Examine the construct validity of the Nextiles arm sleeve by assessing its agreement and convergence with markerless 3D biomechanics used as reference for 1) peak elbow varus torque and 2) pitches thrown with separate analyses for fastballs, breaking balls, and changeups.
Collegiate pitchers were simultaneously assessed through the Nextiles arm sleeve and markerless three-dimensional (3D) motion capture as pitchers threw fastballs, breaking balls, and changeups to a catcher at regulation distance within a laboratory setting. Prior to study inception, a literature synthesis and knowledge users were individually queried for opinions on acceptable limits of agreement. An a priori threshold of 20 Nm was determined for acceptable limits of agreement. A continuity correction was incorporated to account for repeated measures at the individual pitcher level. A total of 18 pitchers were required to pitch three pitches per pitch type, to have an acceptable mean difference confidence interval of 20 Nm. An a priori 95% confidence interval width of 0.15 was determined to be acceptable, with an expected linear association of 0.65, to meet the acceptable lower limit of the 95% confidence interval width of 0.50. Using the design effect method to account for repeated measures at the individual pitcher level, throwing 90 total pitches was required for each pitch type. Nested Bland-Altman limits of agreement (LOA), mixed effect linear regressions, and intraclass correlation coefficients (ICC) were calculated between Nextiles sleeve and 3D biomechanics for elbow varus torque.
29 pitchers were included, throwing 200 pitches (Fastball: 72%; Breaking Ball: 15%; Changeup: 14%), with 182 pitches recorded by the Nextiles sleeve. Agreement and validity for Fastball measures (LOA: 6.9 Nm (-28.5, 42.4); Beta: 0.27 (0.05, 0.51); ICC: Nextiles: 0.96 (0.93, 0.98); KinaTrax: 0.97 (0.94, 0.98)) were higher compared to agreement for breaking balls (LOA: 21.3 Nm (-16.7, 59.31); Beta: 0.14 (-0.04, 0.25); ICC: Nextiles: 0.99 (0.98, 1.00); KinaTrax: 0.86 (0.60, 0.97)) and changeups (LOA: 15.3 Nm (-17.2, 47.7); Beta: -0.01 (-0.47, 0.58); ICC: Nextiles: 0.99 (0.98, 1.00); KinaTrax: 0.86 (0.60, 0.97)).
Although fastballs had improved limits of agreement and convergence between the Nextiles and KinaTrax for all measures, the measures were beyond acceptable thresholds. Repeated measures using Nextiles for breaking ball, and change-up, were more consistent than for KinaTrax. These results suggests that the Nextiles sleeve is not measuring the same construct as KinaTrax.
Pitchers may utilize the Nextiles sleeve within on field sessions to monitor their own individual pitch count. Caution is warranted when using KinaTrax for pitches that do not display similar hand and forearm mechanics to fastball pitches. Our findings do not support the use of the Nextiles arm sleeve or KinaTrax to replace laboratory 3D biomechanical analyses if precise measures are required.
Injury
Wearable Technology
