Prognosis is a key concept in patient care. The methodology of prognostic research is however relatively underdeveloped. This is in contrast to its growing importance in clinical medicine. In the course, principles and methods of non-experimental prognostic research will be discussed. In lectures, practical exercises and discussion of examples, the practice of prognostic research in a clinical setting is addressed. Emphasis will be on design and statistical analysis of prognostic studies, construction and estimation of prediction rules and approaches to validation and generalization of research results. Problems with small datasets will be extensively discussed.
By the end of the course, you should be able to:
- Understand the key characteristics and different types of prognostic research
- Set out the various steps involved in performing prognostic research
In particular, you should be able to:
- Demonstrate an insight into different types of missing values
- Understand different ways of handling missing values in prognostic research
- Propose different modelling approaches for prognostic research, including non-linear models
- Develop a prognostic model
- Show how to derive a prognostic score, and choose adequate score cut-offs
- Know how to apply modelling techniques to deal with over-fitting in small data sets.