Veranstaltung ausschließlich für Fachpublikum
Jan van den Brand, Nijmegen, the Netherlands
Das Berlin Epidemiologic Methods Colloquium (BEMC) bietet Epidemiologinnen und Epidemiologen eine lokale Plattform, auf der wir uns vernetzen, neue Methoden erlernen und aktuelle Themen in diesem Bereich diskutieren können. BEMC bietet interaktive monatliche BEMC Talks und BEMC JClub in englischer Sprache.
The foundations of personalized medicine rest on the accurate prediction of disease course before and after treatment. Patients with chronic diseases have many biomarkers measured at frequent intervals at the out-patient department of a hospital. Moreover, measurements are increasingly taken longitudinally by patients themselves for example using mobile devices. However, such longitudinal data in healthcare remains untapped, as most current prediction models only use cross-sectional measurements and time-to-event data. On the other hand, state-of-the-art joint models that combine longitudinal and time-to-event data show great potential for personalized medicine. The R-package JMBayes by Rizopoulos and the lcmm package by Proust-Lima made the technology widely available and spurred further development of joint models for multiple longitudinal biomarkers, and multiple outcomes. However, developing a joint model is still not straightforward. In this talk I will offer an introduction to joint modeling and the rationale behind the technology. Next, I will go over some of the basic steps required to fit and evaluate a joint model, finally I will discuss common issues faced when building a joint model. In the discussion we could talk about some of the future prospects of joint modeling and prediction modeling in general.
Links
Register here: https://docs.google.com/forms/d/e/1FAIpQLSfHEP2shY1mQx8Z-LS7IfBOyJJ8mp0DAjA8j39dThl3ydYk0A/viewform?usp=sf_link
or visit the BEMC Website to register.