PhD Studentship opportunity at IHI, funded by the NIHR UCLH/UCL BRC

Developing methods for multi-dimensional analysis in health records

Primary supervisors: Dr Reecha Sofat and Dr Dionisio Acosta

Secondary supervisors: Prof Harry Hemingway and Prof Aroon Hingorani

Application deadline: Wednesday 29 August 2018, h.17:00

Interview date: 7 September 2018

The studentship must begin by October 2018

Residency requirements

Studentships are open to all UK applicants. Applicants are also eligible for a studentship if they have been an ordinary resident in the UK for three years prior to the start of the studentship grant. For instance, if the applicant applies for a studentship to start in October 2018, they must have resided in the UK since October 2014. Please note: if the applicant is from an EU-country, these three years may include time spent studying however if the applicant is from outside the EU (international), these three years cannot include time spent studying at a Higher Education institution.

General Description

In medicine each physician does their best to provide care for patients through careful examination, recording of clinical findings and response to treatment. However, in a paper-based system this does not enable the whole system to learn and discover. As a result, we categorise disease according to historic syndromes, yet the biology for sub-types of disease may be very different. This has implications on the way disease is treated, prevented and prognosis is predicted. One way to now overcome these limitations is the use of digital data. Electronic health care records were introduced in the first instance to improve quality and safety of care for patients; however, over time their utility in investigating aetiology, prognosis and treatment response has become clear. Merging primary care and secondary care records, the latter which also include detailing phenotyping of disease, structured data from biochemical analyses, radiological data and unstructured data from clinical notes and adding additional layer of –omics data (including genomics, proteomics and metabolomics) now allows us to redefine the taxonomy of many diseases. This has direct relevance to the underlying biology as it will allow discovery of treatments of disease subtypes and enable trials to be conducted through the electronic health care record allowing not only the system to learn but also the practice of precision medicine.

In this PhD the student will use data sources from primary and secondary care merged with multi-omics to enable development of algorithms to analyse multi-dimensional data, deploy machine learning for discovery, prediction and clinical decision support. Statistical techniques and machine learning methods will be utilised for integration to begin to develop models of defining complex disease syndromes, e.g. heart failure and stroke using biological data rather than physician defined syndromes.  An important aspect in this research is understanding disease progression, in terms of the interaction between different the disease sub-phenotype and the health-care interventions along the clinical care pathways captured in the EHR data (longitudinal phenotypes).  This data driven approach to discovery and delivery in health care is now possible and vital in order to ensure the health of the population. In this way, this PhD is very focussed on achieving translational platforms that can be used within the clinical arena or as tools for the next generation of interventional trials focussing on disease subtypes.

The student will be using both statistical and machine learning methods to develop algorithms therefore an interest in not only the clinical challenge but a good grounding in either statistics, computer sciences and /or machine learning are desirable.

The student will be based at the UCL Institute of Health Informatics (IHI) at 222 Euston Road, London NW1 2DA. UCL IHI has been established to conduct high quality research that leverages health informatics approaches at local, national and international levels, working closely with the UK-wide Health Data Research. 

Person specification

Essential criteria

• Minimum of 2:1 BSc in biomedical, statistics or computing and/or a Master’s degree in computational statistics, machine learning or other quantitative discipline (preferably with a merit or distinction)
• Experience in quantitative data analysis
• Ability to organise and prioritise workload
• Ability to work as part of a team

 Desirable

• Experience in statistical analysis and the use of programmes such as R
• Excellent verbal and written communication skills (ranging from informal 1:1 discussion to formal presentations)
• Experience in analysing electronic health records and/or routine datasets

How to apply

Please submit applications in the following format:

• A CV, including full details of all University course grades to date.
• Contact details for two academic or professional referees (at least one academic).
• A personal statement (750 words maximum) outlining (i) your suitability for the project with reference to the criteria in the person specification, (ii) what you hope to achieve from the PhD and (iii) your research experience to-date.

Please include a contact telephone number and an email address where you can be easily reached. References will be taken up for all short-listed candidates.

Shortlisted candidates will be sent a technical exercise from which they will be asked to present at interview, this will form part of the basis of the selection.

Please send electronic applications to: d.kirkwood@ucl.ac.uk