Serialisation Strategy Matters: How FHIR Data Format Affects LLM Medication Reconciliation

• URL: http://arxiv.org/abs/2604.21076v1
• Date: Wed, 22 Apr 2026 20:39:35 GMT
• Title: Serialisation Strategy Matters: How FHIR Data Format Affects LLM Medication Reconciliation
• Abstract summary: We present the first systematic comparison of four FHIR serialisation strategies across five open-weight models.<n>We find that serialisation strategy has a large, statistically significant effect on performance for models up to 8B parameters.<n>Smaller models plateau at roughly 7-10 concurrent active medications, leaving polypharmacy patients, the patients most at risk from reconciliation errors systematically underserved.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Medication reconciliation at clinical handoffs is a high-stakes, error-prone process. Large language models are increasingly proposed to assist with this task using FHIR-structured patient records, but a fundamental and largely unstudied variable is how the FHIR data is serialised before being passed to the model. We present the first systematic comparison of four FHIR serialisation strategies (Raw JSON, Markdown Table, Clinical Narrative, and Chronological Timeline) across five open-weight models (Phi-3.5-mini, Mistral-7B, BioMistral-7B, Llama-3.1-8B, Llama-3.3-70B) on a controlled benchmark of 200 synthetic patients, totalling 4,000 inference runs. We find that serialisation strategy has a large, statistically significant effect on performance for models up to 8B parameters: Clinical Narrative outperforms Raw JSON by up to 19 F1 points for Mistral-7B (r = 0.617, p < 10^{-10}). This advantage reverses at 70B, where Raw JSON achieves the best mean F1 of 0.9956. In all 20 model and strategy combinations, mean precision exceeds mean recall: omission is the dominant failure mode, with models more often missing an active medication than fabricating one, which changes how clinical safety auditing priorities should be set. Smaller models plateau at roughly 7-10 concurrent active medications, leaving polypharmacy patients, the patients most at risk from reconciliation errors, systematically underserved. BioMistral-7B, a domain-pretrained model without instruction tuning, produces zero usable output in all conditions, showing that domain pretraining alone is not sufficient for structured extraction. These results offer practical, evidence-based format recommendations for clinical LLM deployment: Clinical Narrative for models up to 8B, Raw JSON for 70B and above. The complete pipeline is reproducible on open-source tools running on an AWS g6e.xlarge instance (NVIDIA L40S, 48 GB VRAM).

