Innovative Modeling of Puberty and Substance Use Risk

描述（由申请人提供）：青春期过渡是精神病理发展的最关键时期之一，由于在此期间发生了广泛而同步的激素，身体和社会角色变化，因此具有后期药物使用的风险。目前的发育理论认为，激素浓度对青春期过渡期间的危险行为具有直接的神经化学效应和间接的社会心理效应。评估这些直接和间接影响将为儿童和青少年构建有效的早期药物使用预防或干预方法提供关键知识。然而，由于(a)青春期特征的测量误差和(b)缺乏足够的统计方法来区分直接和间接影响，缺乏这方面的知识。根据PA-03-044，“使用现有数据集的精神病理学风险因素”，我们建议使用来自nih资助的研究（5R01MH058393-03）的特殊数据集来评估激素浓度对青春期过渡时期物质使用风险的直接和间接影响。我们的新分析方法是两种先进统计技术的综合和扩展，以便既纠正青春期的测量误差，又评估基于统计因果推理的直接和间接影响。我们的目标是：(1)利用现有方法和新方法评估激素浓度对青春期过渡期间物质使用风险的直接和间接影响，并比较两者的结果。(2)开发一种更准确和信息丰富的方法和计算工具，用于通过贝叶斯方法模拟青春期成熟过程并校正测量误差。(3)考虑潜在中介和/或潜在结局变量，扩展我们的广义中介模型，以评估直接和间接中介效应。拟议的工作在物质使用和统计应用领域都具有创新性。该研究有望通过检查物质使用和滥用风险与青春期过渡期间激素上升之间的途径，填补青少年早期发育物质使用的知识空白。这也将有助于对纵向青春期成熟过程的建模和复杂的中介机制的统计文献。

DESCRIPTION (provided by applicant): The pubertal transition represents one of the most critical periods for the development of psychopathology and risk for later substance use due to extensive and simultaneous hormonal, physical, and social-role changes occurring during this period. Current developmental theories suggest that hormone concentrations have a direct neurochemical effect and indirect psychosocial effect on risk behavior during the pubertal transition. Assessing these direct and indirect effects will provide critical knowledge for constructing effective early substance use prevention or intervention approaches for children and adolescents. However, such knowledge is lacking due to (a) measurement error in characterization of puberty and (b) lack of adequate statistical methodology to differentiate direct and indirect effects. In response to the PA-03-044, "Risk Factors of Psychopathology Using Existing Data Sets", we propose to assess the direct and indirect effects of hormone concentrations on substance use risk during the pubertal transitional period using an exceptional dataset from an NIH-funded study (5R01MH058393-03). Our novel analytical approach is a synthesis and extension of two advanced statistical techniques in order to both correct for the measurement error in puberty and to assess the direct and indirect effects based on statistical causal inference. Our aims are: (1) To assess direct and indirect effects of hormone concentrations on substance use risk during the pubertal transition using existing and new methods and to compare the results from both. (2) To develop a more accurate and informative method and a computational tool for modeling the pubertal maturation process and correcting for measurement error via a Bayesian approach. (3) To extend our generalized mediational model for assessing direct and indirect mediational effects considering latent mediator and/or latent outcome variables. The proposed work is innovative both in the field of substance use and statistical application. The study is expected to fill the gap of knowledge in early adolescent developmental substance use by examining pathways between substance use and abuse risk and the rise of hormones during the pubertal transition. It will also contribute to the statistical literature on modeling longitudinal pubertal maturational process and complex mediating mechanisms.