Validation of the activPAL activity monitor in children with hemiplegic gait patterns resultant from cerebral palsy

Podium Presentation. Background: Current investigation of treatment outcomes by clinicians is currently hampered by a lack of clinically viable tools. The use of activity monitors specifically validated for a population could help resolve this situation (McAloon MT 2014). Aim: To investigate whether the activPAL monitor is a valid tool for monitoring activity in children with cerebral palsy. Method: The gold standard methodology was applied to measure criterion validity. Activity observed by independent raters of video recordings was compared against the activity measured by the activPAL activity monitor. Step count was monitored and the time periods spent sitting, standing and walking were recorded. Accuracy of the above was measured by, the mean percentage of agreement, mean difference of total times (Bland Altman 1986), sensitivity, and percentage error for number of steps taken in relation to velocity. Repeated measurements were taken to measure reliability. Results: The mean percentage agreement between all samples of observation and the activPAL was 86.46%. The time intervals for standing and walking were accurate with associated good sensitivity. However, the time intervals for sitting were found to be inaccurate affecting sensitivity due to misclassification of the event with two of the participants. Step count accuracy ranged from 78%-100%. The absolute agreement between the raters (0.9111) examined with intraclass correlation was very good (confidence interval 95% 0.9814 to 0.9958). The monitor was found to be reliable. Discussion & Conclusion : The monitor may prove useful to clinicians as a measurement outcome tool for children with hemiplegic gait patterns. Clinicians need to be aware that inaccuracies may occur when measuring sitting activities in smaller children due to them perching on chairs, as it occurred in this study and potential algorithm changes may be necessary. This study also found that the complexity of the individuals’ gait pattern affected step count accuracy to a greater degree than that caused by decreases in velocity. Where internal rotation and overlay of tonal affects to the contralateral limb occur, inaccuracies may also be evident. However, it was felt that the inaccuracies detected were not of a magnitude to prevent the activity monitor being used as a clinical outcome measure. References: Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet i. 1986;307-10