Advanced Topics in Biostatistics (AMC PhD Program)

Information about the course Advanced Topics in Biostatistics.

Go to application form.


The introductory course Practical Biostatistics discusses and provides training in basic statistical tools and models for data­ analysis. These are suitable for answering simple questions based on data structures that have been obtained in standard experimental designs. These techniques are often insufficient, or at least suboptimal.


In the first class, we explain the theory behind statistical principles, such that the participant is better able to understand the information that can be derived from a fitted regression model. In the second class, we explain how to make beautiful and informative figures. Next, we give a general framework of the structure of regression models, which helps in fitting complex relations between predictions and outcomes.

In the final classes, we describe some specific types of models (survival, longitudinal and/or joint models). By the end of this course, the participant has gained a deeper understanding of the assumptions behind several statistical models, also ones that are more complex than the ones taught in the Practical Biostatistics course. As a consequence, the participant will be better able to choose the model that best fits the data at hand and interpret the results.


    This course runs once a year and consists of ten morning sessions on Tuesday and Fridays, spread over a period of five weeks.

    The first two hours of each class explain the theory. Assignments are given, which can be made in a practical immediately after class and which is supervised for one hour. These are mostly exercises in which data have to be analyzed using a statistical program.

    There is no final examination, but active participation is required in order to be able to follow the course. Participants are expected to spend on average three hours around each class on reading of papers, learning of the theory and making of the exercises.

      Scheduled dates:

      • January 14, 17, 21, 24, 28 and 31 and February 4, 7, 11 and 14, 2020 (full until further notice)


      Seven ‘basic’ sessions are devoted to general topics in statistical inference. They will be repeated every year. The specific models that we are explain have a biennial schedule. It is possible to follow only specific topics, in consultation with the course coordinator.

      • Session 1: Main Statistical Concepts:
        Principle of maximum likelihood estimation. Inference: likelihood ratio test, standard error and Wald test, score test; information matrix; multivariate tests; AIC, deviance, bootstrap. All these concepts will be encountered when discussing the regression models in the following sessions.
      • Session 2: Statistical Graphs:
        When and how to use graphs for summarizing data and displaying results . Understanding components of a graph based on the "grammar of graphics". Principles of transparent and informative graph construction.
      • Session 3-4: Regression Models:
        Some frequently used statistical regression models are explained, including their correspondences and the differences between them. Regression strategies. Model fitting. How do we check model assumptions (linearity, homoscedasticity, outliers), and how do we deal with violation of these assumptions.
      • Session 5: Modeling and Understanding Effects from Regression Models:
        Regression models are a very powerful and flexible tool to quantify effects of covariables. We explain how to quantify and interpret interactions, how to model smoothly varying effects via splines an how to deal with "impossible combinations". We also explain how to report effects, both numerically and graphically.
      • Session 6: Why Regression?:
        The pros and cons of hypothesis testing. Exploration, prediction and explanation. Confounding and mediation in causal inference.
      • Session 7: Missing Values:
        Types of missing data. How to deal with missing values: ignore, model, imputation. Missing not at random.
      • Session 8: Introduction to Analysis of Correlated Data:
        Examples of correlated data. Why and when do we need to take correlation into account. Random effects models.
      • Session 9: Longitudinal Data Analysis:
        Some more complicated models for correlated data. Dichotomous outcomes: marginal models (generalized estimating equations) versus random effects models.
      • Session 10: Joint Models:
        A joint model may be used to update prediction of future events using repeated measures of a biomarker or to correct the change-pattern of the repeatedly measured biomarker for non-randomly missing data.

      Curriculum in 2019:

      Sessions 1to7 are similar. The other classes cover prediction, and analysis of survival data.

      Target Audience:

      The course is intended for those who have completed the Practical Biostatistics course, or who have otherwise gained sufficient knowledge of linear, logistic, and Cox regression models. Although problems are approached from a practical perspective, it has a more theoretical content than the Practical Biostatistics course. Therefore, some experience with data analysis is needed in order to be able to relate the statistical principles to one's own practice. Those who have followed the Practical Biostatistics course but have not analyzed data themselves, are advised to postpone participation.

      The techniques and models that are presented during the course are explained and practiced using SPSS (as far as possible) and especially the R statistical program. Some prior knowledge of R is an advantage, but not strictly necessary. Instructions on how to download and install R is distributed in advance of the course.


      The course is given in English, unless all participants understand Dutch.


      To qualify for the certificate, a participant must attend all lectures and workshops, and complete all assignments satisfactorily. Attendance is registered.

      Study load:

      60 hours, which is comparable to 2.1 ECTS points.

      Number of participants:

      Maximum 45 per course.


      No charge for registered AMC PhD candidates. Employees of the AMC or AMC Medical Research BV can participate provided slots are available. All other participants are charged a fee of 1,250 euro.

      Course Coordinator:

      Dr. M.H.P. Hof / Dept. of Clinical Epidemiology / / tel. +31 (0)20 - 566 6880.

      More information:

      From the course coordinator or AMC Graduate School / / room E2-172 / tel. +31 (0)20 - 566 4618.


      AMC Graduate School
      Tel: +31 (0)20 - 566 3108