Causal Inference

Study Course Description

Course Description Statuss:Approved

Course Description Version:5.00

Study Course Accepted:14.03.2024 11:47:32

Study Course Information
Course Code:		SL_114		LQF level:		Level 7
Credit Points:		2.00		ECTS:		3.00
Branch of Science:		Mathematics; Theory of Probability and Mathematical Statistics		Target Audience:		Life Science
Study Course Supervisor
Course Supervisor:		Māris Munkevics
Study Course Implementer
Structural Unit:		Statistics Unit
The Head of Structural Unit:
Contacts:		14 Baložu street, 2nd floor, Riga, statistikarsu[pnkts]lv, +371 67060897
Study Course Planning
Full-Time - Semester No.1
Lectures (count)		6	Lecture Length (academic hours)		2	Total Contact Hours of Lectures		12
Classes (count)		6	Class Length (academic hours)		2	Total Contact Hours of Classes		12
Total Contact Hours								24
Part-Time - Semester No.1
Lectures (count)		6	Lecture Length (academic hours)		1	Total Contact Hours of Lectures		6
Classes (count)		6	Class Length (academic hours)		2	Total Contact Hours of Classes		12
Total Contact Hours								18
Study course description
Preliminary Knowledge:
• Familiarity with probability theory and mathematical statistics. • Basic knowledge in R software. • Basic knowledge of linear models and statistical estimation techniques.
Objective:
The objective of this course is to give students the understanding of the distinction between statistical models and causal models and knowledge of the methodology to assess identifiability of causal effects for a particular study, as well as skills to estimate causal parameters using some specific analysis tools. The software package R will be used for computer practical classes, where mainly simulation methods are used to explore the validity of alternative methods. Also, several specialized R packages for causal inference will be introduced.
Topic Layout (Full-Time)
No.	Topic			Type of Implementation		Number	Venue
1	Understanding and defining causal effects. Study designs that enable causal conclusions. Biases due to confounding and selection bias in observational studies.			Lectures		1.00	auditorium
2	Data simulation in R to test, whether the true causal effects can be identified by classical modelling tools. Issues of confounding and model selection in R.			Classes		1.00	computer room
3	Causal graphs and graphical tools to assess confounding and identifiability of causal effects.			Lectures		1.00	auditorium
4	Causal graphs in R.			Classes		1.00	computer room
5	Propensity score matching and Inverse Probability Weighting and its use for the analysis of epidemiological data.			Lectures		1.00	auditorium
6	Propensity score matching and Inverse probability weighted estimators in R.			Classes		1.00	computer room
7	Causal mediation analysis. The concept of direct and indirect effects.			Lectures		1.00	auditorium
8	Causal mediation analysis in R.			Classes		1.00	computer room
9	The Instrumental Variables (IV) Estimator and its application in clinical trials (analysis of the effect of non-adherence).			Lectures		1.00	auditorium
10	IV analysis in R.			Classes		1.00	computer room
11	Mendelian Randomization – using genes as instruments in epidemiology.			Lectures		1.00	auditorium
12	Mendelian Randomization in R, using summary statistics.			Classes		1.00	computer room
Topic Layout (Part-Time)
No.	Topic			Type of Implementation		Number	Venue
1	Understanding and defining causal effects. Study designs that enable causal conclusions. Biases due to confounding and selection bias in observational studies.			Lectures		1.00	auditorium
2	Data simulation in R to test, whether the true causal effects can be identified by classical modelling tools. Issues of confounding and model selection in R.			Classes		1.00	computer room
3	Causal graphs and graphical tools to assess confounding and identifiability of causal effects.			Lectures		1.00	auditorium
4	Causal graphs in R.			Classes		1.00	computer room
5	Propensity score matching and Inverse Probability Weighting and its use for the analysis of epidemiological data.			Lectures		1.00	auditorium
6	Propensity score matching and Inverse probability weighted estimators in R.			Classes		1.00	computer room
7	Causal mediation analysis. The concept of direct and indirect effects.			Lectures		1.00	auditorium
8	Causal mediation analysis in R.			Classes		1.00	computer room
9	The Instrumental Variables (IV) Estimator and its application in clinical trials (analysis of the effect of non-adherence).			Lectures		1.00	auditorium
10	IV analysis in R.			Classes		1.00	computer room
11	Mendelian Randomization – using genes as instruments in epidemiology.			Lectures		1.00	auditorium
12	Mendelian Randomization in R, using summary statistics.			Classes		1.00	computer room
Assessment
Unaided Work:
1. Individual work with the course material and compulsory literature in preparation to 6 lectures according to plan. 2. Project work – critical assessment of a published paper on causal analysis of biomedical data (mediation analysis, analysis of nonadherence, Mendelian Randomization). Presentation of the project work’s results. In order to evaluate the quality of the study course as a whole, the student must fill out the study course evaluation questionnaire on the Student Portal.
Assessment Criteria:
Assessment on the 10-point scale according to the RSU Educational Order: • Project work and it’s presentation – 50%. • Final written exam – 50%.
Final Examination (Full-Time):		Exam (Written)
Final Examination (Part-Time):		Exam (Written)
Learning Outcomes
Knowledge:		The students will: • compare the distinction between association models and causal models; the problem of confounding and the idea of adjustment and/or standardization to control for confounding. • state terminology and properties of Directed Acyclic Graphs to describe and assess causal association structures in data. • list special methods for estimation of causal effects: propensity score matching, inverse probability weighting, instrumental variables estimation. • explain the essence of the problem of causal mediation, differention between direct and indirect effects.
Skills:		The students who have completed the course, will be able to: • decide, whether a study would lead to estimates with immediate causal interpretation. • sketch a causal graph (a DAG) to understand and discuss identifiability of causal effects of interest. • select an appropriate set of covariates for adjustment in regression analysis. • independently use specialized tools (and corresponding R packages) for causal inference: propensity score matching, inverse probability weighting, instrumental variables estimation. • communicate and present the findings in writing and oral of causal interpretation of the results of data analysis.
Competencies:		• The students will be competent in understanding and critical assessment of the published research that uses causal statements and/or causal inference methods for data analysis. • The students will be competent in causal reasoning based on a study design and available data in an interdisciplinary research team. • In particular, a student who has successfully passed the course, is able to assess (and explain), which of the following is valid in the particular study: a) the causal effect of interest is estimable by standard modelling tools (with adjustment for confounders); b) the causal effect of interest is estimable with specific methodology for causal inference; c) the causal effect of interest cannot be identified; In cases a) and b) the student will be competent to conduct the analysis, and disseminate new knowledge in health-related studies.
Bibliography
No.	Reference
Required Reading
1	Hernan, M. and Robins, J. Causal Inference. What if. Boca Raton: Chapman & Hall/CRC, 2020.
Additional Reading
1	Pearl, J. Causality: Models, Reasoning and Inference. Cambridge university Press, 2009.
2	Pearl, J. and Mackenzie, D. The Book of Why. Penguin Books, 2019.