Can hybrid (unsupervised and supervised) text-mining methods be used for efficient, accurate, automated annotation of companion animal electronic health records to better understand drivers of antibiotic use in veterinary care?

Antimicrobial resistance (AMR) is a critical challenge for human and animal health that requires a coordinated endeavour across the disciplines of clinical and data science. This project seeks to embody the interface between these fields by training a PhD student to understand the features and signals that might be available in a large clinical dataset of companion animal clinical records and to develop and apply cutting-edge machine-learning methodologies to derive important insight from these.

You will join the SAVSNet project (www.liv.ac.uk/savsnet), comprising a multidisciplinary team of clinicians, computer and data scientists, epidemiologists and infectious disease experts, a world leading, and BBSRC award-winning team in animal health data science. SAVSNet collects electronic health records from companion animal veterinary practices across the UK. To combat AMR we need to understand the factors that influence antimicrobial prescription (AMP) by veterinary clinicians. Our database of over seven million records includes the clinical free-text narratives and this is where the clinical syndromes that drive AMU will be documented. Clearly, reading millions of records in subtle detail would take a lifetime. Therefore, we need to develop self-training (unsupervised) computer methods that can screen all of our records for this information. To this purpose we have been evaluating topic-modelling, an unsupervised method to recognise underlying themes or topics in collections of text.

You will learn computer skills for manipulation and processing of narratives, evaluating simple text-mining methods and then move on to developing topic models across larger datasets. Once conversant in these methods you will progress through collaboration with the Innovative Computing Group at Durham University to develop and evaluate machine-learning tools for refining these models. This group excels in the development of computer science theory and its application through leading-edge technologies to challenges including computer vision, healthcare, e-learning, and natural language processing. Throughout the PhD you will be given access to teaching modules at Liverpool and Durham on natural language processing and machine learning. Consequently, you will learn to apply the newest methods in natural language processing and artificial intelligence to one of the major global health threats namely antimicrobial resistance.

This project will imbue you with a strong grasp of both the biological question of AMR and the clinical drivers for AMU alongside training in the intriguing science of machine learning. Building a CV with these features will enable you to pursue a leading role in the emerging field of veterinary or medical health informatics.

Informal enquiries may be made to rtnorle@liverpool.ac.uk

HOW TO APPLY

Applications should be made by emailing bbsrcdtp@liverpool.ac.uk with a CV and a covering letter, including whatever additional information you feel is pertinent to your application; you may wish to indicate, for example, why you are particularly interested in the selected project/s and at the selected University. Applications not meeting these criteria will be rejected. We will also require electronic copies of your degree certificates and transcripts.

 In addition to the CV and covering letter, please email a completed copy of the Newcastle-Liverpool-Durham (NLD) BBSRC DTP Studentship Application Details Form (Word document) to bbsrcdtp@liverpool.ac.uk, noting the additional details that are required for your application which are listed in this form. A blank copy of this form can be found at: https://www.nld-dtp.org.uk/how-apply.