Validation of GAITRite and PROMIS as high-throughput physical function outcome measures following ACL reconstruction

Abstract

New healthcare demands for quality measures of elective procedures, such as anterior cruciate ligament (ACL) reconstructive surgery, warrant the establishment of high throughput outcomes for high volume clinics. To this end, we evaluated the PROMIS and GAITRite as physical function outcome measures to quantify early healing and post-operative complications in 106 patients at pre-operative and 3, 10, 20 and 52 weeks post-ACL reconstruction with bone-tendon-bone autograft, and compared the results to the current IKDC validated outcome measure. The results showed that both PROMIS and GAITRite were significantly quicker to administer versus IKDC (p<0.0001). Additional advantages were that PROMIS and GAITRite detected a significant decrease in physical function at 3 weeks post-operative, and a significant improvement at 10 weeks post-operative versus pre-operative (p<0.001), which were not detected with IKDC. GAITRite was limited by a low ceiling that could not detect improvement of physical function beyond 20 weeks, while both PROMIS and IKDC detected significant improvement out to 52 weeks post-operative (p<0.001). Linear regressions demonstrated a significant relationship between IKDC and PROMIS, with a combined correlation value of 0.8954 (p<0.001) for all time points. Finally, ROC curve analysis demonstrated that PROMIS is a diagnostic test for poor outcomes.

Keywords: ACL reconstruction; PROMIS; gait; physical function.

Figure 1. Longitudinal Assessment of Physical Function Outcome Measures

Average PROMIS physical function scores (A). PROMIS scores are T-scores relative to the reference “normal” population, indicating a mean of 50 and a standard deviation of 10. A significant decrease in perceived physical ability associated with recovery was found 3 weeks after surgery as well as significant functional gains over baseline at 10, 20, and 52 week time points. Averaged IKDC scores (B). Significant increases in perceived physical ability were found at 20, and 52 weeks, although those changes associated with surgical recovery at 3 weeks (p=0.999) and 10 weeks (p=0.122) were not found to be significant. Average gait velocity recorded by GAITRite (C). A Significant decrease was found 3 weeks after surgery, followed by a significant increase at 20 weeks, no further significant velocity increase was found at 52 weeks. The Average FAP scores generated by GAITRite analysis are shown (D). FAP scores indicate an increased risk of falls at baseline (Figure 1D), a significant loss of stability was found 3 weeks post-surgery, and significant gain in function over baseline was found at 10 weeks (p<0.001). The FAP however shows no further improvement beyond the 10 week time point (Figure 1D). Average stance length as a percentage of the gait cycle for each limb (E). The limb with a surgically reconstructed ACL was in stance phase significantly less than the uninjured limb at baseline as well as 3,10, 20 weeks after surgery. At 52 weeks this difference was no longer significant. Error bars indicate standard error of the mean, Asterisk indicate a significant difference from baseline (*p<0.001).

Figure 2. Linear Regression Analysis

Linear regression plots are shown for each time point illustrating the correlation of IKDC with PROMIS (2A), gait velocity (2B), and FAP (2C), associated correlation coefficients as well as those for all time points are given along with their respective p-values in Table 3. A significant correlation between IKDC and PROMIS was found at each time point with a combined correlation value of 0.8954 (p<0.001) for all time points. A significant correlation was found between IKDC and gait velocity at the baseline pre-surgical visit, but at no other individual time point, with a combined correlation value of 0.4185 (p=0.0012). No significant correlation was found between IKDC and FAP at any individual time point, however the combined correlation value was 0.3093 (p=0.0192).

Figure 3. Longitudinal Assessment of Poor Outcomes

A subset of patients with poor outcomes were identified and longitudinally compared to the rest of the cohort. Average PROMIS scores (A), FAP scores (B), gait velocity (C) for each group are shown. Significantly lower scores (p=0.004, p=0.026, p=0.002, p<0.001) were found at baseline, 3, 10, 20, 52 weeks respectively (A). Neither velocity (B) nor FAP (C) showed significant differences at any one time point. Error bars show standard error of the mean; Asterisks indicate a significant difference from baseline.