Abstract MP91: Self-Reported Sitting Time Using a Two-Question Method Is a Good Measure of Sedentary Behavior: Findings in the New York City Adult Population

Introduction: Sedentary behavior is defined as activities that require low energy expenditure (1.0 - 1.5 [METs]). Sitting time is a specific domain of sedentary behavior that is associated with metabolic disturbances, decreased life expectancy, and mortality - outcomes that occur independent of meeting physical activity guidelines. Population-level prevalence estimates of the ‘active couch potato’, or active adults who engage in at least 7+ hours of sitting time a day, are needed.

Hypothesis: We tested that the two-question method of self-reported sitting time would have adequate construct validity for population surveillance.

Methods: The New York City Health Department conducted the 2011 Physical Activity and Transit (PAT) Survey using random-digit dial (cell and landline telephones) methods (n=3811); an additional subset of participants wore accelerometers for a seven-day period (n=679). Sedentary time was defined as accelerometer minutes with <100 counts on valid days (those with 10 hours of wear time). Self-reported sitting time was assessed from two questions on time spent sitting awake for daytime and evening. Construct validity was assessed using Spearman’s Rho and examination of a Bland-Altman plot. Multivariable regression models were used to assess demographic factors associated with sitting time and measured sedentary time. Prevalence of the ‘active couch potato’ was defined as those adults who were meeting the 2008 Physical Activity Guidelines for Americans of at least 150 min/week of physical activity, but who also self-reported sitting for 7+ hours per day, using both accelerometer and self-reported data. All data were weighted and analyzed in SUDAAN 10.

Results: Self-reported mean daily sitting time was 444 min (7.4 hours); accelerometer-based mean daily sedentary time was 494 min (8.2 hours; correlation: r=0.32, p<0.01). The mean difference was 49 min/d (Limits of Agreement: -441-343). In adjusted models, higher measured sedentary time was associated with older age and higher education. Sitting time was higher in women, Asians and those with a higher education, and lower in foreign-born persons and Hispanics, compared to referent groups. The prevalence of being an ‘active couch potato’ was 56.9% and was highest among high-income, highly educated New Yorkers.

Conclusions: Sitting time can be accurately assessed by self-report for population surveillance, but may be limited in characterizing individual-level behavior. Differences in measurement characteristics suggest that accelerometer-based measurements may not be an appropriate way to validate self-reported sitting time. Sedentary behavior appears to be prevalent in two different groups: (1) older adults and; (2) working-age adults of higher socioeconomic status. Targeted interventions to minimize extended periods of sedentary activity among these groups are warranted.

1