Data integration for causal inference in behavioral health

点击翻译按钮获取中文摘要

Behavioral health, broadly defined to include mental health and substance use, includes many of the most pressing public health problems of our time. The transition to a data-rich, web-interconnected society has generated an opportunity to generate solutions, but it necessitates a paradigm shift in workforce training in data analytics. The goal of this training program is to train scholars to become leaders in the use of advanced computational methods and designs to estimate causal effects in behavioral health. To accomplish this goal, we will provide rigorous training and high-quality mentorship in: 1) the science of behavioral health; 2) computational and analytic tools to manage, analyze, and integrate complex data sources; and 3) causal inference methods to take full advantage of these data. Trainees will receive interdisciplinary team-based training and will acquire a deep understanding of all three areas. This training program will capitalize on the rich resources for behavioral health, analytic and computational methods, and biostatistics at the Johns Hopkins Bloomberg School of Public Health (JHSPH) and the broader University. The program will be housed in the Department of Mental Health but the 5 trainees per year will come from any of the four social science oriented departments at JHSPH: 1) Mental Health; 2) Health Behavior & Society; 3) Health Policy & Management; and 4) Population Family & Reproductive Health. Further, the training grant will leverage close connections with data scientists, statisticians, and computer scientists from across the University. Trainees will obtain the skills and experiences needed to lead multi-disciplinary, collaborative research teams. Trainees will undertake a rigorous program of coursework in the core domains of public health and behavioral health including behavioral and social science, epidemiology, biostatistics, data science, population health informatics, causal inference, and research ethics. In addition, each trainee will take additional elective courses in social and behavioral perspectives on mental health and substance use, informatics and computational skills, and causal and statistical inference. Trainees will participate in a year-long seminar on analytics for behavioral health, a bi-weekly seminar to discuss research in progress and professional development, ongoing mentored research projects, and integrative activities to complement their didactic curriculum. The focus area of the program builds on strengths within JHSPH; these areas also are highlighted as priorities by OBSSR, NIMH, and NIDA. The trainees will be supported by an experienced group of 21 core faculty and each trainee will be co-advised by one of 9 affiliated faculty with methodological expertise. The training program director, Dr. Elizabeth Stuart, is a national leader in analytic tools for behavioral health, and will be supported by a 4- member internal Executive Committee and a 5-member external Advisory Committee. The overarching aim of the program is to identify and train scholars who will become leaders in using a diversity of advanced analytic tools and data to answer key questions in behavioral health.