Related papers

• A Federated and Parameter-Efficient Framework for Large Language Model Training in Medicine [59.78991974851707]
  Large language models (LLMs) have demonstrated strong performance on medical benchmarks, including question answering and diagnosis.<n>Most medical LLMs are trained on data from a single institution, which faces limitations in generalizability and safety in heterogeneous systems.<n>We introduce the model-agnostic and parameter-efficient federated learning framework for adapting LLMs to medical applications.
  arXiv  Detail & Related papers  (2026-01-29T18:48:21Z)
• Cross-Representation Benchmarking in Time-Series Electronic Health Records for Clinical Outcome Prediction [44.23284500920266]
  This benchmark standardises data curation and evaluation across two distinct clinical settings.<n>Experiments reveal that event stream models consistently deliver the strongest performance.<n>We find that feature selection strategies must be adapted to the clinical setting.
  arXiv  Detail & Related papers  (2025-10-10T09:03:47Z)
• A systematic review of trial-matching pipelines using large language models [0.9176056742068814]
  Matching patients to clinical trial options is critical for identifying novel treatments, especially in oncology.<n>Large language models (LLMs) offer a promising solution to this problem.<n>This review synthesizes progress in applying LLMs to clinical trial matching, highlighting promising directions and key limitations.
  arXiv  Detail & Related papers  (2025-09-13T21:21:05Z)
• Predicting Length of Stay in Neurological ICU Patients Using Classical Machine Learning and Neural Network Models: A Benchmark Study on MIMIC-IV [49.1574468325115]
  This study explores multiple ML approaches for predicting LOS in ICU specifically for the patients with neurological diseases based on the MIMIC-IV dataset.<n>The evaluated models include classic ML algorithms (K-Nearest Neighbors, Random Forest, XGBoost and CatBoost) and Neural Networks (LSTM, BERT and Temporal Fusion Transformer)
  arXiv  Detail & Related papers  (2025-05-23T14:06:42Z)
• Benchmarking Open-Source Large Language Models on Healthcare Text Classification Tasks [2.7729041396205014]
  This study evaluates the classification performance of five open-source large language models (LLMs)<n>We report precision, recall, and F1 scores with 95% confidence intervals for all model-task combinations.
  arXiv  Detail & Related papers  (2025-03-19T12:51:52Z)
• Large Language Models versus Classical Machine Learning: Performance in COVID-19 Mortality Prediction Using High-Dimensional Tabular Data [0.7472996057387354]
  This study compared the performance of classical feature-based machine learning models (CMLs) and large language models (LLMs) in predicting COVID-19 mortality.<n>XGBoost and RF demonstrated superior performance among CMLs, achieving F1 scores of 0.87 and 0.83 for internal and external validation, respectively.<n>GPT-4 led the LLM category with an F1 score of 0.43, while fine-tuning Mistral-7b significantly improved its recall from 1% to 79%, yielding a stable F1 score of 0.74 during external validation.
  arXiv  Detail & Related papers  (2024-09-02T14:51:12Z)
• Towards a clinically accessible radiology foundation model: open-access and lightweight, with automated evaluation [113.5002649181103]
  Training open-source small multimodal models (SMMs) to bridge competency gaps for unmet clinical needs in radiology. For training, we assemble a large dataset of over 697 thousand radiology image-text pairs. For evaluation, we propose CheXprompt, a GPT-4-based metric for factuality evaluation, and demonstrate its parity with expert evaluation. The inference of LlaVA-Rad is fast and can be performed on a single V100 GPU in private settings, offering a promising state-of-the-art tool for real-world clinical applications.
  arXiv  Detail & Related papers  (2024-03-12T18:12:02Z)
• The effect of data augmentation and 3D-CNN depth on Alzheimer's Disease detection [51.697248252191265]
  This work summarizes and strictly observes best practices regarding data handling, experimental design, and model evaluation. We focus on Alzheimer's Disease (AD) detection, which serves as a paradigmatic example of challenging problem in healthcare. Within this framework, we train predictive 15 models, considering three different data augmentation strategies and five distinct 3D CNN architectures.
  arXiv  Detail & Related papers  (2023-09-13T10:40:41Z)
• A Meta-GNN approach to personalized seizure detection and classification [53.906130332172324]
  We propose a personalized seizure detection and classification framework that quickly adapts to a specific patient from limited seizure samples. We train a Meta-GNN based classifier that learns a global model from a set of training patients. We show that our method outperforms the baselines by reaching 82.7% on accuracy and 82.08% on F1 score after only 20 iterations on new unseen patients.
  arXiv  Detail & Related papers  (2022-11-01T14:12:58Z)
• SANSformers: Self-Supervised Forecasting in Electronic Health Records with Attention-Free Models [48.07469930813923]
  This work aims to forecast the demand for healthcare services, by predicting the number of patient visits to healthcare facilities. We introduce SANSformer, an attention-free sequential model designed with specific inductive biases to cater for the unique characteristics of EHR data. Our results illuminate the promising potential of tailored attention-free models and self-supervised pretraining in refining healthcare utilization predictions across various patient demographics.
  arXiv  Detail & Related papers  (2021-08-31T08:23:56Z)
• An Interpretable Web-based Glioblastoma Multiforme Prognosis Prediction Tool using Random Forest Model [1.1024591739346292]
  We propose predictive models that estimate GBM patients' health status of one-year after treatments. We used total of 467 GBM patients' clinical profile consists of 13 features and two follow-up dates. Our machine learning models suggest that the top three prognostic factors for GBM patient survival were MGMT gene promoter, the extent of resection, and age.
  arXiv  Detail & Related papers  (2021-08-30T07:56:34Z)
• MIA-Prognosis: A Deep Learning Framework to Predict Therapy Response [58.0291320452122]
  This paper aims at a unified deep learning approach to predict patient prognosis and therapy response. We formalize the prognosis modeling as a multi-modal asynchronous time series classification task. Our predictive model could further stratify low-risk and high-risk patients in terms of long-term survival.
  arXiv  Detail & Related papers  (2020-10-08T15:30:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.