An exploration of the activPAL™ activity monitor in measurement of sedentary and physical activity patterns in people with mild to moderate Parkinson’s disease: A validation study.
Background Measurement of physical activity (PA) and sedentary behaviour (SB) may facilitate individually-tailored interventions aimed at modifying the progression of Parkinson’s disease (PD) symptoms. Accelerometry is suggested as a practical and objective measurement tool. Aim This study aims to investigate the validity of the activPAL in measuring PA and SB in people with mild to moderate PD with direct observation (DO) as the criterion measure. Design Cross-sectional, DO, validation study. Methods Eight men and two women (mean age 69 ± 7.3) with mild to moderate PD (median Hoehn and Yahr 1.5, IQR 1-2) were recruited. Wearing an activPAL, they completed a controlled module consisting of 3-minute periods in different postures and a “free-living” module consisting of four randomly selected ADLs. The trial was captured on a video-camera and the time spent in different postures was classified. This was aligned and compared with the activPAL output data. Results The activPAL has a low percentage error for SB in both modules (9.11% and 3.37%) and a higher percentage error for PA (-15.1% and -19.63%). There was no statistically significant correlation between the activPAL and DO in module 1 (r=0.68, p=0.03 and r=0.55, p=0.098), however, high correlation in module 2 (r=0.94, p<0.001 and r=0.99, p<0.001). Wide limits of agreement (LOA) (range -112.6s – 106s) were demonstrated between the activPAL and DO throughout the whole test. Conclusion The activPAL overestimated SB and underestimated PA. Wide LOA may be considered clinically unacceptable. Further research is warranted to validate the activPAL before its use in practice.