Samiran Ghosh, Weill Cornell Medical College

Thomas R. Ten Have (Courtesy of UPENN)

Nearly 60 participants attended the 13th Symposium on Statistics in Psychiatry (PDF download), hosted by the division of biostatistics in psychiatry at Columbia University and New York State Psychiatric Institute on May 2. This is an annual one-day event that brings together academic statisticians involved in collaborative and methodological research and psychiatric researchers to address statistical challenges arising from the design and analysis of psychiatric data.

This year’s symposium was marked by the sudden demise of colleague Thomas R. Ten Have of the University of Pennsylvania (UPENN). Ten Have was a founding member of the conference and an active member of the program committee, despite his poor health. The symposium was his vision of a place for statisticians who work in psychiatry to bring technical questions often posed by mental health research.

The day started with a brief note in memory of Ten Have by Eva Petkova of New York University (NYU). Colleagues from various institutions also shared their memories. Two keynote presentations, three technical talks, and several poster presentations followed.

Xiao-Li Meng from Harvard University presented the first keynote, titled “Mental Exercises for a Mental Health Study: Is This a Simpson’s Paradox?” Meng provided more than one example in which such a paradox exists in published medical literature and its possible solution.

Lassell Lu of NYU, Knashawn H. Morales of UPENN, and Yongtao Guan of Yale University presented technical talks, with topics ranging from variable selection, multilevel models to measurement error model applicable toward data analytic problems related to psychiatry.

Six posters were presented during lunch and coffee breaks, and the technical program concluded with the second keynote presentation by Andrew C. Leon of Weill Cornell Medical College (WCMC), titled “Evolution of Psychopharmacology Trial Design and Analysis: Six Decades in the Making.” Leon presented a sequential history of the clinical trial and some of the major events that guided the path of psychopharmacology to the present.

Finally, in honor of Ten Have, members of the program committee unanimously voted to rename the conference the “Thomas R. Ten Have Symposium on Statistics in Psychiatry.” The conference goes back to 1999, with the first one held at the New York State Psychiatric Institute as an informal joint forum with participating statisticians from Columbia University and UPENN. Since then, Yale University, NYU, and WCMC have joined.