Validation of chest/neck motion sensor site and the traditional Chinese version of the global physical activity questionnaire in the community dwelling Hong Kong Chinese elderly

The World Health Organization recommends that older adults aged 65 and above should perform at least 150 minutes of moderate intensity aerobic physical activity (PA) a week and at least two non-consecutive days of resistance and stretching exercise per week to maintain health and prevent the development of chronic conditions. Motion sensors are an effective means of enhancing daily PA participation, but only if the sensors are valid and users comply with their use. When accelerometry collection is not possible or where additional information such as in which domain the activity occurs, valid, simple, and easily administered questionnaires can help to develop a thorough understanding of older adults’ actual PA pattern. This information is crucial for designing effective health promotion programmes for the Hong Kong elderly population. The primary aim of this study was to investigate the validity and compliance of a novel motion sensor site (chest/necklace), compared with traditional waist and wrist sites. The secondary outcome was to determine the validity and reliability of the new Global Physical Activity Questionnaire (GPAQ) for measuring the PA of Hong Kong elders, compared with the International Physical Activity Questionnaire (IPAQ) and objective measures. A convenience sample of community-dwelling Hong Kong Chinese elderly individuals was studied. PA was measured objectively over four days, including at least one weekend, with three types of portable motion sensors (a three-dimensional Omron pedometer worn on the chest/neck; three tri-axial ActiGraph accelerometers, with one each positioned at the waist, non-dominant side of the wrist and the chest/neck; and the SenseWear multi-sensor) and subjectively with a questionnaire interview using either the GPAQ or IPAQ before and after wearing the sensors. Thirty-nine men and fifth-one women (age range 66-85 years, mean ±SD: 72.9 ± 5.1) were measured. The Spearman’s rho correlation coefficients were moderate to strong for the total step counts measured by the accelerator at the chest/neck site, the waist, the wrist and pedometer (r=0.787, r=0.901, r=0.730 and r=0.903 respectively) and statistically significant (p<0.05) when compared with the SenseWear (SWA) criterion measures. There was no significant difference of steps measured at the chest/neck and the pedometer with 1052 ± 892 and -1389 ± 625 (mean difference ± standard error) respectively. However, the total vector magnitudes (VM) measured at the chest/neck were strongly correlated with the VM measured at the waist (r=0.86, p<0.005) and moderately correlated with the VM measured at the wrist (r=0.698, p<0.05). A repeated measures of ANOVA test showed that there was significant difference (F(1.045, 92.98)= 701.9, p<0.05) between the three sites. The correlation for GPAQ was low and statistically significant (r=0.383, p=0.003) compared with SWA, but moderate (r=0.708, p