Statistical Model Development in Epidemiological Research

Most epidemiological and clinical studies are observational studies; therefore, statistical analyses are required to reduce bias in our observations and to uncover patterns from large volume of data. Statistical models can also help researchers to understand mechanisms that underlie those data. Currently available statistical analysis software packages are not tailored to the needs of epidemiological studies. They are either too complicated or too expensive, and most importantly, do not contain models that are readily applicable to most epidemiological studies. The purpose of the proposed research is to refine modeling strategies most suitable for epidemiological and clinical studies, and to develop a user-friendly software package (Statistical Software for Epidemiology Research or SSER) that empowers researchers to conduct statistical analyses with right tools. The findings and the product of the proposed research will allow researchers build statistical models according to research hypotheses and available experimental data structure, to correct bias, to select suitable model and to interpret research results based on the best model and the hypothesis. The research and the resulting software will have a significant impact on epidemiological and clinical studies, and will benefit health care industry, pharmaceutical industry, food industry, academic institutions, as well as government agencies.