The aim of the thesis is to investigate how to optimize the design and analysis of randomized and non-randomized therapeutic studies, in order to increase the validity and reliability of causal treatment effect estimates, specifically in heterogeneous diseases. The following research questions will be addressed:
1) What are the benefits of more advanced statistical analyses to estimate treatment effects from RTCs in heterogeneous diseases?
a. What is the heterogeneity in acute neurological diseases with regard to baseline severity and further course of the disease?
b. What is the potential gain in efficiency of covariate adjustment and proportional odds analysis in RCTs in Guillain-Barré syndrome (GBS)?
2) What is the validity and reliability of the RD design compared to an RCT to estimate causal treatment effects?
a. What are threats to the validity of the RD design to estimate treatment effects compared to an RCT?
b. How efficient is the RD design to estimate treatment effects compared to an RCT?
c. What are the potential benefits of an alternative assignment approach in an RD design?

E.W. Steyerberg (Ewout) , B.C. Jacobs (Bart)
Erasmus University Rotterdam
Department of Public Health

van Leeuwen, N. (2019, September 11). Design and Analysis of Randomized and Non-randomized Studies: Improving Validity and Reliability. Erasmus University Rotterdam. Retrieved from