Dynamic Prediction of Survival with a Longitudinally Measured Marker
Krithika Suresh, PhD Research Assistant Professor, Department of Biostatistics and InformaticsColorado School of Public Health
Dynamic prediction uses patient information collected during follow-up to produce individualized survival predictions at given time points beyond baseline. This allows clinicians to obtain updated predictions of a patient's prognosis that can be used in making personalized treatment decisions. In this talk, we begin by describing two common approaches for dynamic prediction: joint modeling and landmarking. We then propose a novel alternative approach that aims to overcome some of the limitations of the existing methods. Using a Gaussian copula, we specify the joint distribution of a longitudinal marker and failure time conditional on surviving to a prediction time of interest. We illustrate the utility of our method in an application to predicting death for heart valve transplant patients using longitudinal left ventricular mass index information. We describe extensions of this approach to settings with a binary marker.