Kush is an Assistant Professor of Neurology at Harvard Medical School and a Senior Biostatistician with the Institutional Centers for Clinical and Translational Research (ICCTR) at Boston Children's Hospital. He received his Ph.D. degree in Biostatistics and Master’s degree in Electrical Engineering from the University of Illinois at Chicago.
As a part of statistical methodology development, Kush has published in the field of sample size determination problems for binary longitudinal data, hypothesis testing of skewed data with small sample sizes, analysis techniques for intensive longitudinal data within a Bayesian framework using mixed-effects location and scale models, and novel modeling techniques based on random-effects models for high dimensional dataset such as Functional Magnetic Resonance Imaging (fMRI). His methodological work has appeared in notable statistical journals such as Technometrics and Statistics in Medicine.
Kush is also a co-author of several statistical software applications: fMRIview, which allows users to analyze fMRI event-related designs using random-effects regression models; MIXZIP, which allows the user to estimate mixed-effects Zero-Inflated Poisson (ZIP) regression models using the maximum marginal likelihood approach; and RMASS, which computes sample size for three-level mixed-effects linear regression models for the analysis of clustered longitudinal data.