TUESDAY, 9 JUNE 2026 Day 2 – HealTAC day 1Arundel 1
09:30–10:00
Registration
10:00–10:05
Welcome from the Chairs
10:05–10:15
Opening address – Prof. Goran Nenadic (University of Manchester and Healtex, the UK Health Text Analytics Network)
10:15–11:15
Keynote – Dr Martin Krallinger (Barcelona Supercomputing Center) Speaker profile and abstract Chair: Dr Yamiko Msosa, King's College London
11:15–11:30
Break
11:30–12:00
Oral session 1, group 1 – Mental Health & Social Determinants of Health
Chair: Dr Fahrurrozi Rahman, University of Edinburgh
Paul Howarth. Learning from Tragedy: Structuring Complex Narrative Evidence in Health Systems with Ontology-Guided Hybrid NLP
Matúš Falis, Alice Eaton, Michael Holder, Kieran Sweeney, Matthew Iveson, Samuel McInerney, Franz Gruber, Emily Ball, Heather Whalley and Arlene Casey. Extraction of Antidepressant Response from Primary-Care FreeText Data with Large Language Models (lightning talk)
Sharon Sondh, Christoph Mueller, Lifang Li, Angus Roberts, Harsharon Sondh and Robert Stewart. Application of a BERT NLP model for recorded violence to investigate its associations with emergency department attendance and mental health service use in older adults (lightning talk)
David Chandran, Alice Broadbent, Jyoti Sanyal and Robert Stewart. Utilization of a fine-tuned BERT model to identify instances of the subject of clinical records experiencing job-loss (lightning talk)
12:00–12:30
Oral session 1, group 2 – Evaluation & Reliability
Chair: Dr Fahrurrozi Rahman, University of Edinburgh
Zicheng Li, Honghan Wu and Liana Romaniuk. A Human-Centred Evaluation Framework for Patient-Facing Mental-Health LLMs Using Clinically Grounded Synthetic Dialogues
Liam Barrett, Nikhil Joshi and Nishchay Mehta. LLM reliability in clinical information extraction from ENT electronic health records (lightning talk)
Arooj Hussain, Warren Del-Pinto, Meghna Jani, William G Dixon and Goran Nenadic. Assessing Certainty of Diagnoses in Clinical Text (lightning talk)
Mingyang Li, Viktor Schlegel, Tingting Mu, Wuraola Oyewusi and Goran Nenadic. Evaluation and LLM-Guided Learning of ICD Coding Rationales (lightning talk)
12:30–12:45
Open Community Forum
12:45–13:45
Lunch
13:45–14:45
Industry panel
Moderator:Patrick Schrempf (Canon Medical Research Europe Ltd & University of St Andrews)
Panellists:Natalia Viani (Staff ML Engineer, Flatiron Health), Jack Richmond (Senior AI Scientist, Akrivia Health), Paul Howarth (Director of Research & Development / Founder, Akumen Ltd), Robert Dürichen (Head of ML Analytics, Arcturis Data), Ed Broadhead (CEO, Health Narrator)
14:45–15:15
Oral session 2, group 3 – Engineering, Resources & Governance
Chair: Dr Jaya Chaturvedi, King's College London
Wilson Lukmanjaya, Sarah Cox, Tony Butler, Oscar Perez-Concha, Leah Bromfield and George Karystianis. The Secret is in the Relationships: De-identifying Child Intake Reports
Adam Sutton, Niko Möller-Grell, Thomas Searle and Richard Dobson. Fast, Accurate, and Local Conversion of MIMIC-IV to OMOP with DBT (lightning talk)
Beatrice Alex, Claire Grover, Arlene Casey, Richard Tobin, Heather Whaley and William Whiteley. GS-BrainText: A Multi-Site Brain Imaging Report Dataset for Clinical NLP Development and Validation (lightning talk)
Chao Xu, Xiaolei Diao, Alec Diallo, Luo Mai, Yunsoo Kim and Honghan Wu. PAIR-EHR: Transforming Clinical Case Reports into Structured EHR Representations (lightning talk)
15:15–15:45
Oral session 2, group 4 – Narrative Analysis & Information Extraction
Chair: Dr Jaya Chaturvedi, King's College London
Tao Wang, Zhuoyuan Tang, Kai Yang, Li Yuan and Angus Roberts. Depression Severity Estimation via Speaker Diarization and Multi-Task Learning with Multimodal Cross-Attention
Oscar Windrath-Carr and Maite Arribas. A Scalable Approach to Address the Lack of Labelled Clinical Free Text Data: Case Study for Venous Thromboembolism (lightning talk)
Fahrurrozi Rahman, Aryo Pradipta Gema, Arlene Casey, Honghan Wu, Bruce Guthrie and Beatrice Alex. Adverse Events and Geriatric Syndromes in MIMIC-IV: A Multilabel Document Classification Study (lightning talk)
Daisy Monika Lal, Paul Rayson and Andrew Moore. Do We Need Complex Models? Using Collocations for Metaphor Detection in Cancer Narratives (lightning talk)
15:45–15:50
Demo introductions
15:50–16:00
Break
16:00–17:00
Posters and demos session 1
Neil Smalheiser. Tools for User-focused Mining of the Biomedical Literature
Georgina Cosma, G. Thomas Jun, Mikel Williams-Lekuona and Saul Albert. Multimorbidity Patterns in Adults with Intellectual Disability and Digital Tools for Exploring Care Experiences
Matúš Falis, Alice Eaton, Michael Holder, Kieran Sweeney, Matthew Iveson, Samuel McInerney, Franz Gruber, Emily Ball, Heather Whalley and Arlene Casey. Extraction of Antidepressant Response from Primary-Care FreeText Data with Large Language Models
Sharon Sondh, Christoph Mueller, Lifang Li, Angus Roberts, Harsharon Sondh and Robert Stewart. Application of a BERT NLP model for recorded violence to investigate its associations with emergency department attendance and mental health service use in older adults
David Chandran, Alice Broadbent, Jyoti Sanyal and Robert Stewart. Utilization of a fine-tuned BERT model to identify instances of the subject of clinical records experiencing job-loss
Liam Barrett, Nikhil Joshi and Nishchay Mehta. LLM reliability in clinical information extraction from ENT electronic health records
Arooj Hussain, Warren Del-Pinto, Meghna Jani, William G Dixon and Goran Nenadic. Assessing Certainty of Diagnoses in Clinical Text
Mingyang Li, Viktor Schlegel, Tingting Mu, Wuraola Oyewusi and Goran Nenadic. Evaluation and LLM-Guided Learning of ICD Coding Rationales
Adam Sutton, Niko Möller-Grell, Thomas Searle and Richard Dobson. Fast, Accurate, and Local Conversion of MIMIC-IV to OMOP with DBT
Beatrice Alex, Claire Grover, Arlene Casey, Richard Tobin, Heather Whaley and William Whiteley. GS-BrainText: A Multi-Site Brain Imaging Report Dataset for Clinical NLP Development and Validation
Chao Xu, Xiaolei Diao, Alec Diallo, Luo Mai, Yunsoo Kim and Honghan Wu. PAIR-EHR: Transforming Clinical Case Reports into Structured EHR Representations
Oscar Windrath-Carr and Maite Arribas. A Scalable Approach to Address the Lack of Labelled Clinical Free Text Data: Case Study for Venous Thromboembolism
Fahrurrozi Rahman, Aryo Pradipta Gema, Arlene Casey, Honghan Wu, Bruce Guthrie and Beatrice Alex. Adverse Events and Geriatric Syndromes in MIMIC-IV: A Multilabel Document Classification Study
Daisy Monika Lal, Paul Rayson and Andrew Moore. Do We Need Complex Models? Using Collocations for Metaphor Detection in Cancer Narratives
Omotayo Faluyi. To what extent can existing toxicity- and sentiment-oriented language models reliably distinguish between psychologically harmful and constructive negative comments?
