Psychometric Evaluation of Microsoft Kinect™ in Lumbar Spine Motion Assessment

Background and objective: Microsoft Kinect™ as a non-intrusive motion assessment device may solve the inherent limitations with goniometric and inclinometric approaches in range of motion (ROM) evaluations. This study assessed lumbar spine ROM using the inclinometer and Microsoft Kinect™ devices. It also examined the concurrent validity and intrarater reliability of the Microsoft Kinect™ in lumbar spine ROM assessment.
Methods: Four hundred consenting apparently healthy undergraduates from the Obafemi Awolowo University (OAU), Ile-Ife, Nigeria, participated in this cross-sectional study. Lumbar spine flexion and extension were assessed using the inclinometer and Microsoft Kinect™. Following the standardized procedure, the inclinometer was firmly strapped to each participant’s spine at the level of the anterior superior iliac spine, the values on the inclinometer readings were recorded at the start and after each movement in flexion and extension. Moreover, flexion and extension motion of participants were evaluated with the Microsoft Kinect™ device using the Microsoft Software Development Kit (SDK), version 1.8 and MB ruler, an on screen protractor software solution. Every motion was evaluated twice with both devices and the average value used for statistical analysis. The anthropometric characteristics were recorded and body mass index calculated. Data was summarized using descriptive and inferential statistics. Alpha level was set at p < 0.05.
Results: The mean age, weight, height and body mass index of the participants were 21.0 ± 2.48 years, 61.3 ± 61.3 Kg, 1.60 ± 1.63 m, and 23.4 ± 4.10 Kg/m2 , respectively. The mean values for lumbar flexion and extension were 116.32 ± 13.60 and 25.64 ± 7.90, and 66.10 ± 11.30 and 17.97 ± 5.70 for the Microsoft Kinect and the inclinometer, respectively. There was a significant poor correlation in the two lumbar spine range of motions evaluated for the Microsoft Kinect™ (flexion; r = 0.14, p = 0.001), (extension; r = 0.023, p =- 0.001). The intrarater reliability of both the Microsoft Kinect™ and the inclinometer in lumbar spine range of motion assessment was good. The ICC values for flexion and extension for the Microsoft Kinect™ was 0.77 and 0.87 while for the inclinometer was 0.72 and 0.65, respectively.
Conclusion: The Microsoft Kinect™ has good reliability but poor concurrent validity in comparison with the inclinometer in the assessment of lumbar spine movements in asymptomatic individuals.