Title “Multivariate Modelling: Methodological development and clinical applications”.
Abstract:
In this talk I will describe some of the methodological and collaborative research I work on. There is a need for new statistical methods in clinical research to improve diseases detection and prediction of prognosis by using information collected over time. Delay on the detection of a disease or incorrect prognosis can have a devastating impact on patients, and it is usually linked to economic burdens of the National Health Service and communities. Hence, it is not surprising that in many medical disciplines research focuses on identifying biomarkers and risk factors to correctly predict patient outcomes. The use of clinical and demographic data collected over time can be particularly powerful in discriminant analysis for a better clinical classification. Recent results on sample size calculation for the probability of correct classification in linear discriminant analysis will also be presented and illustrated using biomedical data.