Healthcare Data Science (EPSRC CDT)

This course is jointly run by a range of Oxford departments including the departments of Computer Science, Statistics, Engineering Science, the Nuffield Department of Medicine, and the Nuffield Department of Population Health.

The Oxford EPSRC CDT in Healthcare Data Science is based in the Oxford Big Data Institute (BDI) a purpose-built research institute at the heart of the University's biomedical campus. 

The Institute combines researchers from genomics, epidemiology, population health, and infectious disease alongside those from computer science, statistics and engineering to develop the field of big data as applied to biomedical research. Scientists working in the Institute form an analytical hub, deeply connected to the wider experimental and clinical community in Oxford and beyond, working to solve some of the major challenges in medical research. The BDI aims to develop, evaluate and deploy efficient methods for acquiring and analysing information at scale and for exploiting the opportunities presented by large-scale studies. Its activity includes, the analysis population scale data, derived from health records, genetics and biomarkers, the analysis of images and application of machine learning, and the analysis of single cells and molecular proteomic and transcriptomic data.

Course structure

The course begins with a training year, which consists of two terms of intensive training in core data science principles and techniques followed by a third term where you will usually undertake two ten-week research projects in two of your chosen research areas. One of these projects will usually become the basis of your doctoral research, carried out in the following three years.

During the first year, your day will typically comprise of lectures each morning with practical computational exercises each afternoon.

The taught courses covering core subjects such as computational statistics, machine learning, data engineering, ethics and governance, and health research methodology include the following:

• Ethics
• Software Engineering
• Statistical Methods
• Research Methods
• Machine Learning
• Bayesian Statistics
• Medical Imaging
• Biomedical Image Analysis
• Biomedical Time Series Analysis
• Device and Sensor Data
• Genetics
• Infectious Diseases
• Modelling for Policy Making
• Data Governance
• Data Engineering
• Health Data Quality
• Health Data Standards
• Data-driven Innovation.

In each case, you will develop an understanding of relevant concepts and techniques that is not only enough to enable their application and integration but will also serve as a solid foundation should you choose to pursue research in that area.

Each term of taught modules concludes with an extended, team-based two-week data challenge where you will work in small groups with clinicians and domain experts to address questions using large healthcare datasets.

At the start of the second term you will usually select from a pool of projects. These projects are proposed by Oxford faculty members but you may also contact faculty members to jointly propose projects. There are always more projects than students, and students are typically matched to, at least, their first choice, but it is not possible to guarantee that you will be able to work with a particular member of staff. 

You will usually undertake two ten-week placements with research groups within the University. These will provide you with experience of working as part of an active research group and the opportunity to explore specific areas before writing a proposal for your doctoral research.

At the end of the summer of the first year, you will normally select one of the two projects to become the basis of your DPhil research.

In some cases, we are able to offer studentships that are linked to industrial or partner funding. These studentships are usually linked to a specific research project and supervisor. If you are offered one of these studentships, your DPhil research project will be allocated to you prior to the start of your course. 

In years two to four you will carry out individual research on a project within the scope of the programme, specifically the development of novel statistical, machine learning or computational methods with application to health or healthcare data. Training will continue in academic reading, writing and presentation skills, ethics, responsible research and innovation, and career development and planning.

While working on your research project, you will have the opportunity to participate in a range of activities including an ethics placement, four-week external data challenge, seminar series and annual CDT retreats.

Attendance

Resources to support your study

The Big Data Institute has dedicated teaching spaces for classes, workshops, group exercises, and presentations, as well as study space for students during their first year. The institute has many large and small meeting rooms, a large café, and an open, furnished atrium, affording space for formal and informal interaction with research groups, other programmes, and partner organisations. You will have access to a secure research computing infrastructure that supports containerised processing, and you will be able to push your own applications to cloud infrastructure provided by partner organisations. There is central support for common applications and services, including a JupyterHub server for Jupyter notebooks.

The institute houses internationally recognised research groups in genomic medicine, medical image analysis, mobile and sensor data, infectious diseases, and large-scale clinical trials. It is also home to the Ethox Centre and the Wellcome Centre for Ethics and Humanities.

When you move out to your DPhil research department you will also have access to the facilities provided by that department. You will remain a member of the CDT and will retain access to the Big Data Institute.

Supervision

The allocation of graduate supervision for this course is the responsibility of the Medical Sciences Doctoral Training Centre (MSDTC) and it is not always possible to accommodate the preferences of incoming graduate students to work with a particular member of staff. Under exceptional circumstances a supervisor may be found outside the department.

Teaching on taught modules and subsequent research supervision are provided by leading academics from a range of departments at the University. You will benefit from dual supervision for the duration of your research project; at least one of the members of the supervisory team will have a strong background in core data science.

You will have the opportunity to meet your supervisors on a regular basis. These meetings typically take place at least once every two weeks, averaged across the year and agreed by both parties, to discuss your progress.

Assessment

All modules, data challenges and activities during the taught course component involve some aspect of formal assessment, including written reports, problem solving, and group and individual presentations. At the end of year one, you will submit a short DPhil proposal which will be examined orally by the CDT directorate to evaluate your progress and the suitability of the project.

All students will be initially admitted to the status of Probationer Research Student (PRS). Within a maximum of six terms as a PRS student you will be expected to apply for transfer of status from Probationer Research Student to DPhil status. Students who are successful at transfer will also be expected to apply for and gain confirmation of DPhil status within ten terms of admission, to show that your work continues to be on track.

Both milestones normally involve an interview with two assessors (other than your supervisor) and therefore provide important experience for the final oral examination.

You will be expected to submit an original thesis after, at most, four years from the date of admission.

To be successfully awarded a DPhil in Healthcare Data Science you will need to defend your thesis orally (viva voce) in front of two appointed examiners. 

Graduate destinations

It is expected that graduates will be well placed to take on leading roles in industry, academia and the public sector, including areas where health and health care data is used to direct policy or make decisions about patient care.

College preference

The following colleges accept students on the Healthcare Data Science (EPSRC CDT):