First Annual Thomas R. Ten Have Memorial LectureMay 23, 2012 @ 3:45 p.m. - 5:00 p.m.
Location: BRB II/III Auditorium
Improving Mental Health by Having Statistics Go Viral: Or Should We?
Keynote Speaker: C. Hendricks Brown, PhDProfessor and Director, Prevention Science and Methodology Group
Department of Epidemiology and Public Health
University of Miami
Abstract: Research in mental health is conducted with the ultimate goal of lessening the burdens of mental disorders. But there are major challenges in mental illness prevention, treatment, and service, all due to limitations in existing data, measurement, design, and analysis. Throughout his all too brief life Professor Thomas R. Ten Have pushed through these limits by developing novel statistical methods and their application in collaborative research. It is now left to us to continue down these paths that he laid for us. One of the fundamental challenges he and others recognized was the long lag between the development of statistical methods that could enlighten psychiatric research, and the use of these statistical methods in practice. Too often methods that are inadequate or inappropriate to the task are used. We illustrate the large gap between development and practical use of a number of innovative statistical ideas that have addressed challenges in longitudinal research, community-based randomized trials, mediation analysis, and missing data analysis. We discuss how these methods have naturally diffused and evolved over time and how inappropriate methods, such as "Last Observation Carried Forward" have become entrenched. A change needs to be made to move these new methods into practice more rapidly with precision. We then explore how mixed methods used in implementation science can provide new directions to pursue, as this young field itself is attempting to move evidence based interventions and programs into practice. Finally, some risks are identified, with illustrations of how inappropriate use of statistical methods have sometimes led to widespread societal failures.