[call-for-participation-and-paper] CT-DEB’26 Shared Task on Codabench for CL4Health'26 workshop

[sohrab....@gmail.com]

Dear colleagues,

Apologies for cross-posting.

In conjunction with the CL4Health’26 Workshop at LREC 2026, we are pleased to announce the CT-DEB’26 Shared Task — an open machine learning challenge on predicting medication dosing errors in clinical trials, hosted on CodaBench. The task is open to all with an open leaderboard  and the participants are invited to submit papers describing their methodologies to the CL4Health’26 Workshop.

----------------------------------------------------------------------

Clinical Background

----------------------------------------------------------------------

Medication-related errors, particularly dosing errors, are a major cause of clinical trial delays, increased costs, and patient harm. The goal of CT-DEB’26 is to advance robust, transparent, and explainable ML methods capable of identifying clinical trial protocols at greater risk of such errors before trials begin.

----------------------------------------------------------------------

Task Overview

----------------------------------------------------------------------

Participants will build predictive models based on the ct-dosing-errors-benchmark dataset, a newly curated collection of  more than 40K clinical trial protocols, annotated by clinical research experts. The dataset includes

- Structured metadata,  

- Categorical and numerical variables,  

- Long free-text protocol descriptions,  

- A binary outcome indicating the occurrence of dosing errors.  

The dataset is hosted at the HuggingFace hub and can be found here. The task is to predict the probability that a given trial will experience dosing-related errors during execution.

----------------------------------------------------------------------

Evaluation

----------------------------------------------------------------------

- Primary metric: ROC-AUC.

- Submission is via CodaBench; code repositories are required for leaderboard eligibility.  

- Leaderboards are blind during active phases.

----------------------------------------------------------------------

Timeline

----------------------------------------------------------------------

Phase 1 — Development & Validation:  

- December 1, 2025: Training + validation features release,  

- January 10, 2026: Phase 1 submission deadline,  

- January 12, 2026: Validation leaderboard revealed.  

Phase 2 — Test Evaluation  

- January 13, 2026: Test features release,  

- January 27, 2026: Phase 2 submission deadline,  

- January 31, 2026: Final leaderboard + code verification requests,  

- February 7, 2026: Deadline for responding to code verification.  

CL4Health Workshop Paper Submission  

Participants are invited to submit short papers to the CL4Health’26 workshop summarizing their methods and insights.

- February 18, 2026: Paper submission deadline  

- March 13, 2026: Notification,  

- March 20, 2026: Camera-ready deadline,  

- May 16, 2026: CL4Health Workshop @ LREC 2026 (Palma de Mallorca).

Shared-task link:

https://www.codabench.org/competitions/11891/

Workshop information:  

https://bionlp.nlm.nih.gov/cl4health2026/

Conference information:  

https://www.elra.info/lrec2026/

----------------------------------------------------------------------

Organizing Committee

----------------------------------------------------------------------

Sohrab Ferdowsi  

Félicien Hêche  

Anthony Yazdani  

Douglas Teodoro  

DS4DH — University of Geneva, Switzerland

We warmly invite the ML and NLP communities to participate in advancing safe clinical trial design and improving patient outcomes.

Best regards,  

The CT-DEB’26 Organizing Team