MAM03 - Advanced data analysis in medicine

Advanced Data-Analysis in Medicine

Catalogue number

4604MM003Y

Credits

6

Language of the course

English

Year

1

Time period(s)

Sem. 1 Sem. 2

Educational institute

Medical Informatics

Lecturer(s)

Is part of

Register

Objectives

At the end of the course the student has insight in or is capable of:

  • Recognizing, using and manipulating the mathematical and statistical fundamentals of modern data-analysis techniques;
  • Applying these techniques critically with R and other computer programs to specific datasets.

More information

More information
prof. dr. A.H. Zwinderman Department of KEBB AMC, J1B-226 a.h.zwinderman@amc.uva.nl

Contents

Many medical informaticians are working with large and complex epidemiological databases, and are trying to reconstruct knowledge from such sources. Almost always this requires statistical data-analyses which will be complicated through multidimensionality, nonlinearity, and confounding factors. In this course some modern statistical techniques will be discussed, studied, and applied. Aim is that students are capable of analyzing complex data with minimal guidance: they will recognize how data were sampled, and which consequences the sampling design has for the data-analysis; they will realize which assumptions are made with specific statistical techniques, and how those can be checked; and students are aware of the instability of estimated statistical models and how such instability can be quantified.

Keywords: maximum likelihood theory, (empirical) Bayes methods, modern regression (additive models); prediction, crossvalidation, bootstrapping, multilevel analysis, R.

Recommended prior knowledge

Knowledge of basic statistical concepts is absolutely necessary: estimation, confidence interval, testing, p-value and (linear, logistic, Cox) regression model. The bachelor course 2.2 "Clinical Epidemiology & Biostatistics" covers all these subjects.

Format

In this course we will discuss and study four different subjects (maximum likelihood theory, modern regression, multilevel analysis and bayesian models): one subject per week. Each subject will be introduced in plenary sessions, and studied further by selfstudy assignments (including data-analysis). At the end of the week the results of the assignments are discussed and elaborated upon in a plenary session.

Time

Class times can be found in the course schedule at http://www.rooster.uva.nl.

Study materials

  • Coursebook with papers and assignments, available on Blackboard;
  • Computer program R.

Min/max participants

The maximum number of participants is 25.

Assessment

There will be four self-assignments that will be scored, and the average of the four scores will be the final grade-point.

Comments