XXXIst Conference of the Austro-Swiss Region
(ROeS) of the International Biometric Society
Lausanne, Switzerland, September 9-12, 2019
Conference Program
Scientific Committee :
Valentin Rousson (Chair, University of Lausanne)
Andrea Berghold (University of Graz)
Frank Bretz (Novartis)
Dominik Grathwohl (Nestle)
Georg Heinze (Medical University of Vienna)
Dominik Heinzmann (Roche)
Leo Held (University of Zurich)
Torsten Hothorn (University of Zurich)
Zoltan Kutalik (University of Lausanne)
Patrick Taffe (University of Lausanne)
Local Organizing Committee (Unisante, Lausanne)
Dominique Actis-Datta (Chair)
Romain Piaget-Rossel
Alex Randriamiharisoa
Nathalie Rouge
Internet Sites :
ROeS: https://www.ibs-roes.org
Conference: https://wp.unil.ch/ibs-roes2019
Abstracts: https://www.iumsp.ch/owncloud/index.php/s/CEAczi3yS803Wjc
Wi-Fi :
Network: guest-unil
Password: ROES2019
Contact :
Email: [email protected]
Phone: +41 (0)21 314 73 28
Emergency Numbers :
Internal phone: 115
External phone (mobile): +41 (0)21 692 20 00
2
From Controlled Trials to Big Data and Back
Statistical data analyses are sometimes classified as being either exploratory or confirmatory,
while the reality of statistical practice often lies in between. This middle territory is exemplified
by “model selection” issues and Frank Harrell’s famous words: “Using the data to guide the
data analysis is almost as dangerous as not doing so”.
The most accomplished confirmatory statistical analyses are conducted in the context of con-
trolled (clinical) trials, where regulations and guidelines are to ensure a fully protocoled and
planned statistical analysis. On the other hand, we are now living in the era of “big data”
and “data science”, where extreme forms of exploratory data analyses are encouraged with the
hope that data quantity prevails over data quality.
While data science is currently in vogue, there is also some perception that “those who ignore
statistics are condemned to reinvent it”, as Brad Efron once said. It might be a time to
return from the big data paradigm towards more classical approaches and concerns, and to
land somewhere between the two extremes of the purely confirmatory and purely exploratory
data analyses.
The XXXIst ROeS statistical conference will be a timely occasion to try to define what this
“middle ground” should or could be to best meet the expectations of scientists.
We are looking forward to welcoming you in Lausanne and hope that you will have a pleasant
stay and a fruitful conference!
The conference organizers
September 2019
3
Program Overview
TIME ROOM C ROOM B ROOM A
Monday, September 9
09:00-12:30 short course
13:30-17:00 short course
Tuesday, September 10
08:30-09:30 welcome coffee + registration (main hall)
09:30-11:10 plenary session
coffee break
11:30-12:40 parallel session parallel session
lunch box
12:50-13:40 ROeS general assembly
13:50-15:40 parallel session parallel session
coffee break
16:10-17:10 plenary session
17:30-19:00 poster session + welcome drinks (Montreux Jazz Cafe, Lausanne)
Wednesday, September 11
08:30-10:10 parallel session∗ parallel session parallel session
coffee break
10:40-12:10 plenary session
lunch box
13:00-21:30 excursion + conference dinner (for registered only)
Thursday, September 12
08:30-09:50 parallel session parallel session parallel session
coffee break
10:20-11:50 parallel session parallel session
11:55-12:15 Arthur-Linder Prize
lunch box
13:00-14:10 parallel session∗ parallel session
coffee break
14:30-16:00 parallel session parallel session∗
∗ shorter session4
Detailed Program
Abbreviations
KS Keynote Speaker 45 minutes (+10 minutes questions)
IS Invited Speaker 25 minutes (+5 minutes questions)
CS Contributed Speaker 15 minutes (+3 minutes questions)
YS Young Statistician 15 minutes (+3 minutes questions)
AL Arthur-Linder Prize 5-7 minutes
PO Poster
SC Short Course
Monday, September 9, 2019
09:00-12:30 Short Course 1 Room A
SC1 Introduction to causal analysis
Martin Huber
13:30-17:00 Short Course 2 Room A
SC2 Introduction to big data
Martin Spindler
5
Tuesday, September 10, 2019
09:30-11:10 Keynote Session 1 Room C
Welcome / Chair: Murielle Bochud / Valentin Rousson
IS1 Machine learning versus traditional statistical modeling and medical doctors
Maarten van Smeden
KS1 To infinity and beyond: lessons for big data from small experiments
Stephen Senn
11:30-12:40 Bayesian Analysis Room C
Chair: Maria Laura Gosoniu
IS2 Bayesian variable selection methodology for complex health data
Rianne Jacobs
CS1 Assessment of historical controls benefits in new clinical trials
Nicolas Sauvageot
CS2 Bayesian effect selection in structured additive distributional regression models
Helga Wagner
11:30-12:40 Precision Medicine and Biomarker Assessment (Part 1) Room B
Chair: Kaspar Rufibach
IS3 Bayesian statistical learning for cancer drug screening
Manuela Zucknick
CS3 An adaptive enrichment design to react on emerging biomarker data: a real case
Claude Berge
CS4 Evaluation of functional biomarkers with respect to ordinal disease severity
Amita Manatunga
13:50-15:40 Evidence Synthesis and Meta-Analysis Room C
Chair: Leonhard Held
IS4 The march of evidence synthesis: which limits are we pushing?
