The Statistics Core will ensure that uniform, accurate and complete data are collected across projects. The Core will perform primary analyses for key aims in Projects 1-3 and apply procedures to investigate consistency of results across projects. Cleaned data sets will be provided to investigators for secondary analyses and eventually for use by the public. The Core unit will take responsibility for investigators, internal Steering Committee, and the Advisory Board receiving timely reports on study progress and findings. The core will stay abreast new statistical developments and serve as a resource to investigators on statistical issues. Statistical analyses will use regression methods for translating scores from one instrument into another conditionally on population characteristics. These translations will be validated across projects. Hierarchical models will be used for longitudinal analyses examining time trends in response to treatment, and will emphasize both mean and variance structure. The Core will also apply structural equations models to pool data across measures, and study the joint relationships of health measures on latent dimensions of health and HRQOL All analyses will include careful checking of model fit and examination of potential biases introduced by missing data.
Showing the most recent 10 out of 52 publications
