Ms. Shi's main collaboration has been with the Gastrointestinal Oncology Service, Gastric and Mixed Tumor Service, and currently with members from various services. She has been involved in many projects including 1) assessment of treatment effects on outcomes for various cancer populations, and in vitro and in vivo animal studies, 2) developing prognostic models on outcomes for a variety of cancer patients, 3) assessment of molecular and genetic markers in predicting outcomes for many cancer populations, 4) evaluation of health related quality of life and cognitive functions following treatments, 5) estimation of cumulative incidence rate of secondary or primary malignancies and standardized incidence/mortality ratio for cancer populations, individuals with Bloom's Syndrome and colorectal adenomas, 6) comparisons of accuracy of diagnostic imaging techniques and mortality probability models for cancer patients, and 7) assisting in statistical design for clinical trials and a randomized, placebo-controlled intervention trial to prevent gastric cancer.
Ms. Shi's current research interests lie in 1) developing prognostic models, 2) missing data structure and imputation, and 3) analysis of longitudinal data. Her main focus is on multi-dimensional exploration of data, dimension reduction techniques, data transformation and imputation, categorization of prognostic factor in univariable and/or multivariable settings, verification of model fitting, and multiple testing adjustments.