Thailand

Biomedical question-answering systems remain popular for biomedical experts interacting with the literature to answer their medical questions. However, these systems are difficult to evaluate in the absence of costly human experts. Therefore, automatic evaluation metrics are often used in this space.
Traditional automatic metrics such as ROUGE or BLEU, which rely on token overlap, have shown a low correlation with humans. 
We present a study that uses large language models (LLMs) to automatically evaluate systems from an international challenge on biomedical semantic indexing and question answering, called BioASQ. We measure the agreement of LLM-produced scores against human judgements. We show that LLMs correlate similarly to lexical methods when using basic prompting techniques. However, by aggregating evaluators with LLMs or by fine-tuning, we find that our methods outperform the baselines by a large margin, achieving a Spearman correlation of 0.501 and 0.511, respectively.

ACL 2024

Using Large Language Models to Evaluate Biomedical Query-Focused Summarisation

human correlation

meta evaluation

prompt engineering

summarisation

reranking

large language models

natural language processing

automatic evaluation

evaluation

workshop paper

### Welcome!
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. Our Virtual Poster Sessions will take place online Thursday, August 22, 2024.

You are required to register for this event. **Please register [here](https://2024.aclweb.org/registration). **

If you have already registered, please check your inbox for an email from Underline granting you access to ACL 2024 content.

Please register!

The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. More information will be announced soon.

Large language models (LLMs) have recently become the leading source of answers for users' questions online. Despite their ability to offer eloquent answers, their accuracy and reliability can pose a significant challenge. This is especially true for sensitive domains such as biomedicine, where there is a higher need for factually correct answers. This paper introduces a biomedical retrieval-augmented generation (RAG) system designed to enhance the reliability of generated responses. The system is based on a fine-tuned LLM for the referenced question-answering, where retrieved relevant abstracts from PubMed are passed to LLM's context as input through a prompt. Its output is an answer based on PubMed abstracts, where each statement is referenced accordingly, allowing the users to verify the answer. Our retrieval system achieves an absolute improvement of 23% compared to the PubMed search engine. Based on the manual evaluation on a small sample, our fine-tuned LLM component achieves comparable results to GPT-4 Turbo in referencing relevant abstracts. We make the dataset used to fine-tune the models and the fine-tuned models based on Mistral-7B-instruct-v0.1 and v0.2 publicly available.

How do you know that? Teaching Generative Language Models to Reference Answers to Biomedical Questions

Domain adaptation of Large Language Models (LLMs) leads to models better suited for a particular domain by capturing patterns from domain text which leads to improvements in downstream tasks. To the naked eye, these improvements are visible; however, the patterns are not so. How can we know which patterns and how much they contribute to changes in downstream scores? Through a Multilevel Analysis we discover and quantify the effect of text patterns on downstream scores of domain-adapted Llama 2 for the task of sentence similarity (BIOSSES dataset). We show that text patterns from PubMed abstracts such as clear writing and simplicity, as well as the amount of biomedical information, are the key for improving downstream scores. Also, we show how another factor not usually quantified contributes equally to downstream scores: choice of hyperparameters for both domain adaptation and fine-tuning.

Multilevel Analysis of Biomedical Domain Adaptation of Llama 2: What Matters the Most? A Case Study

The automatic construction of knowledge graphs (KGs) is an important research area in medicine, with far-reaching applications spanning drug discovery and clinical trial design. These applications hinge on the accurate identification of interactions among medical and biological entities. In this study, we propose an end-to-end machine learning solution based on large language models (LLMs) that utilize electronic medical record notes to construct KGs. The entities used in the KG construction process are diseases, factors, treatments, as well as manifestations that coexist with the patient while experiencing the disease. Given the critical need for high-quality performance in medical applications, we embark on a comprehensive assessment of 12 LLMs of various architectures, evaluating their performance and safety attributes. To gauge the quantitative efficacy of our approach by assessing both precision and recall, we manually annotate a dataset provided by the Macula and Retina Institute. We also assess the qualitative performance of LLMs, such as the ability to generate structured outputs or the tendency to hallucinate. The results illustrate that in contrast to encoder-only and encoder-decoder, decoder-only LLMs require further investigation. Additionally, we provide guided prompt design to utilize such LLMs. The application of the proposed methodology is demonstrated on age-related macular degeneration.

Large Language Models for Biomedical Knowledge Graph Construction: Information extraction from EMR notes

Electronic Health Records (EHR) serve as a valuable source of patient information, offering insights into medical histories, treatments, and outcomes. Previous research has developed systems for detecting applicable ICD codes that should be assigned while writing a given EHR document, mainly focusing on discharge summaries written at the end of a hospital stay. In this work, we investigate the potential of predicting these codes for the whole patient stay at different time points during their stay, even before they are officially assigned by clinicians. The development of methods to predict diagnoses and treatments earlier in advance could open opportunities for predictive medicine, such as identifying disease risks sooner, suggesting treatments, and optimizing resource allocation. Our experiments show that predictions regarding final ICD codes can be made already two days after admission and we propose a custom model that improves performance on this early prediction task.

Continuous Predictive Modeling of Clinical Notes and ICD Codes in Patient Health Records

Healthcare professionals often manually extract information from large clinical documents to address patient-related questions. The use of Natural Language Processing (NLP) techniques, particularly Question Answering (QA) models, is a promising direction for improving the efficiency of this process. However, document-level QA from large documents is often impractical or even infeasible (for model training and inference). In this work, we solve the document-level QA from clinical reports in a two-step approach: first, the entire report is split into segments and for a given question the most relevant segment is predicted by a NLP model; second, a QA model is applied to the question and the retrieved segment as context. We investigate the effectiveness of heading-based and naive paragraph segmentation approaches for various paragraph lengths on two subsets of the emrQA dataset. Our experiments reveal that an average paragraph length used as a parameter for the segmentation has no significant effect on performance during the whole document-level QA process. That means experiments focusing on segmentation into shorter paragraphs perform similarly to those focusing on entire unsegmented reports. Surprisingly, naive uniform segmentation is sufficient even though it is not based on prior knowledge of the clinical document's characteristics.

