A Longitudinal Study of Real-World Physical Activity Assessed Using Cut—point Free Metrics, Mobility Capacity, and Mobility Perception Among Older Adults Recovering from Proximal Femoral Fracture

<jats:title>Abstract</jats:title> <jats:p>Physical activity (PA) is measured objectively through daily wearable monitoring and mobility capacity tests, and subjectively via patient reported outcomes (mobility perception). This study investigated longitudinal changes in, and relationships between, different measures of PA among older adults recovering from proximal femoral fracture (PFF). Participants (N=201) were classified into four groups by time since surgery at baseline (T1) and followed over two assessments (T2, T3). They wore an accelerometer for 7 consecutive days. Daily PA was measured using cut-point free metrics including Average Acceleration, Intensity Gradient, and intensity of the most active accumulated X minutes (MX: M1-M90). Mobility capacity and perception of participants were evaluated using clinical tests (e.g., 6-minute Walking Test (6MinWT)) and questionnaires (Late-Life Function and Disability Instrument (LLFDI)). MX metrics, particularly M1-M15, increased significantly across the first three groups with higher sensitivity in group 1 (p&lt;0.001).</jats:p> <jats:p>Distance covered during the 6MinWT increased significantly (p&lt;0.01). Three of the seven LLFDI’s domains showed the largest significant changes. Overall, sustained, moderate-strong positive correlations were observed between the clinical tests, LLFDI, and short-duration MX metrics in group 3 and 4 at T1, and across all participants at T2 and T3. Thus, MX metrics (M1-M15) can reveal change for daily PA intensities, especially among PFF groups in early recovery groups at T1 and reached the late stage at follow-ups. Clinicians may focus on specific LLFDI’s domains to maximize assessment efficiency. The direct links between mobility capacity, perceived mobility, and short-duration MX metrics indicate the potential of these metrics to monitor patients remotely.</jats:p>