This proposal represents a request for a NIDA Research Scientist Development Award (K01) to provide training and research experience that will integrate complex developmental theory surrounding the etiology of substance abuse and dependence with newly emerging statistical methods. The candidate's goal is to advance her growth as an independent investigator and to apply appropriate and powerful quantitative methods to the refinement of substance theory that may optimally guide the development of intervention efforts. Thus, the applicant seeks additional training in the technical aspects of newly developed quantitative methods and in the consideration of state-of-the-art statistical procedures in the planning and design phases of longitudinal research aimed at uncovering the heterogeneity of developmental pathways to substance use disorder. The combination of training and supervised research provides the groundwork for the candidate's ability to independently bridge different disciplines and to ultimately function as an independent researcher and as the leading member of a multidisciplinary research team. The proposed research project will apply state-of-the-art statistical methods to the analysis of newly available longitudinal and cross-sectional data that include measurement of the widest array of individual and contextual risk factors available to date. The goal will be to identify multiple and prominent risk pathways that will guide the theoretical characterization of individual differences in risk for substance abuse and dependence. To achieve this goal, the following specific aims will be addressed: 1) Distinguish between substance use trajectories leading to substance use disorder vs. less severe levels of use 2) Identify prominent and unique risk constellations/pathways that predict the most severe outcome trajectories 3) Determine the attributable risk associated with each of the prominent risk pathways in order to evaluate the potential for reduction of substance abuse and dependence Existing data from three data sets will be utilized that include both nationally representative and case-control samples. Multiple statistical techniques for studying both growth and mediational effects will be used including group based semi-parametric techniques, tree-based methods, and diverse latent class approaches.
Showing the most recent 10 out of 29 publications
