Construct validity of scores from the Godin leisure-time exercise questionnaire in adults with Crohn's disease and healthy controls.

Abstract

To examine the construct validity of scores from the Godin Leisure-Time Exercise Questionnaire Health Contribution Score (GLTEQ HCS) as a measure of moderate-to-vigorous physical activity (MVPA) in adults with Crohn's disease (CD) and healthy controls. Thirty-eight adults with CD (78.9% female; 41.3 ± 12.0 years) and 41 controls (73.2% female; 39.3 ± 12.0 years) completed the GLTEQ and wore an ActiGraph GT3X-BT accelerometer for 7 days. Mean differences in accelerometer outcomes across HCS categories were examined (known-groups test of construct validity). Correlations between GLTEQ HCS and accelerometer-derived MVPA (convergent validity), and light PA (LPA) and sedentary behavior (divergent validity), were examined separately by group. Significant differences in accelerometer-derived MVPA across the three HCS categories were observed in controls [<i>F</i>(2,40)=3.90, <i>p</i>=.03, <i>η²</i>=.17], but not CD. When dichotomized (active ≥24 units versus insufficiently active <24 units), medium-to-large differences in MVPA emerged for both CD (<i>d</i>=.62) and controls (<i>d</i>=.87), though this difference did not reach statistical significance in CD. GLTEQ HCS correlated with MVPA in controls (<i>r_s</i>=.40, <i>p</i>=.01) but not in CD (<i>r_s</i>=.28, <i>p</i>=.09), and was not associated with LPA or sedentary behavior. GLTEQ HCS demonstrated modest, context-dependent construct validity for MVPA assessment, with stronger evidence using dichotomized classification and in controls than in adults with CD.