arXiv subjects

Louise Matthews

Publications and source records attributed to Louise Matthews.

5 recordsLinked to original sources

arXiv2022

Visualization for Epidemiological Modelling: Challenges, Solutions, Reflections & Recommendations

We report on an ongoing collaboration between epidemiological modellers and visualization researchers by documenting and reflecting upon knowledge constructs -- a series of ideas, approaches and methods taken from existing visualization research and practice -- deployed and developed to support modelling of the COVID-19 pandemic. Structured independent commentary on these efforts is synthesized through iterative reflection to develop: evidence of the effectiveness and value of visualization in this context; open problems upon which the research communities may focus; guidance for future activity of this type; and recommendations to safeguard the achievements and promote, advance, secure and prepare for future collaborations of this kind. In describing and comparing a series of related projects that were undertaken in unprecedented conditions, our hope is that this unique report, and its rich interactive supplementary materials, will guide the scientific community in embracing visualization in its observation, analysis and modelling of data as well as in disseminating findings. Equally we hope to encourage the visualization community to engage with impactful science in addressing its emerging data challenges. If we are successful, this showcase of activity may stimulate mutually beneficial engagement between communities with complementary expertise to address problems of significance in epidemiology and beyond. https://ramp-vis.github.io/RAMPVIS-PhilTransA-Supplement/

cs.HC↗

arXiv2022

FAIR Data Pipeline: provenance-driven data management for traceable scientific workflows

Modern epidemiological analyses to understand and combat the spread of disease depend critically on access to, and use of, data. Rapidly evolving data, such as data streams changing during a disease outbreak, are particularly challenging. Data management is further complicated by data being imprecisely identified when used. Public trust in policy decisions resulting from such analyses is easily damaged and is often low, with cynicism arising where claims of "following the science" are made without accompanying evidence. Tracing the provenance of such decisions back through open software to primary data would clarify this evidence, enhancing the transparency of the decision-making process. Here, we demonstrate a Findable, Accessible, Interoperable and Reusable (FAIR) data pipeline developed during the COVID-19 pandemic that allows easy annotation of data as they are consumed by analyses, while tracing the provenance of scientific outputs back through the analytical source code to data sources. Such a tool provides a mechanism for the public, and fellow scientists, to better assess the trust that should be placed in scientific evidence, while allowing scientists to support policy-makers in openly justifying their decisions. We believe that tools such as this should be promoted for use across all areas of policy-facing research.

q-bio.QM↗

arXiv2019

A hypothesis testing framework for the ratio of means of two negative binomial distributions: classifying the efficacy of anthelmintic treatment against intestinal parasites

Over-dispersed count data typically pose a challenge to analysis using standard statistical methods, particularly when evaluating the efficacy of an intervention through the observed effect on the mean. We outline a novel statistical method for analysing such data, along with a statistically coherent framework within which the observed efficacy is assigned one of four easily interpretable classifications relative to a target efficacy: "adequate", "reduced", "borderline" or "inconclusive". We illustrate our approach by analysing the anthelmintic efficacy of mebendazole using a dataset of egg reduction rates relating to three intestinal parasites from a treatment arm of a randomised controlled trial involving 91 children on Pemba Island, Tanzania. Numerical validation of the type I error rates of the novel method indicate that it performs as well as the best existing computationally-simple method, but with the additional advantage of providing valid inference in the case of an observed efficacy of 100%. The framework and statistical analysis method presented also allow the required sample size of a prospective study to be determined via simulation. Both the framework and method presented have high potential utility within medical parasitology, as well as other fields where over-dispersed count datasets are commonplace. In order to facilitate the use of these methods within the wider medical community, user interfaces for both study planning and analysis of existing datasets are freely provided along with our open-source code via: http://www.fecrt.com/framework

stat.ME↗

arXiv2019

Estimation of temporal covariances in pathogen dynamics using Bayesian multivariate autoregressive models

It is well recognised that animal and plant pathogens form complex ecological communities of interacting organisms within their hosts. Although community ecology approaches have been applied to determine pathogen interactions at the within-host scale, methodologies enabling robust inference of the epidemiological impact of pathogen interactions are lacking. Here we developed a novel statistical framework to identify statistical covariances from the infection time-series of multiple pathogens simultaneously. Our framework extends Bayesian multivariate disease mapping models to analyse multivariate time series data by accounting for within- and between-year dependencies in infection risk and incorporating a between-pathogen covariance matrix which we estimate. Importantly, our approach accounts for possible confounding drivers of temporal patterns in pathogen infection frequencies, enabling robust inference of pathogen-pathogen interactions. We illustrate the validity of our statistical framework using simulated data and applied it to diagnostic data available for five respiratory viruses co-circulating in a major urban population between 2005 and 2013: adenovirus, human coronavirus, human metapneumovirus, influenza B virus and respiratory syncytial virus. We found positive and negative covariances indicative of epidemiological interactions among specific virus pairs. This statistical framework enables a community ecology perspective to be applied to infectious disease epidemiology with important utility for public health planning and preparedness.

stat.ME↗

arXiv2016

How to partition diversity

Diversity measurement underpins the study of biological systems, but measures used vary across disciplines. Despite their common use and broad utility, no unified framework has emerged for measuring, comparing and partitioning diversity. The introduction of information theory into diversity measurement has laid the foundations, but the framework is incomplete without the ability to partition diversity, which is central to fundamental questions across the life sciences: How do we prioritise communities for conservation? How do we identify reservoirs and sources of pathogenic organisms? How do we measure ecological disturbance arising from climate change? The lack of a common framework means that diversity measures from different fields have conflicting fundamental properties, allowing conclusions reached to depend on the measure chosen. This conflict is unnecessary and unhelpful. A mathematically consistent framework would transform disparate fields by delivering scientific insights in a common language. It would also allow the transfer of theoretical and practical developments between fields. We meet this need, providing a versatile unified framework for partitioning biological diversity. It encompasses any kind of similarity between individuals, from functional to genetic, allowing comparisons between qualitatively different kinds of diversity. Where existing partitioning measures aggregate information across the whole population, our approach permits the direct comparison of subcommunities, allowing us to pinpoint distinct, diverse or representative subcommunities and investigate population substructure. The framework is provided as a ready-to-use R package to easily test our approach.

q-bio.QM↗