A comparison of treatment signatures of high and low performing physical therapists for patients with lower back pain: analysis of spine care from a physical therapy outcomes registry.

INTRODUCTION: Lower back pain (LBP) is the most common orthopedic complaint in the United States. Physical therapy is recommended as a conservative, non-pharmacological intervention for LBP. While it is thought that skill level and effectiveness of physical therapists differ, there is little understanding regarding characteristics that distinguish high and low performing physical therapists. The purpose of this study was to compare differences in care delivery, termed treatment signatures, between high and low performing physical therapists previously differentiated by a risk-adjusted performance measure. METHODS: Using previously published methodology, 1,240 physical therapists were classified as "outperforming", "meeting expectations", or "underperforming" relative to predicted change in Modified Low Back Pain Disability Questionnaire (MDQ) across patients receiving care for LBP. Patients were divided into quartiles of baseline disability per initial MDQ. Two-way analyses of variance were used to compare billed (1) active, exercise-based units per visit (UPV), (2) manual therapy UPV, (3) modality UPV, and (4) the combination of active and manual therapy UPV (broadly termed skilled UPV) by performance cohort and baseline patient disability quartile among physical therapists deemed "outperforming" and "underperforming". Tukey's post hoc tests established mean differences with 95% confidence intervals. RESULTS: Physical therapists that "outperformed" (n=120; 17,404 patients) used more active UPV (mean difference (diff) = 0.1, p<.001), manual therapy UPV (mean diff = 0.2, p<.001), and skilled UPV (mean diff = 0.3, p<.001), and less modality UPV (mean diff = 0.1, P < 0.001) than those that "underperformed" (n=139; 21,800 patients). Tukey's post hoc tests showed that while differences in care were negligible in patients with low baseline disability, the highest performing PT cohort delivered skilled (0.4 UPV), active (0.2 UPV), and manual therapy (0.2 UPV) UPV at a significantly higher mean rate in patients with the highest baseline disability. CONCLUSIONS: Clinically effective physical therapists incorporated a treatment signature that included a consistent blend of skilled active and manual therapy interventions that was distinct from lower performing physical therapists. While group mean differences were relatively small, a consistent pattern emerged in which high performing physical therapists maintained a high level of skilled, one-on-one interventions across their entire caseload-while their lower performing counterparts significantly decreased use of the same interventions as baseline disability increased. These differences highlighted a treatment signature that was associated with clinically important improvements for patients with greater baseline disability. Future guideline recommendations should consider the importance of baseline disability and the consistent application of skilled active and manual therapy interventions.