0858 A Predictive Model of Composite Sleep Health for Cardiovascular Disease Based on the RU-SATED Framework

Abstract Introduction Sleep health is a multidimensional construct. Previous studies have focused on single dimensions of sleep health and their association with cardiovascular disease (CVD). This study aimed to generate a composite multidimensional sleep health (MSH) score by combining subjective and objective measurements of sleep based on the RU-SATED framework, which consisted of Regularity, Satisfaction, Alertness, Timing, Efficiency, and Duration, and examine the association between the composite MSH score and the prevalence of CVD. Methods This was a secondary analysis using data from the multicenter, prospective, population-based Hispanic Community Health Study/Study of Latinos (HCHS/SOL). A total of 1,912 participants with complete baseline and actigraphy data were included in the analysis. The primary outcome was the prevalence of CVD, defined as self-reported coronary heart disease or stroke. Covariates included demographics (age, gender, education, and income) and major CVD risk factors (hypertension, dyslipidemia, diabetes, obesity, and smoking). Each domain in RU-SATED framework was measured using both self-reported questions and actigraphy, and converted to binary values with unfavorable conditions coded as 1. Receiver operating characteristic (ROC) curve was used to select the measurements with higher predictive accuracy for each domain. The composite MSH score was obtained by summing the scores from each single domain. Logistic regression analyses were performed to analyze the relationship between the composite MSH score and CVD prevalence. Results Subjectively measured sleep efficiency, timing, alertness, and satisfaction showed higher predictive values for CVD, and actigraphy-measured sleep duration and regularity were associated with better predictive performance. The odds ratio of the composite MSH associated with CVD was 1.261-fold (95% CI 1.060-1.500, p=0.009) after adjusting for covariates. Neither standalone subjective measures nor actigraphy measures of MSH were able to uncover this relationship (OR=1.161, 95% CI 0.978-1.378; OR=1.152, 95% CI 0.966-1.374). Conclusion In this study, actigraphy-measured or subjectively measured MSH did not predict the prevalence of CVD independently, whereas the composite MSH combined actigraphy-measured and subjective measurements could predict the prevalence of CVD. The results have inspired us to further explore standardized methods and reach a consensus in measuring MSH, using the RU-SATED framework as a guide. Support (if any)