The estimation of causal effects is often hampered by the presence of incomplete data either due to death or other reasons. Incompleteness raises challenges across all steps of causal inference: estimand definition, identifiability, and estimation. This session aims to showcase methodological innovations to tackle incomplete data across the three key steps. Firstly, the challenges of defining an estimand in the context of truncation by death will be discussed with a dual perspective (pragmatic evaluation in clinical research and mechanistic insights into disease processes). Solutions proposed in recent years include the survivor average causal effect, composite estimands such as survival-incorporated quantiles or the win ratio, and hypothetical estimands inspired by causal mediation analyses. Secondly, recent developments regarding identifiability of causal effects in multivariable missingness settings depicted by so-called “missingness” directed acyclic graphs (DAGs) will be exposed. Finally, innovative estimation techniques will be addressed through causal machine learning methods for missing data.

In the era of precision medicine, the paradigm of the “average population trial” is challenged by the need to adopt individualized therapies where treatment choices are dynamically linked to the course of the disease and tailored to the single patient. Biomarker discovery is critical for precision medicine because it reveals molecular features that predict response or risk with targeted treatments. In this setting, the development of modern trial designs and of computationally intensive statistical methods, e.g. for integrating multi-omics data or drug response profiles, is becoming central for predicting a patient’s prognosis or treatment response, and for accelerated drug discovery and companion diagnostics, for instance in oncology.

The traditional fragmented clinical development of treatments are being optimised by master protocols, which are set up to evaluate multiple treatments and (rare) diseases in the same trial. The session will cover recent statistical developments in early to late phase platform trials, precision medicine designs such as basket and umbrella trials used extensively in oncology, and multi-arm adaptive trials, with both Bayesian and frequentist approaches. This session will be highly valuable for conference participants working on trials.

This session advances methods that make modern AI/ML decision-ready for biomedical research where data are censored, heterogeneous, and small in key subgroups. We showcase approaches that (i) identify causal effects of intervening variables even under unmeasured confounding, extending inference beyond standard assumptions; (ii) transfer information from large external cohorts to under-represented populations via principled weighting that aligns covariates and outcomes while down-weighting mismatched sources; and (iii) estimate individualized causal pathways with survival outcomes using deep generative models. Across these contributions, common pillars are explicit estimands, assumption transparency, and valid uncertainty with theory-backed guarantees, paired with scalable computation for longitudinal and high-dimensional settings. The importance is practical and immediate: clinicians and policymakers increasingly rely on AI trained on imperfect data. By addressing censoring, sample-size limitations, domain shift, and hidden bias, this session delivers tools that convert complex real-world data into credible, transportable causal evidence for precision medicine and public-health decisions.