Course with Hein Putter & Theodor A. Balan (3 Nov 2017)

The Czech National Group of the

International Society for Clinical Biostatistics

in cooperation with

Institute of Computer Science, Czech Academy of Sciences in Prague &

Faculty of Mathematics and Physics, Charles University in Prague

(i) Informing the Course Secretary (CS) about your intent to pay the registration fee and asking her for an approval to register for the course

Note: Please, note that the course capacity is limited to a maximum of 30 participants (see the CS' contact details below). Should the course reach its capacity, a notification will be placed here.

All domestic (CZK) and international transfers (EUR), including Slovakia:

Account number: 2100009829 / 2010

Bank account owner: Mezinárodní společnost pro klinickou biostatistiku v České

republice, z.s.

IBAN: CZ3820100000002100009829

BIC code/SWIFT: FIOBCZPPXXX

Bank name and address: Fio banka, a.s., V Celnici 1028/10, Praha 1, Czech Republic

Course Venue:

Institute of Computer Science AS CR

Pod Vodarenskou vezi 2

182 07 Prague

Czech Republic

Web: http://www.ustavinformatiky.cz/?id_jazyk=en

Course Secretary:

Ms. Lenka Semeráková

Dept. of Medical Informatics & Biostatistics

Institute of Computer Science AS CR

Tel. +420 266 053 640

E-mail: semerakova@cs.cas.cz

Course Outline

Frailties are random effects in models for time-to-event data. They represent unobserved heterogeneity and can be used for two purposes: modelling dependence in clustered failures or recurrent events (shared frailties), and explaining lack of fit of univariate survival models, such as deviation from the proportional hazards assumption (individual frailties). Shared frailty models are increasingly used in applied research; a Google Scholar search of “shared frailty” yields 4490 results for 2016, compared to 3100 for 2011 and 1930 in 2006. Individual frailties on the other hand have recently been used to explain selection effects in epidemiological studies and to shed light on a number of paradoxes in epidemiology, like the obesity paradox.

The aim of this course is 1) to explain the selection effects of unobserved heterogeneity (individual frailty) in epidemiological studies and to illustrate how many strange artefacts can be explained from a frailty point of view, and 2) to show how to use shared frailty models in practice. With respect to the latter, we will review classical results and discuss practical aspects, such as software to fit frailty models, properties of different frailty distributions, and possible ways of extending such models.

Course' Content

TBA

Tentative Course Schedule

8:00 - 9:00 Registration

9:00-12:00 Morning session (includes 20 min coffee break)

12:00-13:00 Lunch break

13:00-16:00 Afternoon session (includes 20 min coffee break)

16:10-16:30 Closing discussion and course evaluation