Validity of a children's physical activity questionnaire (cPAQ) for the study of bone health

BACKGROUND: The aim of this cross-sectional study was to examine the ability of a children's physical activity questionnaire (cPAQ) to assess physical activity levels and bone health status of school children.
METHODS: Subjects consisted of 90 pre-pubertal and early pubertal children aged 9-10 years. Components of physical activity were assessed using metabolic intensity (METPA) scores and mechanical bone strain (MECHPA) scores. An Actical accelerometer was used to validate METPA scores among a sub-sample of 57 children. Reliability was assessed by test-retesting all children after a 7 day interval. Whole body bone mineral content (BMC) was measured using dual-energy X-ray absorptiometry.
RESULTS: The reliability of cPAQ for assessment of various categories of physical activity was moderate to high (r ranged from 0.55 to 0.68, P < 0.001). Agreement was fair for repeated use of the cPAQ (Cohen's kappa = 0.32, P < 0.001). Bland-Altman plots show cPAQ had fair agreement only for moderate activity (mean difference 35.4 min/week; 95% limits of agreement -434.0 to +504.9 min/week). Approximately 69.6% of children were correctly classified (into the same or adjacent quartiles) according to the quartiles of BMC for METPA score, and 58.7% were correctly classified according to MECHPA score. Only 10.9% and 12.0% of children were grossly misclassified as compared to METPA and MECHPA scores, respectively.
CONCLUSIONS: The cPAQ has reasonable validity in assessing moderate physical activity, and it demonstrates good ability to accurately classify children according to BMC. It fails, however, to assess other activity levels, suggesting that objective measurement is still a better method of assessment of physical activity among primary school children.