Wearable sensors boost adherence and early recovery in home-based hip replacement rehabilitation

At the primary outcome measurement of 24 weeks, the sensor group showed significantly greater improvement in the Hip Disability and Osteoarthritis Outcome Score (HOOS), the main measure of hip-specific function (P=.01). Early differences were more pronounced: at 6 weeks, 4 of 5 HOOS subscales, timed up-and-go performance, and all patient-reported outcomes favored sensors (all P<.001 after correction for multiple comparisons). By 24 weeks, hip-specific functional differences had narrowed, but improvements in SF-36 physical and mental quality-of-life scores and Hospital Anxiety and Depression Scale scores remained significant in the sensor group (all P<.001). Adherence was substantially higher with sensors: the intervention group completed more exercise sessions (P=.002) and participated more frequently in weekly assessments and follow-up calls (both P<.001). Total 24-week healthcare and intervention costs were nearly identical between groups (¥87,967 vs ¥94,396; approximately US$12,880 vs $13,821; P=.32), with the sensor group averaging numerically lower costs. Adverse events were rare and minor in both groups (5.8% vs 7.5%).

The between-group difference in HOOS at 24 weeks, though statistically significant, was below the established minimal clinically important difference threshold, suggesting the practical magnitude of hip-function improvement was modest by clinical standards. However, the persistent improvements in SF-36 physical and mental scores and HADS anxiety/depression scores point to meaningful gains in overall quality of life and psychological well-being beyond hip-specific function alone. The proposed mechanism is that real-time sensor feedback enhances exercise execution quality and corrects movement patterns during unsupervised home sessions, which accelerates early postoperative neuromotor recovery—explaining why differences were largest in the first 6 weeks when motor re-education is most critical. The cost-neutrality is noteworthy: the sensor intervention generated a negative incremental cost-effectiveness ratio, meaning it cost less on average and produced better outcomes, though the cost difference did not reach statistical significance, likely due to sample variability. Subgroup outcomes for age, BMI, or baseline function are not reported, so individual responses may vary. The study was conducted in a single urban center in China with relatively high patient engagement (high adherence in both groups), which may limit generalizability to lower-resource or less digitally connected settings.