Grouping Growth Trajectories Leading to Early Childhood Obesity: An Application of Mixed Effects Model
This seminar will present an innovative use of the mixed effects model in grouping children with distinct trajectories leading to early childhood obesity. The heterogeneity in the subject-specific effects of the mixed effects model was used to classify 488 children with differential pathways to early childhood obesity. Further investigation was performed to explore the association of the co-morbid conditions and correlates with the trajectory based groups. The presentation covers methodological concepts of clustering longitudinal data as well as clinical effectiveness through the exploration of underlying co-morbid conditions of differential pathways to obesity.
Dr. Hossain is the manager of the Nemours Biostatistics Core, a Research Scientist in Nemours Biomedical Research, and an adjunct faculty of statistics at the University of Delaware. Since joining Nemours in 2007, Dr. Hossain has been collaborating with Nemours clinicians/researchers on designing clinical trials, pre-clinical, epidemiological and survey studies, calculating power and sample size, randomization, authoring statistical section in research protocols, managing and analyzing data, as well as disseminating results through manuscripts, abstracts, presentations and posters. He teaches graduate and undergraduate statistics courses at the University of Delaware, as well as provides training to research fellows at Nemours on statistical and design skills. Dr. Hossain has expertise in variety of study designs, and in effective application of linear and nonlinear regressions and mixed effects modeling, generalized estimating equation, multilevel analysis, survival analysis, time series analysis, cluster analysis, principal component analysis, factor analysis, propensity score estimation, bootstrapping, cross-validation and G-estimation.
Thursday Tech Talks, sponsored by the Delaware Clinical & Translational Science ACCEL program and the Christiana Care Value Institute, are monthly seminars intended to present statistical, epidemiological and bioinformatics methods and how they can be applied to medical research studies. Presentations focus on the technical aspects of methods as well as their application to real world situations.
These free talks are held the first Thursday of each month at noon at Christiana Hospital but can be viewed from your home or office computer. Earn CMEs by participating in-person or online. Lunch is served and all are welcome to attend.
To see the full calendar of events, visit the Value Institute Events page or the ACCEL website, or subscribe to the ID Series mailing list.
Contact Sarahfaye Dolman at sarahfaye.f.dolman@christianacare.org with any questions.