Admission requirements
Basic knowledge of biostatistical procedures. Specifically, knowledge about standard regression techniques in epidemiology: logistic regression and Cox regression.
Description
Predictive analytics is an area of statistics concerned with developing and evaluating tools for predicting a health outcome based on individual characteristics. This course aims to improve understanding of the role of predictive analytics for prevention, diagnosis, and intervention. We discuss how to build, validate, and update prediction tools. We also consider the task of assessing their predictive performance in targeted settings. Key concepts and issues in prediction research will be covered, relating for example to adequate study design, sample size and statisytical overfitting. Case studies are used throughout for illustration. The core material of this course will be introduced through predominantly online lectures, which will be interspersed with quiz questions and exercises. The course also gives a hands-on experience in prediction research through R programming and example datasets.
Course objectives
Upon successful completion of this course, students should be able to:
Understand the roles that diagnostic and prognostic models may play
in risk stratification, and ultimately medical decision-making;Know the critical factors that determine the validity of predictions
provided by diagnostic and prognostic models;Have insight in the pitfalls of model development with standard
statistical techniques;Have both theoretical and practical knowledge on advanced methods
in model development and validation, specifically on regression
modelling;Understand the possibilities of using Electronic Health Records data
for risk stratification;
Timetable
The timetable is published on the LUMC roostersite or can be found via the LUMC scheduling app.
Mode of instruction
Lecture, research, online education, group work
Assessment method
Students are assessed according to the following three obligatory components:
Week 1-2 – Online: (20%) Peer review assessment
Week 3 – On campus: (30%) Group presentation
Week 4 – Final week:
(50%) Take-home assignment
All components together make up the grade for the course. It is compulsory to participate in each of the components in order to receive a grade
Details on the assessment can be found in the assessment plan on Brightspace
A minimum result of 5,5 for the overall assessment is required to pass.
Resit
If the result is less than 5,5 or if the student did not participate in one of the components, the student is given the opportunity to resit the assessment as one assignment that covers all the learning goals of the course.
A final grade of 5,5 minimum is considered sufficient.
Reading list
The reading list can be found on Brigthspace. These are given as presentations and pdf files. There is no need to purchase literature, as the presented material is not commercialized.
Extra in depth material is available in the book ‘Clinical prediction models’ by E.W. Steyerberg, published by Springer, for sale at Amazon, Bol.com, Springer, etc.
Registration
Registration must be done via uSis at the latest 5 days before the start of the course. Registration in uSis gives you automatic access to the course in Brightspace.
Contact
Prof. dr. Ewout Steyerberg e.w.steyerberg@lumc.nl
Dr Bas Penning de Vries b.b.l.penning_de_vries@lumc.nl
Remarks
This course is a combination of online ecucation and on campus Education at Leiden University Campus The Hague.