Validity of the GPAQ for Measuring Sedentary Behaviour in a Chilean Adult Population

Objective: To assess the validity of the Global Physical Activity Questionnaire (GPAQ) for measuring sedentary behaviour (SB) in the Chilean adult population.

Methods: 204 adults (88/116 male/female, 45.2±14.0) who were randomly selected during National Health Survey 2009-2010 completed the protocol. The participants wore an Actigraph GT3X (AG) for at least 7 consecutive days and then completed the Spanish version of the GPAQ which includes one single-item question for measuring time spent sitting in a usual day. For AG data, at least 10 hours of wearing time and 4 days were required to be valid. Validity was examined using Spearman’s correlation, mean bias and limits of agreement (LoA), with AG (vertical axis <100 counts/minute) as the reference standard. Agreement between the GPAQ and AG for classifying data into quartiles and tertiles was assessed with kappa method.

Results: The GPAQ showed poor correlation with AG (r=0.25, p<0.001), with mean bias of -0.55 h/day and 95% LoA ranging from -1.58 to 0.48 h/day. Agreement between the GPAQ and AG was fair but significant for categorising time spent in SB into tertiles (43.78%, k=0.16, p<0.001) and quartiles (31.84%, k=0.09, p=0.01). For tertiles, nearly 1 out of 2 participants were correctly classified into the lowest and highest quantiles, while for quartiles 1 out of 3 and 2 out of 5 were correctly classified into the lowest and highest quantiles, respectively.

Conclusion: Despite a relatively low mean bias the GPAQ has poor validity when compared with AG. Adjustments by participants’ characteristics may be necessary to improve its validity.