Classification of Substance Use Severity and Treatment Outcomes: NIDA CTN Studies

摘要 本R21/R33应用是对RFA-DA-09-020的响应。它寻求支持，以解决一项研究 本RFA推荐的领域<$NIDA临床试验中多项试验的次要数据分析 网络（CTN）。通过利用CTN从多个试验中获得的丰富的多中心治疗数据， 在R21阶段（第1年），我们寻求：1）从原始数据中识别和创建可分析的平面数据文件 试验数据（堆叠格式），2）通过编程评分生成研究变量和码本 多种标准化临床仪器的算法，3）评估缺失数据的程度，以及4） 检查基线临床概况（物质使用和诊断，心理健康和生活质量） 按性别和种族/民族列出试验参与者，以指导R33阶段的工作。在R33 阶段（2-4年），我们计划评估分类的结构，同时和预测效度 和物质使用诊断的维度模型。多中心CTN试验提出了一种 无与伦比的，极好的机会，探索和评估这些临床和科学上重要的 与治疗结果和新出现的《诊断和统计手册》直接相关的问题 第四版（DSM-V）。R33的研究目的是1）研究DSM-IV的测量等效性 按性别和种族/民族列出的诊断问题，2）评价 DSM-IV使用分类、维度和整合混合模型进行物质使用诊断 诊断分类，3）确定经验定义的药物依赖亚型（综合 混合模型）在预测随后的结果，并探讨调解人和主持人的 相关性，以及4）评估临床定义的共病药物依赖亚型在预测 随后的结果，并探讨协会的调解人和主持人。性别 在所有分析中，将通过分层分析或纳入 适当的互动方式。

ABSTRACT This R21/R33 application is in response to RFA-DA-09-020. It seeks support to address one research area recommended by this RFA ¿Secondary data analyses of multiple trials in the NIDA Clinical Trials Network (CTN).¿ By capitalizing on the CTN¿s wealth of multisite treatment data from multiple trials, we seek to, in the R21 phase (Year 1), 1) identify and create analyzable flat data files from the raw trial data (stacked format), 2) generate study variables and codebooks by programming scoring algorithms for multiple standardized clinical instruments, 3) evaluate the extent of missing data, and 4) examine baseline clinical profiles (substance use and diagnoses, mental health, and quality of life) of trial participants by gender and race/ethnicity in order to guide the R33 phase of work. In the R33 phase (Years 2-4), we plan to evaluate the construct, concurrent, and predictive validity of categorical and dimensional models of substance use diagnoses. The multisite CTN trials present an unparalleled, excellent opportunity to explore and evaluate these clinically and scientifically important issues directly related to treatment outcomes and to the emerging Diagnostic and Statistical Manual Version-IV (DSM-V). Study Aims for R33 are to 1) investigate measurement equivalence of DSM-IV diagnostic questions by gender and race/ethnicity, 2) evaluate the construct and concurrent validity of DSM-IV substance use diagnoses using categorical, dimensional, and integrated hybrid models of diagnostic classification, 3) determine empirically defined drug dependence subtypes (an integrated hybrid model) in predicting subsequent outcomes and to explore the mediator and moderators of the associations, and 4) evaluate empirically-defined comorbid drug dependence subtypes in predicting subsequent outcomes and to explore the mediator and moderators of the associations. Gender/sex differences will be examined in all analyses either through stratified analyses or the inclusions of interaction terms as appropriate.