Self-report Pregnancy Physical Activity Questionnaire overestimates physical activity.

OBJECTIVES Physical activity (PA) research during pregnancy relies heavily on indirect/subjective measures of PA, which may be less accurate than directly measured PA. We tested whether the Pregnancy Physical Activity Questionnaire (PPAQ) could accurately estimate PA by comparing PPAQ results to directly measured PA. METHODS In a sample of 29 women who completed the PPAQ, PA was directly measured in the second trimester of pregnancy using Actical® accelerometers (valid day = 10+ hours; 4-7 valid days). Activity variables from the PPAQ were calculated using all questions, and also by only considering the leisure time section. Women were classified as 'active' or 'non-active' using Canadian PA guidelines for adults (150 minutes moderate to vigorous PA (MVPA)/ week, bouts of 10+ minutes). Bonferroni corrections were used to adjust for multiple comparisons. Data presented as mean ± standard deviation or median (interquartile range). RESULTS The PPAQ overestimated MVPA by 12.12 (14.34) hours/week in the combined sample, and the difference remained substantial when investigating the non-active [overestimate = 11.54 (10.10) hrs/wk] and the active women [overestimate = 16 ± 11 hrs/wk] separately. PPAQ-measured PA variables did not correlate with any of their respective Actical®-measured variables (p > 0.008). The leisure time PPAQ questions overestimated MVPA by 1 ± 3 hrs/wk, with a positive correlation between PPAQ-leisure time MVPA and Actical®-measured MVPA (r = 0.565, p = 0.001). CONCLUSION The PPAQ significantly overestimates MVPA and does not provide an accurate estimate of PA in pregnancy. While PPAQ leisure time questions may help distinguish trends in PA, data from subjective questionnaires may result in misinterpretation of relationships between prenatal PA and health outcomes.