Workplace Sedentary Behavior Questionnaire

Abstract

Self-report remains an important measurement method for physical activity and sedentary behaviors because it provides the contextual information needed in targeted behavioral interventions. PURPOSE: This study aimed to evaluate the validity and responsiveness to change of a workplace sedentary behavior questionnaire (SBQ) in assessing total sedentary behavior (1) at work and (2) outside of work during workdays, and during (3) non-workdays. METHODS: Participants (N= 359) were recruited from 16 worksites in the greater Minneapolis and Phoenix regions. Participants wore an activPAL3c accelerometer for 7 days at two timepoints (baseline and 12-week follow-up). They also completed a sedentary behavior questionnaire on the 7th day of each assessment timepoint. Time spent in each behavior from both measures was standardized to a 16-hour day on non-work days and 8-hours to both work hours and non-work hours on work days. The agreement between the two measures was assessed through single-measure with absolute definition ICC using a two-way random effects model. Bias estimates (mean difference [] and root of mean square error [RMSE= were used to compare reported behaviors to the activPAL values. Responsiveness to change following the intervention was assessed using the responsiveness statistic (RS). RESULTS: There was poor agreement between the SBQ and activPAL (ICC from 0.06 - 0.29) across all time periods. On average, participants over-reported their time spent sitting (Mean [95% CI]= 39.9 [31.1, 48.7] min/day, RMSE= 112.5) and moving (12.9 [9.39, 16.5] min/day, RMSE= 44.3), but under-reported standing (-52.9 [-60.5, -45.2) min/day, RMSE= 105.5) at work. Moreover, sitting time during their non-work hours was over-reported (51.7 [34.5, 68.8] min/day, RMSE= 211.8). Sitting on non-work days was underestimated (-87.2 [-111.6, -62.7] min/day, RMSE= 298.8). SBQ was similarly responsive to change compared to the activPAL except for sedentary time during non-work days (0.13 and 0.49 for SBQ and AP, respectively). CONCLUSION: Despite the poor absolute agreement between the SBQ and activPAL, overall bias estimates and responsiveness during work days were acceptable. Further studies are needed to explore ways to improve accuracy in sedentary behavior reporting during non-working days.