arXiv subjects

Bruce Guthrie

Publications and source records attributed to Bruce Guthrie.

7 recordsLinked to original sources

arXiv2026

Explainable AI for Data-Driven Design of High-Dimensional Predictive Studies

Predictive modelling is important for health data analysis and data-driven clinical decision-making. However, predictive studies are challenging to design optimally by hand when tens or even hundreds of features require selection, transformation, or interaction modelling. While complex machine learning models offer high performance, their "black-box" nature limits the clinical trust, transparency, and interpretability required for decision-making. We developed and evaluated an Exploratory AI Recommender that provides data-driven recommendations to improve predictive performance of existing interpretable statistical models. The developed framework uses flexible AI modelling to capture complex data patterns and explainable AI techniques to translate the patterns into three recommendation types: feature exclusion, non-linear terms, and feature interactions. We evaluated the framework by comparing predictive performance of a baseline (i.e., no interactions or non-linear terms) Cox Proportional Hazards (CPH) model against an augmented CPH incorporating recommendations suggested by our method. The primary analysis predicts the time to the first occurrence of a fall or related injury in 245,614 patients. Our method recommended excluding 23 features, including non-linear terms for two features, and including 221 suggested feature interactions. The C-index improved from 0.805 (95% CI 0.798-0.812) to 0.815 (95% CI 0.809-0.822), and so did calibration (intercept: -0.006 to 0.003; slope: 1.063 to 0.950). All recommendations were supported by existing literature. The method also proved effective on two additional public datasets, demonstrating wider applicability. The proposed Exploratory AI Recommender demonstrates the potential of explainable AI and data-driven study design to improve the process of developing, and the performance of high-dimensional transparent predictive models.

cs.LG↗

arXiv2025

Adverse Event Extraction from Discharge Summaries: A New Dataset, Annotation Scheme, and Initial Findings

In this work, we present a manually annotated corpus for Adverse Event (AE) extraction from discharge summaries of elderly patients, a population often underrepresented in clinical NLP resources. The dataset includes 14 clinically significant AEs-such as falls, delirium, and intracranial haemorrhage, along with contextual attributes like negation, diagnosis type, and in-hospital occurrence. Uniquely, the annotation schema supports both discontinuous and overlapping entities, addressing challenges rarely tackled in prior work. We evaluate multiple models using FlairNLP across three annotation granularities: fine-grained, coarse-grained, and coarse-grained with negation. While transformer-based models (e.g., BERT-cased) achieve strong performance on document-level coarse-grained extraction (F1 = 0.943), performance drops notably for fine-grained entity-level tasks (e.g., F1 = 0.675), particularly for rare events and complex attributes. These results demonstrate that despite high-level scores, significant challenges remain in detecting underrepresented AEs and capturing nuanced clinical language. Developed within a Trusted Research Environment (TRE), the dataset is available upon request via DataLoch and serves as a robust benchmark for evaluating AE extraction methods and supporting future cross-dataset generalisation.

cs.CL↗

arXiv2025

Age-stratified clustering of multiple long-term conditions

Background: Most people with any long-term condition have multiple long-term conditions, but our understanding of how conditions cluster is limited. Many clustering studies identify clusters in the whole population, but the clusters that occur in people of different ages may be distinct. The aim of this paper was to explore similarities and differences in clusters found in different age-groups. Method: We present a method for finding similar clusters in multiple age-groups, referred to as cluster sets, using Latent Class Analysis (LCA) and Chebyshev distance metric. We analyse a primary care electronic health record (EHR) dataset recording the presence of 40 long-term conditions (LTCs) in 570,355 people aged 40-99 years with at least one of these conditions, analysing in five-year age-groups. Findings: We find that the 600 clusters found separately in 12 age-strata can be summarised by 342 cluster sets with 263 cluster sets only being found in a single age-group (singleton cluster sets), and 79 cluster sets being present in multiple age-groups. We observe that 31 conditions of the 40 conditions studied appear in cluster sets with the respective condition being the only condition present with a very high prevalence of more than 0.9 whereas the remaining cluster sets typically contain two to four conditions present with a high prevalence of more than 0.7. Interpretation: Multimorbidity profiles in different age-groups are often distinct (singleton cluster sets observed only in that age-group), but similar clusters with small variations in their composition are also found in multiple age-groups. This demonstrates the age dependency of MLTC clusters and presents a case for age-stratified clustering.

stat.AP↗

arXiv2025

What do we mean when we say we are clustering multimorbidity?

Clustering multimorbidity has been a global research priority in recent years. Existing studies usually identify these clusters using one of several popular clustering methods and then explore various characteristics of these clusters, e.g., their genetic underpinning or their sociodemographic drivers, as downstream analysis. These studies make several choices during clustering that are often not explicitly acknowledged in the literature, e.g., whether they are clustering conditions or clustering individuals, and thus, they lead to different clustering solutions. We observe that, in general, clustering multimorbidity might mean different things in different studies, and argue that making these choices more explicit and, more importantly, letting the downstream analysis, or the purpose of identifying multimorbidity clusters, guide these choices, might lead to more transparent and operationalizable multimorbidity clusters. In this study, we discuss various purposes of identifying multimorbidity clusters and build a case for how different purposes can justify the different choices in data and methods.

stat.AP↗

arXiv2024

Probabilistic Modelling of Multiple Long-Term Condition Onset Times

The co-occurrence of multiple long-term conditions (MLTC), or multimorbidity, in an individual can reduce their lifespan and severely impact their quality of life. Exploring the longitudinal patterns, e.g. clusters, of disease accrual can help better understand the genetic and environmental drivers of multimorbidity, and potentially identify individuals who may benefit from early targeted intervention. We introduce $\textit{probabilistic modelling of onset times}$, or $\texttt{ProMOTe}$, for clustering and forecasting MLTC trajectories. $\texttt{ProMOTe}$ seamlessly learns from incomplete and unreliable disease trajectories that is commonplace in Electronic Health Records but often ignored in existing longitudinal clustering methods. We analyse data from 150,000 individuals in the UK Biobank and identify 50 clusters showing patterns of disease accrual that have also been reported by some recent studies. We further discuss the forecasting capabilities of the model given the history of disease accrual.

stat.AP↗

arXiv2024

Infusing clinical knowledge into tokenisers for language models

This study introduces a novel knowledge enhanced tokenisation mechanism, K-Tokeniser, for clinical text processing. Technically, at initialisation stage, K-Tokeniser populates global representations of tokens based on semantic types of domain concepts (such as drugs or diseases) from either a domain ontology like Unified Medical Language System or the training data of the task related corpus. At training or inference stage, sentence level localised context will be utilised for choosing the optimal global token representation to realise the semantic-based tokenisation. To avoid pretraining using the new tokeniser, an embedding initialisation approach is proposed to generate representations for new tokens. Using three transformer-based language models, a comprehensive set of experiments are conducted on four real-world datasets for evaluating K-Tokeniser in a wide range of clinical text analytics tasks including clinical concept and relation extraction, automated clinical coding, clinical phenotype identification, and clinical research article classification. Overall, our models demonstrate consistent improvements over their counterparts in all tasks. In particular, substantial improvements are observed in the automated clinical coding task with 13\% increase on Micro $F_1$ score. Furthermore, K-Tokeniser also shows significant capacities in facilitating quicker converge of language models. Specifically, using K-Tokeniser, the language models would only require 50\% of the training data to achieve the best performance of the baseline tokeniser using all training data in the concept extraction task and less than 20\% of the data for the automated coding task. It is worth mentioning that all these improvements require no pre-training process, making the approach generalisable.

cs.CL↗

arXiv2024

Frailty or Frailties: Exploring Frailty Index Subdimensions in the English Longitudinal Study of Ageing

Background: Frailty, a state of increased vulnerability to adverse health outcomes, has garnered significant attention in research and clinical practice. Existing constructs aggregate clinical features or health deficits into a single score. While simple and interpretable, this approach may overlook the complexity of frailty and not capture the full range of variation between individuals. Methods: Exploratory factor analysis was used to infer latent dimensions of a frailty index constructed using survey data from the English Longitudinal Study of Ageing (ELSA), wave 9. The dataset included 58 self-reported health deficits in a representative sample of community-dwelling adults aged 65+ (N = 4971). Deficits encompassed chronic disease, general health status, mobility, independence with activities of daily living, psychological wellbeing, memory and cognition. Multiple linear regression examined associations with CASP-19 quality of life scores. Results: Factor analysis revealed four frailty subdimensions. Based on the component deficits with the highest loading values, these factors were labelled "Mobility Impairment and Physical Morbidity", "Difficulties in Daily Activities", "Mental Health" and "Disorientation in Time". The four subdimensions were a better predictor of quality of life than frailty index scores. Conclusions: Distinct subdimensions of frailty can be identified from standard index scores. A decomposed approach to understanding frailty has potential to provide a more nuanced understanding of an individual's state of health across multiple deficits.

stat.AP↗