Institution: Universität Hamburg, Hamburg Center for Health Economics
Course Instructors: Prof. Dr. Marco Caliendo / Prof. Dr. Tom Stargardt
Course Value: 5 ECTS
24.09.2018: 9:00 am ‐ 12:30 pm / 01:30 pm ‐ 05:00 pm
25.09.2018: 9:00 am ‐ 12:30 pm / 01:30 pm ‐ 05:00 pm
26.09.2018: 9:00 am ‐ 12:30 pm / 01:30 pm ‐ 05:00 pm
Place: Universität Hamburg
Classroom: 4029, Esplanade 36
Language of instruction: English
Registration: Please contact Elena Phillips, firstname.lastname@example.org (first come, first-served)
The course will cover methods for drawing causal inference in interventional, non‐
experimental/non‐randomized studies on quality of care with administrative data. In order to control for confounders between intervention and control group, at first simple methods (such as stratification and standardization) as well as advanced methods (Propensity Score Matching, Difference‐in‐Differences, Regression‐Discontinuity Designs) are taught. The course will also give an overview on common risk‐adjustment instruments (generic and disease specific risk‐adjustment scores based on diagnoses or ATC codes) for use with health outcomes.
The course will be split in theoretical and practical sessions. During the practical sessions we are going to implement the discussed estimators with STATA. Hence, a basic knowledge of STATA (data handling, running do‐files, etc.) is a prerequisite for the course. If you are not familiar with STATA you might want to check the online introduction from the UCLA Institute for Digital Research and Education https://stats.idre.ucla.edu/stata/. The relevant estimation commands and ado‐files will be explained during the course; some of them require STATA 13 or higher.
Assessment: Students will have to complete an assignment doing (statistical) analyses of a dataset. Results have to be presented in the form of a short summary paper.