Unit information: Advanced Epidemiology in 2019/20

Unit name	Advanced Epidemiology
Unit code	BRMSM0015
Credit points	20
Level of study	M/7
Teaching block(s)	Teaching Block 2 (weeks 13 - 24)
Unit director	Dr. Savovic
Open unit status	Not open
Pre-requisites	Introduction to Epidemiology and Statistics
Co-requisites	NONE
School/department	Bristol Medical School
Faculty	Faculty of Health Sciences

Description including Unit Aims

The aims of this unit are to:

• Explain how confounding, selection bias, and information bias can arise within different epidemiological studies, and how they can affect findings
• Use causal diagrams (directed acyclic graphs: DAGs) to summarise assumptions about causal relationships, identify sources of bias, and select which variables to adjust for
• Apply statistical methods to address confounding
• Explain the concept of mediation and how it can be explored using causal inference methods
• Describe the application of causal inference methods to the analysis of randomized trials
• Design epidemiological studies including cohort studies, case-control studies, Mendelian randomization studies and regression discontinuity designs, and apply these designs to studies based on routine data and electronic health records
• Describe statistical methods designed to strengthen causal inferences in epidemiology including propensity scores, inverse probability weights and instrumental variable methods
• Explain time-varying confounding and describe statistical methods to address it
• Use sensitivity analyses and negative control approaches to assess whether results are robust to the presence of different biases
• Explain the concept of triangulation.

Intended Learning Outcomes

By the end of this unit, students should be able to:

1. Draw and interpret causal diagrams (directed acyclic graphs: DAGs) to summarise assumptions about causal relationships and identify sources of bias
2. Design and conduct statistical analyses of epidemiological data in order to estimate causal effects
3. Describe challenges in observational epidemiology that limit its ability to establish causal effects
4. Interpret results of statistical analyses in the light of potential confounding, selection, and information biases, and conduct sensitivity analyses to assess the potential of these biases to change conclusions
5. Explain the causal effects that can be estimated from randomized trials, and use appropriate methods for their estimation
6. Describe time varying confounding and outline statistical methods to adjust for this
7. Critically appraise published epidemiological studies with regards to confounding, selection and information biases
8. Describe key features of the important epidemiological study designs, and explain their strengths, limitations and role in modern epidemiology
9. Explain the concept of triangulation and how it can be applied to interpret multiple epidemiological findings

Teaching Information

There will be 10 teaching weeks, plus reading week and revision week.

Face to face teaching for a total of 50 hours will include lectures and tutorials. Directed and self-directed learning (150 hours) will include activities such as reading and preparation for assessment.

Assessment Information

Formative assessment will support learning by using informal questioning, quizzes and group exercises in lectures and tutorials. These will form assessments for learning and will not contribute to the final unit mark. Feedback will be provided in the form of model answers and through group discussion.

Summative assessments This unit is assessed by a 1-hour closed book exam (70% of total mark; ILOs 1-8) and two pieces of coursework (30%):

• Group exercise to develop and justify a DAG for a study addressing a proposed research question. Students will prepare a brief presentation with their causal diagram and an explanation of the reasoning behind it. Students will be provided with a model answer to the exercise and will be asked to carry out peer-marking in groups (15% of total mark; ILO 1).
• Critical appraisal of a published epidemiological study, in the form of a written assignment, to identify key threats to the validity of the study findings (15% of total mark; ILO 1, 2, 6).

A mark of 50% on each assessment task is required to pass the unit.

Reading and References

Reading and References*

There is no compulsory unit text book.

Recommended reading:

1. Rothman, Greenland, Lash. Modern Epidemiology. Third edition.

Hernan & Robins. Causal inference. 2017. (https://cdn1.sph.harvard.edu/wp-content/uploads/sites/1268/2017/10/hernanrobins_v1.10.33.pdf)