Are Accelerometer-based Functional Outcome Assessments Feasible and Valid After Treatment for Lower Extremity Sarcomas?

Furtado, Sherron, Godfrey, Alan, Del Din, Sylvia, Rochester, Lynn and Gerrand, Craig (2020) Are Accelerometer-based Functional Outcome Assessments Feasible and Valid After Treatment for Lower Extremity Sarcomas? Clinical Orthopaedics and Related Research, 478 (3). pp. 482-503. ISSN 0009-921X

Preview

Text Furtado et al - Accelerometer-based Functional Outcome Assessments AAM.pdf - Accepted Version Download (283kB) | Preview

Abstract

Background: Aspects of physical functioning, including balance and gait, are affected after surgery for lower limb musculoskeletal tumors. These are not routinely measured but likely are related to how well patients function after resection or amputation for a bone or soft tissue sarcoma. Small, inexpensive portable accelerometers are available that might be clinically useful to assess balance and gait in these patients, but they have not been well studied.

Questions/purposes: In patients treated for lower extremity musculoskeletal tumors, we asked: (1) Are accelerometer-based body-worn monitor assessments of balance, gait, and timed up and go tests (TUG) feasible and acceptable? (2) Do these accelerometer-based body-worn monitor assessments produce clinically useful data (face validity), distinguish between patients and controls (discriminant validity), reflect findings obtained using existing clinical measures (convergent validity) and standard manual techniques in clinic (concurrent validity)?

Methods: This was a prospective cross-sectional study. Out of 97 patients approached, 34 adult patients treated for tumors in the femur/thigh (19), pelvis/hip (3), tibia/leg (9), or ankle/foot (3) were included in this study. Twenty-seven had limb-sparing surgery and seven underwent amputation. Patients performed standard activities while wearing a body-worn monitor on the lower back, including standing, walking, and TUG tests. Summary measures of balance (area [ellipsis], magnitude [Root Mean Square (RMS)], jerkiness [jerk], frequency of postural sway below which 95% of power of acceleration power spectrum is observed [f95 of postural sway], gait [temporal outcomes, step length and velocity], and TUG time were derived. Body-worn monitor assessments were evaluated for feasibility by investigating data loss and patient-reported acceptability and comfort. In addition, outcomes in patients were compared with datasets of healthy controls collected in parallel studies using identical methods as in this study to assess discriminant validity. Body-worn monitor assessments were also investigated for their relationships with routine clinical scales [(Musculoskeletal Tumour Society Scoring system (MSTS), Toronto Extremity Salvage Score (TESS), Quality of life-Cancer survivors [QoL-CS)] to assess convergent validity and their agreement with standard manual techniques (video and stopwatch) to assess concurrent validity.

Results: Although this was a small patient group, there were initial indications that body-worn monitor assessments were well-tolerated, feasible to perform, acceptable to patients who responded (19 of 20 [95%] found the body-worn monitor acceptable and comfortable and 17 of 20 [85%] found it user-friendly), and produced clinically useful data comparable to the evidence. Balance and gait measures distinguished patients and controls (discriminant validity), for instance balance outcome (ellipsis) in patients (0.0475; 95% confidence interval [CI] 0.0251–0.0810 m2/s4) was affected compared with controls (0.0007; 95% CI, 0.0003–0.0502 m2/s4; p = 0.001). Similarly gait outcome (step time) was affected in patients (0.483; 95% CI, 0.451–0.512 seconds (s)) compared with controls (0.541; 95% CI, 0.496–0.573 s; p < 0.001). Moreover, body-worn monitor assessments showed significant relationships with existing clinical scales (convergent validity), for instance ellipsis with MSTS (r = -0.393; p = 0.024). Similarly, manual techniques showed excellent agreement with body-worn monitor assessments (concurrent validity), for instance stopwatch time 22.28 +/- 6.93 s with iTUG time 21.18 +/- 6.23 s (ICC agreement = 0.933; p < 0.001). P < 0.05 was considered statistically significant.

Conclusions: Although we had a small, heterogeneous study patient population, this pilot study suggests that body-worn monitors might be useful clinically to quantify physical functioning in patients treated for lower extremity tumors. Balance and gait relate to disability and quality of life. These measurements could provide clinicians with useful novel information on balance and gait, which in turn can guide rehabilitation strategies.

Level of Evidence: Level III, diagnostic study.

Actions (login required)

View Item

Downloads