Georgia Salanti
CS6 Empirical evaluation of ranking metrics in network meta-analysis
Virginia Chiocchia
CS7 A Bayesian two-step dose-response meta-analysis model
Tasnim Hamza
CS8 Comparing methods for variable selection in individual patient data meta-analysis
Michael Seo
CS5 A two-stage prediction model of heterogeneous effects for many treatment options
Konstantina Chalkou
6
13:50-15:40 Innovations in Early and Late Clinical Trials Room B
Chair: Hans Ulrich Burger
IS5 Opportunities for small data: multistate models in clinical trials
Kaspar Rufibach
CS9 Using a non-interventional study to strengthen the evidence collected in Phase III pro-
gram: a hemophilia A case study
Elina Asikanius
CS10 Efficient cut-point analyses
Dominik Grathwohl
CS11 Design and analysis considerations for outcome-based treatment escalation in treat-to-
target studies
Wei Wei
CS12 Comparison of time-to-first event and recurrent event methods in multiple sclerosis trials
Marcel Wolbers
16:10-17:10 Keynote Session 2 Room C
Chair: Patrick Taffe
KS2 Statistics as a condemned building: demolition and reconstruction
Sander Greenland
17:30-19:00 Poster Session∗ Montreux Jazz Cafe, Lausanne
PO1 Machine learning and artificial intelligence in life sciences
Daniel Christen
PO2 Possible association between disruptive sleep patterns in children at six months and their
later development assessed at 18 month, 3 and 5 years: possibilities and challenges when
analyzing large longitudinal data from a population based national registry
Milada Cvancarova Smastuen
PO3 Heterogeneous effects of poverty on cognition
Helmut Farbmacher
PO4 Multivariate matching and propensity scores in retrospective data to conclude about the
benefit of a given medical intervention in ICU patients
Irina Irincheeva
PO5 Survival probability estimation and group comparison for an exogenous binary time-
dependent covariate
Martina Mittlbock
PO6 Exact parametric causal mediation analysis for a binary outcome with a binary mediator
Martina Raggi
∗A price will be awarded to the best poster (each conference participant can vote).
