One of the main limitations of commercial games is the inability to determine improvements in the mental and physical health of the players. Although high game scores might provide an indication of higher cognitive and physical abilities, these are not sufficient to reliably determine the improvement in health outcomes. The work presented in this paper hence focuses on determining whether the collection of clinical measures during gameplay could potentially be used as a reliable indicator of improvement. For this study, the author uses the StepKinnection game for fall prevention, a Kinect-based game which builds on a hybrid version of the Choice Stepping Reaction Time (CSRT) task, a validated test that has been shown to prospectively predict older fallers. A group of 10 independent-living older adults was recruited and asked played with the game for a 12 weeks period of time. Assessments were conducted at baseline and every four weeks. Stepping performance data collected through gameplay was compared to the validated CSRT test. Statistical analysis proved that the stepping performance data collected by the game correlated and agreed with the validated measures of the CSRT test, suggesting that this could be used as a reliable indicator for health improvements.