Paragraph Retrieval for Enhanced Question Answering in Clinical Documents

The success of pretrained language models (PLMs) across a spate of use-cases has led to significant investment from the NLP community towards building domain-specific foundational models. On the other hand, in mission critical settings such as biomedical applications, other aspects also factor in‚Äîchief of which is a model's ability to produce reasonable estimates of its own uncertainty. In the present study, we discuss these two desiderata through the lens of how they shape the entropy of a model's output probability distribution. We find that domain specificity and uncertainty awareness can often be successfully combined, but the exact task at hand weighs in much more strongly.

Domain-specific or Uncertainty-aware models: Does it really make a difference for biomedical text classification?

Developing imaging models capable of detecting pathologies from chest X-rays can be cost and time-prohibitive for large datasets as it requires supervision to attain state-of-the-art performance. Instead, labels extracted from radiology reports may serve as distant supervision since these are routinely generated as part of clinical practice. Despite their widespread use, current rule-based methods for label extraction rely on extensive rule sets that are limited in their robustness to syntactic variability. To alleviate these limitations, we introduce RadPert, a rule-based system that integrates an uncertainty-aware information schema with a streamlined set of rules, enhancing performance. Additionally, we have developed RadPrompt, a multi-turn prompting strategy that leverages RadPert to bolster the zero-shot predictive capabilities of large language models, achieving a statistically significant improvement in weighted average F1 score over GPT-4 Turbo. Most notably, RadPrompt surpasses both its underlying models, showcasing the synergistic potential of LLMs with rule-based models. We have evaluated our methods on two English Corpora: the MIMIC-CXR gold-standard test set and a gold-standard dataset collected from the Cambridge University Hospitals.

Can Rule-Based Insights Enhance LLMs for Radiology Report Classification? Introducing the RadPrompt Methodology.

Instruction-tuned Large Language Models (LLMs) can perform a wide range of tasks given natural language instructions to do so, but they are sensitive to how such instructions are phrased. 
This issue is especially concerning in healthcare, as clinicians are unlikely to be experienced prompt engineers and the potential consequences of inaccurate outputs are heightened in this domain. 

This raises a practical question: How robust are instruction-tuned LLMs to natural variations in the instructions provided for clinical NLP tasks?
We collect prompts from medical doctors across a range of tasks and quantify the sensitivity of seven LLMs---some general, others specialized---to natural (i.e., non-adversarial) instruction phrasings. 
We find that performance varies substantially across all models, and that---perhaps surprisingly---domain-specific models explicitly trained on clinical data are especially brittle, compared to their general domain counterparts. 
Further, arbitrary phrasing differences can affect fairness, e.g., valid but distinct instructions for mortality prediction yield a range both in overall performance, and in terms of differences between demographic groups.

Open (Clinical) LLMs are Sensitive to Instruction Phrasings

Large language models (LLMs) have shown remarkable performance on many tasks in different domains. However, their performance in contextual biomedical machine reading comprehension (MRC) has not been evaluated in depth. In this work, we evaluate GPT on four contextual biomedical MRC benchmarks. We experiment with different conventional prompting techniques as well as introduce our own novel prompting method. To solve some of the retrieval problems inherent to LLMs, we propose a prompting strategy named Implicit Retrieval Augmented Generation (RAG) that alleviates the need for using vector databases to retrieve important chunks in traditional RAG setups. Moreover, we report qualitative assessments on the natural language generation outputs from our approach. The results show that our new prompting technique is able to get the best performance in two out of four datasets and ranks second in rest of them. Experiments show that modern-day LLMs like GPT even in a zero-shot setting can outperform supervised models, leading to new state-of-the-art (SoTA) results on two of the benchmarks.

Can GPT Redefine Medical Understanding? Evaluating GPT on Biomedical Machine Reading Comprehension

The consequences of a healthcare data breach can be devastating for the patients, providers, and payers. The average financial impact of a data breach in recent months has been estimated to be close to USD 10 million. This is especially significant for healthcare organizations in India that are managing rapid digitization while still establishing data governance procedures that align with the letter and spirit of the law. Computer-based systems for de-identification of personal information are vulnerable to data drift, often rendering them ineffective in cross-institution settings. Therefore, a rigorous assessment of existing de-identification against local health datasets is imperative to support the safe adoption of digital health initiatives in India. Using a small set of de-identified patient discharge summaries provided by an Indian healthcare institution, in this paper, we report the nominal performance of de-identification algorithms (based on language models) trained on publicly available non-Indian datasets, pointing towards a lack of cross-institutional generalization. Similarly, experimentation with off-the-shelf de-identification systems reveals potential risks associated with the approach. To overcome data scarcity, we explore generating synthetic clinical reports (using publicly available and Indian summaries) by performing in-context learning over Large Language Models (LLMs). Our experiments demonstrate the use of generated reports as an effective strategy for creating high-performing de-identification systems with good generalization capabilities.

Premium content

Using Large Language Models to Evaluate Biomedical Query-Focused Summarisation

Next from ACL 2024

How do you know that? Teaching Generative Language Models to Reference Answers to Biomedical Questions

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES