Third Workshop on Patient-Oriented Language Processing (CL4Health) @ LREC 2026

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

LREC 2026 Palma, Mallorca (Spain)

SCOPE

CL4Health fills the gap among the different biomedical language processing workshops by providing a general venue for a broad spectrum of patient-oriented language processing research. The third workshop on patient-oriented language processing follows the successful CL4Health workshops (co-located with LREC-COLING 2024 and NAACL 2025), which clearly demonstrated the need for a computational linguistics venue focused on language related to public health.

CL4Health is concerned with the resources, computational approaches, and behavioral and socio-economic aspects of the public interactions with digital resources in search of health-related information that satisfies their information needs and guides their actions. The workshop invites papers concerning all areas of language processing focused on patients' health and health-related issues concerning the public. The issues include, but are not limited to, accessibility and trustworthiness of health information provided to the public; explainable and evidence-supported answers to consumer-health questions; accurate summarization of patients' health records at their health literacy level; understanding patients' non-informational needs through their language, and accurate and accessible interpretations of biomedical research. The topics of interest for the workshop include, but are not limited to the following:

* Health-related information needs and online behaviors of the public; * Quality assurance and ethics considerations in language technologies and approaches applied to text and other modalities for public consumption; * Summarization of data from electronic health records for patients; * Detection of misinformation in consumer health-related resources and mitigation of potential harms; * Consumer health question answering (Community Question Answering)(CQA); * Biomedical text simplification/adaptation; * Dialogue systems to support patients' interactions with clinicians, healthcare systems, and online resources; * Linguistic resources, data, and tools for language technologies focusing on consumer health; * Infrastructures and pre-trained language models for consumer health;

IMPORTANT DATES (Tentative) February 18, 2026 -Workshop Paper Due Date️ March 13, 2026 - Notification of acceptance March 20, 2026 - Camera-ready papers due April 10, 2026 - Pre-recorded video due (hard deadline) May 16, 2026 - Workshop SHARED TASKS

Detecting Dosing Errors from Clinical Trials. Medication errors constitute a significant threat to public health worldwide. Although various types of errors may occur, dosing errors have been identified as one of the most frequent types. The objective of the shared task is to develop and evaluate machine learning methods capable of analyzing clinical trial data (including structured metadata and free-text protocol descriptions) to identify trials that are likely to experience unusually high rates of dosing errors. Such predictive tools could serve as early-warning systems, supporting more reliable trial design and enhancing medication safety. A human-annotated dataset comprising 40,000 clinical trials will be used for the training and validation set. Submissions will be evaluated primarily using the F1-score, with AUROC and AUPRC reported as complementary metrics. To avoid participants using unauthorized data for training, only submissions of fully reproducible, open methods will be considered.

Automatic Case Report Form (CRF) Filling from Clinical Notes. Case Report Forms are standardized instruments in medical research used to collect patient data consistently and reliably. They consist of predefined items to be filled with patient information. Automating CRF filling from clinical notes would accelerate clinical research, reduce manual burden on healthcare professionals, and create structured representations that can be directly leveraged to produce accessible, patient-friendly, and practitioner-friendly summaries. The shared task focuses on developing systems that take clinical narratives as input and automatically populate the relevant slots in a CRF. Two different (synthetic and real clinical data) multilingual datasets covering English and Italian will be shared with the participants to develop the system. The evaluation will be performed in terms of F1-score by comparing the system's outputs with ground truth labels.

Grounded Question Answering from Electronic Health Records. While there have been studies on answering general health-related queries, few have focused on their own medical records. Furthermore, grounding (linking responses to specific evidence) is critical in medicine. Yet, despite extensive studies in open domains, its application in the clinical domain remains underexplored. To foster research in these sparsely studied areas of clinical natural language processing, the ArchEHR-QA (“Archer”) shared task was introduced as part of the BioNLP Workshop at ACL 2025. Given a patient-posed natural language question, the corresponding clinician-interpreted question, and the patient's clinical note excerpt, the task is to produce a natural language answer with citations to the specific note sentences. The ArchEHR-QA dataset is based on real-life patients' questions from public health forums aligned with clinical notes from publicly accessible EHR databases (MIMIC-III/IV) to form a cohesive question-answer source case. Submissions will be evaluated for evidence use (“Factuality”) and answer quality (“Relevance”). Factuality is measured via Precision, Recall, and F1 Scores between the cited evidence sentences in systems' answers and ground truth labels. Relevance is measured against ground truth answers using BLEU, ROUGE, SARI, BERTScore, AlignScore, and MEDCON.

SUBMISSIONS

Two types of submissions are invited:

- Full papers: should not exceed eight (8) pages of text, plus unlimited references. These are intended to be reports of original research. - Short papers: may consist of up to four (4) pages of content, plus unlimited references. Appropriate short paper topics include preliminary results, application notes, descriptions of work in progress, etc.

Electronic Submission: Submissions must be electronic and in PDF format, using the Softconf START conference management system. Submissions need to be anonymous. Papers should follow LREC 2026 formatting. LREC provides style files for LaTeX and Microsoft Word at https://lrec2026.info/authors-kit/.

Submission site: https://softconf.com/lrec2026/CL4Health/ Dual submission policy: papers may NOT be submitted to the workshop if they are or will be concurrently submitted to another meeting or publication.

Share your LRs: When submitting a paper from the START page, authors will be asked to provide essential information about resources (in a broad sense, i.e. also technologies, standards, evaluation kits, etc.) that have been used for the work described in the paper or are a new result of your research. Moreover, ELRA encourages all LREC authors to share the described LRs (data, tools, services, etc.) to enable their reuse and replicability of experiments (including evaluation ones).

MEETING

The workshop will be hybrid. Virtual attendees must be registered for the workshop to access the online environment.

Accepted papers will be presented as posters or oral presentations based on the reviewers’ recommendations.

ORGANIZERS - Deepak Gupta, US National Library of Medicine - Paul Thompson, National Centre for Text Mining and University of Manchester, UK - Dina Demner-Fushman, US National Library of Medicine - Sophia Ananiadou, National Centre for Text Mining and University of Manchester, UK --

Paul Thompson Research Fellow Department of Computer Science National Centre for Text Mining Manchester Institute of Biotechnology University of Manchester 131 Princess Street Manchester M1 7DN UK http://personalpages.manchester.ac.uk/staff/Paul.Thompson/