This Independent Scientist Award (K02) is proposed for the applicant to acquire necessary skills to apply the developmental psychopathology conceptual framework and innovative statistical techniques for categorical longitudinal data to creating a model for the relationships among alcohol use disorders (AUD) and other mental disorders in adolescence. Based on empirical findings to date, the specific hypotheses focus on antisocial disorders (i.e., conduct disorder, oppositional defiant disorder) and negative affect disorders (i.e., mood and anxiety disorders) as possible predictors, consequences, or moderators of the structure, course and consequences of adolescent AUD. The applicant, trained as a child clinical psychologist and adult psychiatrist, is the Scientific Director of the NIAAA-funded Pittsburgh Adolescent Alcohol Research Center (PAARC). The career development plan focuses on the acquisition of a thorough foundation in statistical modeling techniques and related methodological issues, including the implications of sampling strategies, missing data imputation, model selection, and controversies concerning causal inference from observational data. The applicant will learn statistical methods based on regression for modeling time-dependent relationships among continuous and categorical variables. The focus on methods for categorical variables is relevant to longitudinal research involving symptom and diagnostic categories. Relevant statistical techniques include methods for observed variables, including proportional hazards and random regression modeling and methods for latent variables, including latent class analysis, latent transition analysis, and growth mixture modeling with latent trajectory classes. Bayesian approaches to model selection and causal inference will also be considered. These methods will be specifically applied to examining the relationships among AUD, antisocial disorders, and negative affect disorders using longitudinal data from PAARC (n=1000 adolescents). Methods for evaluating the extent and the influence of sampling bias will assessed through comparison of the model generated using PAARC data with models generated with other data sets, including studies using high-risk and community sampling approaches. The integration of the concepts of developmental psychopathology with innovative longitudinal statistical modeling methods will contribute to the applicant's long-term career goal to advance research on adolescent AUD by clarifying the importance of psychopathology in determining the structure, course and consequences of adolescent alcohol abuse and dependence.
Showing the most recent 10 out of 72 publications
