Abstract
Advances in algorithms developed with sensor data from smart phones demonstrates the capacity to passively collect qualitative gait metrics. The purpose of this feasibility study was to assess the recovery of these metrics following joint reconstruction. A secondary data analysis of an ethics approved global, multicenter, prospective longitudinal study evaluating gait quality data before and after primary total knee arthroplasty (TKA, n=476), partial knee arthroplasty (PKA, n=139), and total hip arthroplasty (THA, n=395). A minimum 24 week follow-up was required (mean 45±12, range 24 - 78). Gait bouts and gait quality metrics (walking speed, step length, timing asymmetry, and double support percentage) were collected from a standardized smartphone operating system. Pre- and post-operative values were compared using paired-samples t-tests (p<0.05).
A total of 595 females and 415 males with a mean age of 61.9±9.3 years and mean BMI of 30.2±6.1 kg/m2 were reviewed. Walking speeds were lowest at post-operative week two (all, p<.001). Speeds exceeded pre-operative means consistently by week 21 (p=0.015) for PKA, and week 13 (p=0.007) for THA. The average weekly step length was lowest in post-operative week two (all, p<0.001). PKA and THA cases achieved pre-operative step lengths by week seven (p=0.064) and week 9 (p=0.081), respectively. The average weekly gait asymmetry peaked at week two post-operatively (all, p <0.001). Return to pre-operative baseline asymmetry was achieved by week 11 (p=0.371) for TKA, week six (p=0.541) for PKA, and week eight (p=.886) for THA. Double limb support percentages peaked at week two (all, p<0.001) and returned to pre-operative levels by week 24 (p=0.089) for TKA, week 12 (p=0.156) for PKA, and week 10 (p=0.143) for THA.
Monitoring gait quality in real-world settings following joint reconstruction using smartphones is feasible, and may provide the advantage of removing the Hawthorne effect related to typical gait assessments and in-clinic observations.
