Statistics Seminars: Modelling of biomarkers to predict clinical outcomes
23 November 2009 14:15 in CM212
Biomarkers are biometric or biochemical quantities in the human body that are frequently measured to; (1) identify those individuals at risk of disease, (2) those individuals with disease but at an earlier stage than would be possible clinically, and (3) those individuals not responding to treatment. They therfore offer an opportunity to target therapy or preventative measures more appropriately than might otherwise be possible. Clearly their sucess depends crucially on developing an appropriate statistical model relating biomarker levels (or changes in levels) to clinical events/outcomes. This talk will discuss some of the statistical approaches currently taken, particularly Bayrsian methods, and illustrate them using a study on yolk sac development during pregnancy to predict first trimester miscarriage.
Contact email@example.com for more information