The Biostatistics &Data Management core for The University of Texas M. D. Anderson Cancer Center Program Project """"""""Mechanisms of Symptoms of Multiple Myeloma and Its Therapy"""""""" will provide comprehensive support for the planning and conduct of research implemented under this program. This resource will provide guidance for hypothesis testing and refinement, experimental design, implementation of clinical trials, data collection and management, data analysis, and data modeling. It will provide statistical expertise required for the preparation of reports summarizing program achievements. This core will also support the development of innovative statistical methodology required for the analysis of multivariate mixed ordinal and continuous longitudinal data. This methodology will be based on hierarchical, joint longitudinal modeling of symptom response and neuropsychiatric, neurocognitive, and quantitative sensory variables as functions of an underlying state space. The framework developed as part of this research program will facilitate the analysis of missing data and joint modeling of both continuous and discrete data. Finally, this core will support the development, analysis, and interpretation of psychometric instruments designed for the measurement of cancer symptoms and treatment-related symptom burden, and will apply multidimensional scaling and clustering methodology to identify groups of symptoms that occur concurrently with changes in cytokine-immunological activations. The specific, primary objectives of this core are as follows: 1. To develop appropriate statistical models for the analysis of data collected under each of the program projects. 2. To develop and maintain a central database management system for integrating data across all projects and cores. 3. To provide descriptive analysis, hypothesis testing, estimation, and innovative statistical modeling needed by the projects, developmental projects, and other cores to achieve their objectives. 4. To ensure that the results of all projects are based on well-designed experiments and that the results from these experiments are appropriately interpreted.
Showing the most recent 10 out of 133 publications