randomized-clinical-trial-of-group-vs-individual-physical-therapy-for-knee-osteoarthritis/