Investigation of a Low-cost Portable Inline Dynamometer for Measuring Isometric Knee Extensor Strength
Amantej Sran, Elira Leake, Sydney Yott, Julie E Brandon, Geoffrey A Power, Scott C.E. BrandonBackground and Purpose
Isometric strength testing is an important tool for detecting deficits in injured populations and monitoring strength changes during rehabilitation and training. However, it is unclear whether strength measurements obtained from low-cost, inline force sensors match measurements from large, expensive, lab-based dynamometers. Therefore, the purpose of this study was to investigate agreement in maximum voluntary isometric knee extension torque measurements between a lab-based dynamometer, a low-cost wireless inline force sensor, and a wired inline force sensor.
Design
Prospective cross-sectional study
Setting
Laboratory
Methods
Fourteen healthy adult participants (age >18 years) were recruited from the university community using posters, email, and social media advertisements. Peak isometric knee extension torque was measured for two repetitions at each of three knee angles (90°, 45° and 15°) using a lab-based dynamometer (Humac, CSMI Medical Solutions, USA). Then, all six trials were repeated while simultaneously recording strength using a low-cost wireless inline force sensor (Tindeq, Trondheim, Norway) and a wired inline force sensor wired (MLP, Transducer Techniques, CA, USA). Peak electromyography (EMG) magnitude was recorded for knee extensors and flexors, and electrical stimulation was used to quantify the level of voluntary activation (VA) for knee extensors. Agreement between torque measurement systems was assessed using Pearson’s Correlation (r), Interclass Correlation Coefficients (ICC3,1, absolute agreement), Minimum Detectable Change, and Bland-Altman Limits of Agreement (LOA). Additionally, peak torque was compared at each knee angle using a repeated measures ANOVA, and peak EMG outputs were compared using paired t-tests.
Results
Fourteen participants completed the study (3 male, 11 female; 26 ± 8.97 years; 1.71 ± 0.11 m; 66.75 ± 15.9 kg). The mean (95% CI) difference in torque between Tindeq and Humac dynamometers was 0.16 Nm/kg (−0.07 to 0.26 Nm/kg) at a knee angle of 15 degrees, −0.02 Nm/kg (−0.19 to 0.14 Nm/kg) at 45 degrees, and −0.10 Nm/kg (−0.32 to 0.13 Nm/kg) at 90 degrees. Bland-Altman LOA were −0.86 to 0.67 Nm/kg at 15 degrees, −0.60 to 0.55 Nm/kg at 45 degrees, and −0.86 to 0.67 Nm/kg at 90 degrees. Bias and LOA were similar for the MLP device. All participants achieved >90% VA. Results were more similar between Tindeq and MLP inline force sensors (r = 1.0, ICC3,1 = 0.94-1.00) than between inline force sensors and the Humac device (r <0.81, ICC3,1 = 0.37-0.80).
Conclusions
Although sample size was small and only a single testing session was performed, the Tindeq device showed small measurement bias, underestimating the Humac torque by less than 0.1 Nm at 45 and 90 degrees. However, relatively large LOA indicated that individual measurement errors could exceed 25%. Low-cost wireless force sensors (e.g. Tindeq) show promise for clinical use, but individual measurements should be interpreted with caution.
Level of Evidence
2 (measurement agreement and validity study)