Reliability of an International Physical Activity Questionnaire (IPAQ) Modified to Capture Neighbourhood-specific Physical Activity ( poster )

Abstract

Background Few self-report tools exist that accurately capture neighbourhood-based physical activity (PA) including walking. Our study aimed to modify PA items from the widely-used International Physical Activity Questionnaire (IPAQ) to capture neighbourhood-based PA, and to estimate the reliability of these modified items. Methods In 2016,75 adults from four Calgary neighbourhoods (2 high-walkable n=34 and 2 low-walkable n=41) completed the same questionnaire twice, 7-days apart. Modified-IPAQ items captured days/week and usual minutes/day of transportation cycling (TC), transportation walking (TW), leisure walking (LW), moderate PA (MPA), and vigorous PA (VPA) undertaken during the last 7-days inside the neighbourhood. To estimate the reliability for these items between the two questionnaires, we undertook Spearman’s rank correlations (r). Moreover, percent of overall agreement and Kappa statistics (κ) estimated the consistency in reported participation (any or none). Dependent t-tests, also compared differences in neighbourhood PA outcomes of the high and low walkable neighbourhoods. Spearman’s rank correlations and Kappa values approaching 1 suggest excellent agreement or consistency while values closer to 0 suggest poor agreement or consistency. Results The mean age (±SD) of our sample was 54±14 years, consisted mostly of women (65.3%), those with university-level education (80.0%), non-dog owners (65.3%), and those with access to a motor vehicle (100.0%) or bicycle (81.3%). Participation in neighbourhood PA ranged from poor (κ=0.21 for MPA) to moderate (κ=0.55 for VPA), while proportion of overall agreement was excellent and ranged from 77.3% for MPA to 81.3% for TC. Reported neighbourhood PA between the two questionnaires ranged from poor (r=0.20 for MPA) to moderate (r=0.62 for TW) for days/week, poor (r=0.34 for MPA) to moderate (r=0.55 for VPA) for minutes/day, and poor (r=0.19 for MPA) to moderate (r=0.54 for TW and VPA) for minutes/week. TC and TW participation, TC days/week, duration/day, and duration/week, and TW days/week and duration/week was significantly (p Conclusions Our findings suggest that, overall, IPAQ items can be modified to provide reliable self-report estimates of neighbourhood-based physical activity in adults. The modified-IPAQ items could be used to capture neighbourhood-based physical activity in studies investigating the relations between neighbourhood built environments and walking.