The activPAL™ and Actigraph activity monitors show a low level of agreement in measuring sedentary behaviour in people with Parkinson's disease: a 7-day validation study

Abstract

Background People with Parkinson's disease (pwPD) have high levels of sedentary behaviours (SB) which are associated with adverse health-outcomes. Validated instruments are required to objectively quantify SB, establish dose-response relationships and SB intervention strategies. Aims To investigate the concurrent validity of the activPAL™ and ActiGraph activity-monitors in pwPD during 7-days of free-living. Methods 9 pwPD completed a 7-day period of free-living while concurrently wearing the activPAL™ and ActiGraph activity-monitors and completed a wear-time diary. activPAL™ SB was determined from total sit/ly time per day. ActiGraph SB was calculated using an activity-count cut-points of vertical-axis<100cpm (AG100) and vector-magnitude<200cpm (AG200). Descriptive statistics, correlations and agreement were compared for average SB/day and an individual-day basis. Results Participants recorded 13.2 hours/day of wear-time. Average SB/day for the activPAL™ was 519 minutes (SD 72). The AG100 reported 133 minutes (SD 40) greater and AG200 reported 43 minutes (SD 40) greater. Very strong, positive correlations were found between all activPAL™ and ActiGraph comparison (r=0.84-0.93, p<0.01). Bland-Altman plots showed large mean differences (AG100 bias 133 minutes and AG200 bias 43). 95% Limits of agreement were >2.5 hours/day for both AG100 and AG200. Conclusion The concurrent validity between the ActivPAL™ and ActiGraph is low, characterised by noticeable differences in reported SB/day and large levels of variation of greater than 2.5 hours per day. Results suggest that the activity-monitors are not interchangeable and SB values reported are not directly comparable. Different cut-points show a large influence on the ActiGraph reported SB therefore further investigation is needed for pwPD.