Comparison of physical activity questionnaires in a large population based study (The HUNT Study, Norway) and the International Physical Activity Questionnaire (IPAQ), short form

Objective: To study the comparability of the International Physical Activity Questionnaire (IPAQ) for the assessment of physical activity (PA) with two differently structured PA questionnaires used in one the largest population based health studies ever performed, the Nord-Trøndelag Health Study (HUNT 1 and HUNT 2), Norway. Methods: The questionnaires were administered twice to a random sample of 108 men aged 20-39, and in the same individuals levels of VO2max (measured on treadmill) and PA and energy expenditure (EE) (measured by ActiReg), were obtained. ActiReg records the main body positions (stand, sit, bent forward and lie) together with the motion of the trunk and/or one leg each second. Results: The HUNT 1 questionnaire on PA showed intraclass correlation coefficients (ICC) ranging from 0.81 to 0.88, indicating very good reliability. In HUNT 2 ICC were 0.08 for light activity and 0.39 for hard activity, showing poor and fair reliability, respectively, while ICC for the IPAQ ranged from low 0.30 for moderate activity hours, to high 0.80 for sitting hours. In all three questionnaires we found moderate, significant correlation between the index based on questionnaire responses and VO2max (HUNT 1 r=0.48 (p≤0.01)) and for hard PA (HUNT 2, r=0.46 (p≤0.01)) and for total vigorous activity IPAQ (r=0.41 (p≤0.01)). In HUNT 1 metabolic equivalent (MET) values of 6 or more estimated from ActiReg and total vigorous activity from IPAQ most strongly correlated with the index r=0.39, r =0.55, respectively and in HUNT 2 with hard PA r=0.31, r=0.48 (p≤0.01) respectively. Conclusions: In young men the HUNT 1questionnaire on PA showed very good reliability and moderate validity, while HUNT 2 showed poor to fair reliability measuring light and hard PA, and respectively poor and moderate validity. Both questionnaires are very short and compared favorably with much longer instruments for assessment of more vigorous PA. The HUNT 1 questionnaire was best reproducible and yielded the most significant data on PA behaviour, making it useful also in future epidemiological studies.