7
PO7 Automated spatio-temporal outbreak detection in low-count settings
Kelly Reeve
PO8 Evaluation of the DMP “Therapie aktiv - Diabetes im Griff” - results for the established
program phase
Regina Riedl
PO9 A comparison of statistical methods for allocating disease costs in the presence of inter-
actions
Jean-Benoıt Rossel
PO10 Machine learning based prediction of insufficient herbage allowance with automated feed-
ing behavior and activity data
Abu Zar Shafiullah
PO11 Gender-related aging trajectories in Western Europe
Valentina Shipovskaya
PO12 Bring more data! - a good advice? Removing separation in logistic regression by increasing
sample size
Hana Sinkovec
PO13 Methods to analyze continuous outcomes by incorporating baseline data in individual
participant data meta-analysis of non-randomized studies
Lamprini Syrogiannouli
PO14 Generalizing effect sizes for differences with an alternative to Cohen’s d coefficient
Carl Taswell
PO15 Performance evaluation of regression splines for propensity score adjustment in post-
market safety analysis with multiple treatments
Yuxi Tian
PO16 Outcomes truncated by death in RCTs: an application on preterm infants to estimate the
survivor average causal effect
Stefanie von Felten
PO17 Identification of features before adrenal surgery indicating adrenal insufficiency following
surgery
Xiao Wang
PO18 Assessing the impact of a heterogeneity prior in Bayesian hierarchical models in terms of
added or subtracted samples to the data (effective sample size)
Manuel Wiesenfarth
PO19 Model-averaged confidence distribution
Jimmy Zeng
8
Wednesday, September 11, 2019
08:30-10:00 Reproducibility in Biomedical Research Room C
Chair: Georg Heinze
CS13 Impact of a collaborating biostatistician on the quality of research: a meta-science study
protocol
Eva Furrer
CS14 A new standard for the analysis and design of replication studies
Leonhard Held
IS6 Reproducibility of animal trials: what can be done about it?
Florian Frommlet
CS15 Bland & Altman quo vadis?
Patrick Taffe
08:30-10:10 Machine Learning and Artificial Intelligence Room B
Chair: Maarten Van Smeden
CS16 Insights of plant electrophysiology - using signal processing techniques and machine learn-
ing algorithms to associate tomatoes reaction to external stimuli
Elena Najdenovska
CS17 Statistical modeling versus machine learning: can flexible methods provide accurate pre-
dictions AND interpretable effects?
Christine Wallisch
IS7 Machine learning in pharma: where are we on this journey?
Markus Lange
IS8 Uniform inference in high-dimensional Gaussian graphical models
Martin Spindler
08:30-10:10 Multiple Testing and Adaptive Designs1 /
Causal Inference in Epidemiology (Part 1)2 Room A
Chair: Marcel Wolbers
CS18 Treatment selection in multi-arm multi-stage designs: an application to surgical trials1
Alexandra Blenkinsop
CS19 Optimized multiple testing procedures for confirmatory subgroup analysis based on a
continuous biomarker1
Alexandra Graf
CS20 A closed omnibus test1
Sonia Zehetmayer
CS21 Sufficient dimension reduction for feasible and robust estimation of average causal effect2
Xavier de Luna
9
CS22 Using higher moments to test requirements for causal inference2
Wolfgang Wiedermann
10:40-12:10 Keynote Session 3 Room C
Chair: Torsten Hothorn
IS9 Plea for a marriage of machine learning and causal inference
Els Goetghebeur
KS3 Experimenting in equilibrium
Stefan Wager
10
Thursday, September 12, 2019
08:30-09:50 Young Statisticians Room C
Chair: Andrea Berghold
YS1 Isotonic regression for growth charts of children
Alexandre Mosching
YS2 New statistical methods to analyse evolve and resequence genetic data
Marta Pelizzola
YS3 Rare event meta-analysis of count data: a journey across adverse settings
Romain Piaget-Rossel
YS4 A nonparametric approach of interpreting the ABC-algorithm
Victoria Racher
08:30-09:50 Model Selection, Prediction and Overfitting (Part 1) Room B
Chair: Michael Schemper
CS23 Robust and unbiased estimation framework in high-dimensional setting
Elise Dupuis
CS24 Quantifying degrees of necessity and of sufficiency in cause-effect relationships for cate-
gorical and survival outcomes
Andreas Gleiss
CS25 A modified Firth correction for Poisson regression model
Ashwini Joshi
CS26 Comparison of likelihood penalization and variance decomposition approaches for the
derivation of binary logistic regression based low-dimensional clinical prediction models
Anna Lohmann
08:30-09:50 Precision Medicine and Biomarker Assessment (Part 2) Room A
Chair: Susanne Strohmaier
CS27 Tree-based search for predictive factors based on observational studies
Julia Krzykalla
CS28 Sequential models in regression of vaccine-induced antibody titers in stem cell transplant
recipients
Janina Linnik
CS29 Inference in ROC surface analysis via a trinormal model-based testing approach
Christos Nakas
CS30 Challenges and possibilities when analyzing big data from several national registries - use
of antibiotics in patients treated for morbid obesity (outside main session topic)
Milada Cvancarova Smastuen
11
10:20-11:50 Causal Inference in Epidemiology (Part 2) Room C
Chair: Patrick Taffe
IS10 Direct and indirect effects based on changes-in-change
Martin Huber
CS31 Sensitivity analysis after propensity score matching - how strong would an unmeasured
confounder have to be to explain away the treatment effect?
