04th Apr 2024

Introduction: Questionnaires and device-based measures are used for understanding daily behaviors (sleep, sedentary behavior (SB), physical activity (PA)). The recently developed Daily Activity Behavior Questionnaire (DABQ) showed a good agreement for sleep but poor agreements for SB, light-intensity PA (LPA), and moderate-to-vigorous PA (MVPA) compared to ActivPAL. While the thigh-worn ActivPAL is best known for classifying posture-related behaviors, hip-worn devices like Actigraph-GT3x+ are preferred to classify activity intensities. Therefore, the aim of this study is to compare the DABQ with Actigraph-GT3X+, and to explore associations of sociodemographic factors with measurement differences.

Methods: Adults (n=105, 54% female, ±46 y/o) wore an Actigraph-GT3X+ for seven consecutive days (hip/day; wrist/night). Afterwards, participants filled in the DABQ considering the past seven days. Raw accelerometer data were processed by the R-package GGIR comparing different metrics (ENMO, MAD). Single movement analysis (intraclass correlation coefficient (ICC), interclass correlations (r), Bland Altman plots) and compositional analyses (multilevel regression models) were used to compare the measurement techniques.

Results: There was moderate agreement between DABQ and Actigraph for sleep (ICC 0.55; r=0.55) across different metrics. Poor agreements were found for SB, LPA and MVPA, however with varying results across metrics (SBenmo: ICC=0.01, r=0.06; SBmad: ICC=0.24, r=0.42; LPAenmo: ICC=0.01, r=0.11; LPAmad: ICC=0.17, r=0.5; MVPAenmo: ICC=0.22, r=0.28; MVPAmad: ICC=0.24, r=0.38). Using ENMO, Actigraph estimated +400 minutes SB/day, -350 minutes LPA/day and +15 minutes MVPA/day vs DABQ. Using MAD, Actigraph estimated +100 minutes SB/day, -150 minutes LPA/day and +40 minutes MVPA/day vs DABQ. Compositional models found that age, sedentary job and education were associated with differences between DABQ/Actigraph (p<0.001).

Conclusions:. Compared to Actigraph, the DABQ underestimates SB and MVPA, but overestimates LPA. Differences between Actigraph and DABQ measurement differ by metric, age, job and education. Understanding the validity of the DABQ questionnaire is helpful to improve assessments of 24-hour movement behaviors.