Skip to main content
ISOP ACoP 2025 Annual Conference Main banner
ViewAttendees 4
Symposium

Big Data/AI/Machine Learning

C5C - Bridging Causal Inference and Pharmacometric Modeling

Tuesday, October 21, 2025
1:45 PM - 3:15 PM MDT
Location: Willow Lake 3 - 5

Disclosure(s):
Jonathan French, ScD, FISoP: No relevant disclosure to display
James A. Rogers, PhD: No financial relationships to disclose
Mats O. Karlsson, PhD, FISoP, FACCP: No relevant disclosure to display
Manuela R. Zimmermann, PhD: No relevant disclosure to display
Francois Mercier, PhD: No relevant disclosure to display

Description of session (include background & scientific importance): Estimation of causal dose-exposure-response effects, along with understanding causal effect pathways, lies at the core of model-informed drug development and decision-making. In the past two decades, substantial advances in causal inference have emerged across diverse fields, including epidemiology, statistics, and economics, fueled by a drive to derive causal effects from observational data where traditional randomization is absent.

However, similar complexities can arise within randomized trials, particularly when our interest lies in understanding the impact of post-randomization factors—such as drug exposure or surrogate biomarkers—on downstream responses. While biases resulting from confounded exposure-response relationships have received some attention in the pharmacometrics literature, the broader application of causal inference methodologies in model-informed drug development (MIDD) and decision-making remains limited.

This session will feature three insightful talks that link innovative causal inference methods with pharmacometric modeling. By exploring these connections, we seek to show a future landscape where contemporary causal inference concepts enhance traditional pharmacometric modeling, providing a richer understanding of causal dose-exposure-response relationships.

Learning Objectives:

  • Explore how to incorporate various causal inference methodologies into pharmacometric models, enhancing their capability to inform drug development decisions and clinical practice 2. Learn the principles of causal mediation analyses with longitudinal biomarkers and understand the implications for interpreting surrogate endpoints. 3. Learn how to apply instrumental variable techniques to assess causality and obtain quantitative measures of how different origins of exposure variability impact response. 4. Understand how non-linear mixed effects modeling can be employed to adjust for unmeasured confounders in longitudinal settings and provide unbiased estimates of hypothetical estimands.

Presentations:

  • 1:45 PM - 2:05 PM MDT
    Landmarking Strategies for Causal Mediation Analyses with Longitudinal Biomarkers

    Speaker: James A. Rogers, PhD (he/him/his) – Metrum Research Group

  • 2:05 PM - 2:25 PM MDT
    Instrumental variable analyses of exposure – response relationships

    Speaker: Mats O. Karlsson, PhD, FISoP, FACCP (he/him/his) – Uppsala University

  • 2:25 PM - 2:45 PM MDT
    Non-linear mixed effects modeling as a tool for causal inference

    Speaker: Manuela R. Zimmermann, PhD (she/her/hers) – Novartis

  • 2:45 PM - 3:15 PM MDT
    Panel - Bridging Causal Inference and Pharmacometric Modeling

    Panelist: James A. Rogers, PhD (he/him/his) – Metrum Research Group

    Panelist: Mats O. Karlsson, PhD, FISoP, FACCP (he/him/his) – Uppsala University

    Panelist: Manuela R. Zimmermann, PhD (she/her/hers) – Novartis

    Panelist: Francois Mercier, PhD – Roche-Genentech