Abstract MP15: Different Questions Lead to Different Answers: Comparing Different Survey Methods for Assessing Relationships between Sleep Duration and Cardiometabolic Health Risk Factors

Introduction: Sleep Duration is part of Life’s Essential 8, and sleep duration items are included in surveys. Yet, sleep duration can be assessed in several ways. The present study examined the degree to which assessing sleep duration using different survey items produces different results when examining associations with cardiometabolic health. Methods: Data from the Sleep and Healthy Activity, Diet, Environment, and Socialization (SHADES) study was used. SHADES is a community-based study of adults in southeastern Pennsylvania (N=1007). Habitual sleep duration was assessed in four ways: the National Health and Nutrition Examination Survey (NHANES) item assessed hours on a typical weeknight, the Behavioral Risk Factor Surveillance System (BRFSS) item assessed typical hours in 24h, the Pittsburgh Sleep Quality Index (PSQI) item assessed hours on a typical night, and the Sleep Timing Questionnaire (STQ) computed typical weeknight and weekend sleep from self-reported bed and wake times. These were evaluated continuously and categorized as short (≤6h), normal (7-8h, reference), and long (≥9h). Correlations among variables were computed. Binary logistic regression analyses adjusted for age and sex examined each of these sleep continuously and categorically relative to obesity, hypertension, hypercholesterolemia, and arrhythmia. Results: Although questions were asked of the same participants at the same time, there was notable variability in responses. No pairs were collinear in either Pearson or Spearman correlations; most correlations were weak to moderate. The Table shows results of regression analyses. As a continuous variable, increased sleep duration was associated with decreased likelihood of obesity and hypertension, but this was only observed using some variables. Also, short and long sleep were associated with obesity and hypertension, and long sleep was associated with increased likelihood of hypertension, hypercholesterolemia, and arrhythmia, but these results were only observed using certain variables. Conclusions: Despite similarly-worded self-report items, there was a high degree of inconsistency regarding whether sleep duration variables were associated with cardiometabolic disease risk factors. This has implications for epidemiologic studies attempting to assess Life’s Essential 8. Assessments that differentiate time in bed vs sleep, and nighttime vs 24-hour sleep, may produce differing results.