The core novelty of our approach lies in the addition of entropy regularisation to self-critical sequence training. This helps maintain a higher entropy in the token distribution, preventing overfitting to common phrases and ensuring a broader exploration of the vocabulary during training, which is essential for handling the diversity of the radiology reports in the RRG24 datasets. We apply this to a multimodal language model with RadGraph as the reward. Additionally, our model incorporates several other aspects. We use token type embeddings to differentiate between findings and impression section tokens, as well as image embeddings. To handle missing sections, we employ special tokens. We also utilise an attention mask with non-causal masking for the image embeddings and a causal mask for the report token embeddings.

Clinical documentation is an important aspect of clinicians’ daily work and often demands a significant amount of time. The BioNLP 2024 Shared Task on Streamlining Discharge Documentation (Discharge Me!) aims to alleviate this documentation burden by automatically generating discharge summary sections, including brief hospital course and discharge instruction, which are often time-consuming to synthesize and write manually. We approach the generation task by fine-tuning multiple open-sourced language models (LMs), including both decoder-only and encoder-decoder LMs, with various configurations on input context. We also examine different setups for decoding algorithms, model ensembling or merging, and model specialization. Our results show that conditioning on the content of discharge summary prior to the target sections is effective for the generation task. Furthermore, we find that smaller encoder-decoder LMs can work as well or even slightly better than larger decoder-based LMs fine-tuned through LoRA. The model checkpoints from our team (aehrc) are openly available.

In the field of information retrieval, Query Likelihood Models (QLMs) rank documents based on the probability of generating the query given the content of a document. Recently, advanced large language models (LLMs) have emerged as effective QLMs, showcasing promising ranking capabilities. This paper focuses on investigating the genuine zero-shot ranking effectiveness of recent LLMs, which are solely pre-trained on unstructured text data without supervised instruction fine-tuning. Our findings reveal the robust zero-shot ranking ability of such LLMs, highlighting that additional instruction fine-tuning may hinder effectiveness unless a question generation task is present in the fine-tuning dataset. Furthermore, we introduce a novel state-of-the-art ranking system that integrates LLM-based QLMs with a hybrid zero-shot retriever, demonstrating exceptional effectiveness in both zero-shot and few-shot scenarios. We make our codebase publicly available at https://github.com/ielab/llm-qlm.

We describe the participation of team e-Health CSIRO in the BioNLP RadSum task of 2023. This task aims to develop automatic summarisation methods for radiology. The subtask that we participated in was multimodal; the impression section of a report was to be summarised from a given findings section and set of Chest X-rays (CXRs) of a subject’s study. For our method, we adapted an encoder-to-decoder model for CXR report generation to the subtask. e-Health CSIRO placed seventh amongst the participating teams with a RadGraph ER F1 score of 23.9.

We investigated the development of a Machine Learning (ML)-based classifier to identify abnormalities in radiology reports from Emergency Departments (EDs) that can help automate the radiology report reconciliation process. Often, radiology reports become available to the ED only after the patient has been treated and discharged, following ED clinician interpretation of the X-ray. However, occasionally ED clinicians misdiagnose or fail to detect subtle abnormalities on X-rays, so they conduct a manual radiology report reconciliation process as a safety net. Previous studies addressed this problem of automated reconciliation using ML-based classification solutions that require data samples from the target institution that is heavily based on feature engineering, implying lower transferability between hospitals. In this paper, we investigated the benefits of using pre-trained BERT models for abnormality classification in a cross-institutional setting where data for fine-tuning was unavailable from the target institution. We also examined how the inclusion of synthetically generated radiology reports from ChatGPT affected the performance of the BERT models. Our findings suggest that BERT-like models outperform previously proposed ML-based methods in cross-institutional scenarios, and that adding ChatGPT-generated labelled radiology reports can improve the classifier’s performance by reducing the number of misdiagnosed discharged patients.

This paper investigates the significant impact different prompts have on the behaviour of ChatGPT when used for health information seeking. As people more and more depend on generative large language models (LLMs) like ChatGPT, it is critical to understand model behaviour under different conditions, especially for domains where incorrect answers can have serious consequences such as health. Using the TREC Misinformation dataset, we empirically evaluate ChatGPT to show not just its effectiveness but reveal that knowledge passed in the prompt can bias the model to the detriment of answer correctness. We show this occurs both for retrieve-then-generate pipelines and based on how a user phrases their question as well as the question type. This work has important implications for the development of more robust and transparent question-answering systems based on generative large language models. Prompts, raw result files and manual analysis are made publicly available at https://github.com/ielab/drchatgpt-health_prompting.