Abstract Body: Background: Self-reported measurements have been used as a monitoring approach because it is difficult to objectively assess physical activity levels at a population level. Our aims were to assess the validity of self-reported physical activity data compared with accelerometer-based measurements in adults and to investigate the association between self-reported physical activity (Study 1) and the prevalence risk of hypertension, diabetes, and obesity, had physical activity been measured by accelerometer instead (Study 2).
Methods: We used data from the 2014-2017 Korea National Health and Nutrition Examination Survey, which involved a nationally representative sample of n=20,059 Korean adults aged 20-79 years. Participants in our Global Physical Activity Questionnaire (GPAQ) survey wore an accelerometer (Actigraph GT3X+) for a period of seven days. We used a regression model to calculate validity coefficients and attenuation factors using accelerometer-measured physical activity as the criterion measure. Associations between moderate-to-vigorous physical activity status and the prevalence of hypertension, diabetes, and obesity were determined using multiple logistic regression calibration method.
Results: The results from Study 1 showed that participants tended to overreport their physical activity levels when measured through GPAQ, with this tendency being more pronounced among those with lower levels of physical activity. Additionally, women and individuals with higher education or income were more accurate in reporting their physical activity levels than others. Study 2 revealed that the associations between the prevalence of hypertension, diabetes, and obesity and adjusted physical activity were considerably deattenuated, after correcting for measurement error in self-reported physical activity. The prevalence ratio (PR) for hypertension and diabetes for higher levels of self-reported physical activity, using the inverse hyperbolic sine function (IHS) transformation, was 0.79 (95% CI: 0.78, 0.80) and 0.75 (95% CI: 0.74, 0.77), respectively.
Conclusions: The associations between physical activity and the prevalence risk of hypertension, and diabetes are significantly attenuated by the overestimation of self-reported physical activity, suggesting that the beneficial effects of physical activity may have been underestimated in the past. Further research will analyze the association between physical activity and mortality.