Causal Diagrams, Target Trial Emulation and Causal Inference for Modeling and Medical Decision Making

28 June 2026, 1:30 PM – 4:55 PM | 13:30 – 16:55 (local time)

Faculty

  • Uwe Siebert, MD, MPH, MSc, ScD – Professor of Public Health, Medical Decision Making, UMIT TIROL – University for Health Sciences and Technology, Austria; Adjunct Professor of Epidemiology and Health Policy & Management, Harvard T.H. Chan School of Public Health, United States
  • Lára Hallsson, PhD – Researcher, UMIT TIROL – University for Health Sciences and Technology, Austria

Course Overview

One of the most important tasks of decision makers is to derive causal interpretations from comparative effectiveness and cost-effectiveness studies. Often, an intervention or risk factor is modeled to have a causal effect on model parameters such as probabilities, utilities, or costs. Therefore, both data scientists and decision analysts need tools to determine when effect estimates have a valid causal interpretation, which study designs and methods can identify causal effects rather than statistical associations, and which estimand appropriately answers the research question.

This course provides foundational knowledge in causal thinking and introduces visual, structural, and statistical tools to support causal inference. Participants will learn about causal diagrams (directed acyclic graphs, DAGs), causal inference methods, and how structural elements influence whether a model can be interpreted causally and applied across different settings.

The course also introduces the framework of target trial emulation using counterfactual “clones” and links this approach to estimand frameworks and causal decision-analytic modeling. Through published case examples from areas such as oncology, cardiovascular disease, HIV, nutrition, and obstetrics, participants will learn how causal methods are applied in real-world contexts.

Learning Objectives

Participants will learn how to:

  • Define causal interventions and actions, draw and interpret causal diagrams, and apply rules to distinguish causal from non-causal associations and identify bias
  • Understand target trial emulation and the use of counterfactual “clones,” and develop a target trial protocol
  • Select appropriate study designs and analytical methods (e.g., propensity score methods, g-methods) to estimate causal effects
  • Describe how causal methods are applied to real-world observational data, single-arm trials, and pragmatic randomized trials
  • Derive causal model inputs for decision-analytic models used in benefit–harm assessment and health economic evaluation
  • Distinguish between causal and non-causal elements in decision-analytic model structures

Course Format

The course consists of lectures and applied published case examples, with a strong emphasis on interaction and discussion. Participants will engage in exercises, polls, and Q&A sessions, and are encouraged to bring their own research questions for discussion.

Course materials will include session handouts, exercises with solutions, background readings, and software recommendations.

Participant Requirements

Participants should have a basic understanding of epidemiologic methods, particularly confounding.

Three pre-read papers will be distributed prior to the course to support preparation.