Ulrike Held
CS32 Modified causal forests for estimating heterogeneous causal effects
Michael Lechner
CS33 Paradoxical findings in observational research - a new example from coronary artery bypass
surgery
Hanno Ulmer
10:20-11:50 Statistical Genomics Room B
Chair: Zoltan Kutalik
IS11 Two-sample tests on deep learning embeddings
Christoph Lippert
CS34 Simultaneous estimation of heritability, genetic confounding, and bi-directional causal
effect from GWAS summary statistics
Liza Darrous
CS35 Leveraging correlated risks to increase power in genome-wide association studies
Ninon Mounier
CS36 A global-local variational approach for detecting hotspots in multiple-response regression
Helene Ruffieux
11:55-12:15 Arthur-Linder Prize Ceremony Room C
Chair: Valentin Rousson
AL1 Subgroup identification in clinical trials via the “predicted individual treatment effect”
Nicolas Ballarini
AL2 Re-estimation improved two Framingham cardiovascular risk equations and the pooled
cohort equations: nationwide registry analysis
Christine Wallisch
13:00-14:00 Miscellaneous Room C
Chair: Elise Dupuis
CS37 How to interpret over/under-dispersion when modeling person-time incidence rates with
Poisson models
Rossella Belleli
CS38 Non-normality of the error term in linear regression: least squares estimator and testing
Caroline Giacobino
12
CS39 Modeling extremes of flu episodes and detecting hospital congestion
Setareh Ranjbar
13:00-14:10 Model Selection, Prediction and Overfitting (Part 2) Room B
Chair: Martina Mittlbock
CS40 Transparent statistical models in the times of machine learning
Georg Heinze
CS41 New approaches for selective inference: an independent comparison
Michael Kammer
IS12 Searching for truth or profit in data
Marcus Hudec
14:30-16:00 Survival and Event History Analysis Room C
Chair: Hanno Ulmer
CS42 Evaluating daily adherence to drug prescription from censored observations
Isabella Locatelli
CS43 Probability of random cancers as supported by the data
Janez Stare
CS44 Additive and multiplicative hazard models in practice - a series of case studies from clinical
epidemiology
Susanne Strohmaier
IS13 Sampling where the events are
Jan Beyersmann
14:30-15:30 Model Selection, Prediction and Overfitting (Part 3) Room B
Chair: Dominik Grathwohl
CS45 Fitting linear mixed-effect models to right-skewed data in a surgical trial with an unbal-
anced design
Stefanie Hayoz
CS46 Evaluating the cost of simplicity of a score to predict a binary gold standard
Katia Iglesias
CS47 From classical statistics to machine learning and back - leveraging evidence in clinical
research
Cheng Chen
13
Conference Venue
Building Genopode (ground floor)
Campus of Dorigny
University of Lausanne (UNIL)
Metro m1 stop: UNIL-Sorge (5 minutes walk to Genopode)
From Lausanne main railway station:
1. use metro m2 from Lausanne-Gare to Lausanne-Flon, and then
2. use metro m1 from Lausanne-Flon to UNIL-Sorge.
14
Social Program
• Welcome drinks and poster session venue (September 10, 17:30-19:00)
Montreux Jazz Cafe
Artlab EPFL, Place Cosandey, 1015 Lausanne
Phone: +41 (0)21 693 89 00
10 minutes walk from the conference venue
• Excursion and conference dinner (September 11, 13:00-21:30, for registered only)
Meeting point: 13:00 at the conference desk (to leave by coach at 13:15)
Destinations: Montreux (5 km walk along the lake), Chillon Castle (guided tour),
Lavaux (UNESCO World Heritage, 2.5 km walk in the vineyards)
Dinner: 19:00-21:00, Auberge de Rivaz, Route de Sallaz 6, 1071 Rivaz,
Phone: +41 (0)21 946 10 55
Ending: 21:30 at Lausanne main railway station
See the internet site of the conference for more details (https://wp.unil.ch/ibs-roes2019/excursion).
15
We thank our sponsors!