Daisy Monika Lal, Paul Rayson and Andrew Moore. Tracing Annotation Bias in Patient Narratives through LLM-Based Role-Conditioned Emotion Detection
Ebrahim Alharbi and Mark Stevenson. Predicting Systematic Review Conclusion Change
Keiran Tait, Joseph Cronin and Robert Durichen. Clinical data enrichment using LLM ensemble approaches
Joseph Cronin, Keiran Tait and Robert Durichen. Agentic system for research specific real world data quality checks
Yunsoo Kim, Xiaolei Diao, Chao Xu, Sandosh Padmanabhan and Honghan Wu. PAIR-SUM: Summarisation of MIMIC hypertension discharge notes
Frida Hæstrup, Jakob Grøhn Damgaard, Sara Kolding, Erik Perfalk, Andreas Aalkjær Danielsen and Søren Dinesen Østergaard. Metric-Dependent Optimisation of Clinical Prediction Models in Psychiatry
Jakob Grøhn Damgaard, Kenneth Enevoldsen, Sara Kolding, Frida Hæstrup, Erik Perfalk, Andreas A. Danielsen and Søren D. Østergaard. Predicting Clinical Outcomes for Patients with Mental Illness using NLP on Electronic Health Records
Sara Kolding, Jakob Grøhn Damgaard, Frida Hæstrup, Erik Perfalk, Rebekah Baglini, Andreas Aalkjær Danielsen and Søren Dinesen Østergaard. Detection of Bias in Prediction Models for Clinical Psychiatry based on Data from Electronic Health Records
Erik Perfalk, Jakob Grøhn Damgaard, Kenneth Enevoldsen, Andreas Aalkjaer Danielsen and Søren Dinesen Østergaard. Estimating severity of psychiatric symptoms via natural language processing of electronic health record data
TUESDAY, 9 JUNE 2026 19:00 – Conference DinnerArundel 1
WEDNESDAY, 10 JUNE 2026 Day 3 – HealTAC day 2Arundel 1
09:00–09:15
Welcome to day 2
09:15–10:15
PhD and fellowship forum
Chair: Dr Matúš Falis, University of Edinburgh
Faezeh Safari, Hang Dong, Zeyu Fu and Aline Villavicencio. A Neuro-Symbolic Approach to Graph-Verified and Interpretable Chest X-Ray Report Generation
Linglong Qian and Zina Ibrahim. TIMELY-Agent: An Agentic Framework for Multimodal Clinical Reasoning Benchmark Construction
Mel Ramasawmy. Addressing inequalities in patient-facing AI symptoms checkers: Planned fellowship application
10:15–10:45
PhD and fellowship lightning talks
Chair: Dr Tao Wang, King's College London
Agathe Zecevic, Angus Roberts and Sebastian Zeki. Exploring limitations of guideline-grounded Clinical Decision Support Systems by comparison with clinical practice
Adam Williams, Pj Noble and Kirsten McMillan. Comparison of Transformer Encoder-Based Text Classifiers in Identifying Canine Involvement in Road Traffic Accidents
Chloe Hutton, Chris Fox, Carol Brayne, Hang Dong and David Llewellyn. AI-enhanced dementia prevention: precision risk reduction using large language models
Aditya Borakati, Jack Wu, Christopher Callaghan, Miriam Cortes Cerisuelo, James Teo and David Wallace. Developing a Natural Language Processing Enhanced Liver Transplant Registry
Shubham Agarwal, Jaya Chaturvedi, Julia Ive, Robert Stewart, Thomas Searle and Richard Dobson. Investigating Bias in Mental Health Clinical Notes
10:45–11:00
Break
11:00–12:00
Panel – Human-Centered AI & NLP: Bridging Research and Real-World Practice
Moderator:Prof. Rob Stewart (King's College London and South London and Maudsley NHSFT)
Panellists:Matthew Broadbent (Clinical Informatics Lead, South London and Maudsley NHSFT), Rudolf Cardinal (Consultant Psychiatrist, Cambridgeshire and Peterborough NHSFT and Professor of Psychiatry and Informatics, University of Cambridge), Ian Roberts (Research Software Engineer, University of Sheffield), Arlene Casey (Principal NLP Data Scientist, DataLoch), Mohammad Al-Agil (Senior Developer Analyst, King's College Hospital NHSFT)
12:00–13:00
Keynote – Prof. Aline Villavicencio (University of Exeter & University of Sheffield) Speaker profile and abstract Chair: Dr Sneha Jha, Queen's University Belfast
