The Microsoft Kinect is a reliable method but lacks agreement with the Perometer when measuring lower limb volume: a test-retest reliability and betweenmethods agreement study.
(2017) The Microsoft Kinect is a reliable method but lacks agreement with the Perometer when measuring lower limb volume: a test-retest reliability and betweenmethods agreement study., no. 26.
Question: What is the test-retest reliability of the Microsoft Kinect sensor and the perometer? What is the level of agreement between the Kinect and the perometer? Design: Between-methods agreement and test-retest reliability study. Participants: Fifteen university students were recruited through convenience sampling. Intervention: All participants' dominant lower limb volumes were measured using the vertically oriented perometer and the Microsoft Kinect sensor using Skanect and Meshlab software. Outcome measures: A one-sample t-test investigated the difference between the methods. A Bland-Altman plot determined the bias between the methods along with the limits of agreement (LOA). Intraclass correlation coefficient (ICC) was used to show the agreement between the methods (3,k), and determine the test-retest reliability of each (3,1). Results: The one-sample t-test showed that there was a statistically significant difference in the measurements captured by the two methods (p=0.002). The Bland-Altman plot showed that the Kinect overestimated lower limb volume by 626 ml compared to the perometer and wide LOA were observed (-615 to 1867ml). The ICC value for agreement was high (0.095), however, the 95% CIs were wide (0.6606 to 0.987). High ICC values were reported for test-retest reliability for both the Kinect (0.956) and perometer (0.994). Conclusion: There was a lack of agreement between the perometer and Kinect thus, the two methods should not be used interchangeably to measure lower limb volume. Further research should be carried out on the Kinect, and investigation is required into testing the Kinect on clinical populations.