Title: Differential Impact of School Start Times: Exploring Heterogeneous Treatment Effects of Delaying School Start Times on Sleep and BMI via Glmm Trees in the Start Study
Presented by Emma Billmyer
Masters Candidate in Biostatistics
Plan B Adviser: Ashley Petersen
Abstract: Adequate sleep is crucial for adolescent health and development. Delaying school start times has been proposed as a policy measure to promote healthier sleep patterns among teenagers. Research has shown that later school start times (e..g, 8:30 am) are associated with longer sleep duration in adolescents compared to early school start times (e..g, 7:30 am). However, limited attention has been given to examining the effects of delaying school start times among subgroups. This study aimed to identify subgroups of students that experienced heterogeneous treatment effects of delaying school start times on weekday sleep duration and body mass index (BMI) using the START study. The START study was a three-year evaluation of a natural experiment involving delayed high school start times in the Minneapolis metropolitan area. Generalized linear mixed model trees were used to identify subgroups of students that experienced differential effects of the policy change on weekday sleep duration and BMI. This decision tree-based method accounted for the clustered nature of the data while detecting potential treatment-subgroup interactions without a-priori specifying subgroups of interest. Multiple imputation was used to account for missing data, and the stability of variable selection was assessed using a subsampling approach. Results showed that delaying high school start times had substantially heterogeneous effects on sleep and minimally heterogeneous effects on BMI, with racial minority and lower socioeconomic subgroups experiencing the largest sleep benefits. These findings highlight the importance of examining intervention effects across subpopulations to ensure equitable impacts and identify students who may benefit the most from the policy.