13:00–13:05
Demo introductions
13:05–14:00
Lunch
14:00–15:00
Posters and demos session 2
Samuel Thio, David Tang, James Teo, Thomas Searle and Richard Dobson. Cogstack Coder: Agentic Medical Coding Assistant for EHR Systems
Faezeh Safari, Hang Dong, Zeyu Fu and Aline Villavicencio. A Neuro-Symbolic Approach to Graph-Verified and Interpretable Chest X-Ray Report Generation
Linglong Qian and Zina Ibrahim. TIMELY-Agent: An Agentic Framework for Multimodal Clinical Reasoning Benchmark Construction
Mel Ramasawmy. Addressing inequalities in patient-facing AI symptoms checkers: Planned fellowship application
Agathe Zecevic, Angus Roberts and Sebastian Zeki. Exploring limitations of guideline-grounded Clinical Decision Support Systems by comparison with clinical practice
Adam Williams, Pj Noble and Kirsten McMillan. Comparison of Transformer Encoder-Based Text Classifiers in Identifying Canine Involvement in Road Traffic Accidents
Chloe Hutton, Chris Fox, Carol Brayne, Hang Dong and David Llewellyn. AI-enhanced dementia prevention: precision risk reduction using large language models
Aditya Borakati, Jack Wu, Christopher Callaghan, Miriam Cortes Cerisuelo, James Teo and David Wallace. Developing a Natural Language Processing Enhanced Liver Transplant Registry
Shubham Agarwal, Jaya Chaturvedi, Julia Ive, Robert Stewart, Thomas Searle and Richard Dobson. Investigating Bias in Mental Health Clinical Notes
Yusuf Yildiz, Goran Nenadic, Meghna Jani and David A. Jenkins. Infusing Medical Hierarchies into Transformers: A Study of Ontology Infusion Methods in Clinical Transformers
Haoyu Wang, Zitong Li, Linglong Qian and Zina Ibrahim. TIMELY-Bench: Quantifying Temporal Leakage in Multimodal ICU Prediction
Elizabeth Ford, Rob Dickinson and Edmund Broadhead. Challenges in Generating Realistic Synthetic Clinical Records Using Large Language Models for Evaluating AI Summarizing Tools
Franz Gruber, Stuart Dunbar, Matúš Falis, Samuel McInerney, Stephen Powell and Arlene Casey. De-identifying patient records using a combination of regular expressions and large language models
Shrankhla Pandey, Sarah Morgan, Ben Laws, Stefan Rennick-Egglestone, Mike Slade and Graham Murray. Annotation data collection study for recovery narratives
Joe Davies, Beata Fonferko-Shadrach, Ben Holgate, Arron Lacey, Huw Strafford, Owen Pickrell and Mark P Richardson. Creating A Gold-Standard Annotated Epilepsy EHR Dataset
Thomas Rowlands, Yamiko Msosa, Claire Newman, Phil Quinlan, Angus Roberts, Robert Stewart, Simon Thompson, Grazziela Figueredo and Tim Beck. TRExt: Demonstrating text analytics capabilities for Trusted Research Environments
Matúš Falis, Matthew Iveson, Samuel McInerney, Franz Gruber, Emily Ball, Heather Whalley and Arlene Casey. Systematic Review of Natural Language Processing for Extracting Psychiatric Medication Response from Clinical Free Text (Antidepressants, Antipsychotics, and Mood Stabilisers)
Leo Zhang, Agathe Zecevic, Sebastian Zeki and Angus Roberts. A Comparison Study of Three Pipelines for Barrett's Oesophagus Surveillance Prediction
Rebecca Goulding, Gail Davidge, Sarah Markham, Warren Del-Pinto, Brian McMillan and Goran Nenadic. Synthetic free-text healthcare data: informing research design through public involvement
Yamiko Joseph Msosa, Angus Roberts and Richard J. Dobson. Supporting Extraction of Biopsychosocial Factors from Routine Electronic Health Records with Natural Language Processing
Zitong Li, Haoyu Wang and Linglong Qian. Large Language Models for Sparse Clinical Time Series: A MIMIC-IV Benchmark of Imputation, Calibration, and Explanation Reliability
15:00–15:45
Panel – Free-text healthcare data and Trusted Research Environments
