Round 2 of the EPSRC Centre for Predictive Modelling in Healthcare's seed corn projects begins

The EPSRC Centre for Predictive Modelling in Healthcare commenced its second round of seed corn projects in July 2017.

The Centre’s seed corn projects are multidisciplinary research projects that bring together researchers with complimentary expertise, to work on a research problem associated with the expertise of researchers within Centre. A budget of up to £10k is provided for essential consumables for each project.

The projects began with a seed corn incubator event held on 25-26 July in Torquay, which brought together the research teams and other experts from the University's commercial team, impact team, public engagement team, and Research Development Managers. Three seed corn projects were funded and a summary of each is provided below:

Predicting appointment numbers in general practice based on characteristics of the registered population - pilot study

General practices/ local doctors’ surgeries are currently under enormous pressure, and have been for some time. On average, people see or speak to their GP or practice nurse around 5 times per year. However, numbers of appointments per year (including telephone appointments) are very different for people of different ages and genders.

At the moment, most general practices do not use any kind of mathematical system, based on the patients registered with them, to predict the number of appointments they should provide. This project will use the information we have collected about appointments together with the mathematical abilities of researchers at the University of Exeter to produce a system for making these kinds of predictions.

If general practices could predict appointment numbers in this way it would help them to make better management decisions on new-patient registrations, recruitment, rotas and the appointment system. It could also help the government and the NHS decide how much money practices should get, based on their registered population.

The project is led by Dr Kate Sidaway-Lee from St Leonard's General Practice, Exeter, and her colleagues Associate Professor Philip Evans and Professor Sir Denis Pereira Gray. They will team up with Professor Krasimira Tsaneva-Atanosova, one of the co-investigators of the Centre, and Research Fellows Dr Margaritis Voliotis and Dr Hossein Mohammadi.

Prospective validation and automated generation of clinical scorecards for the assessment of C. difficile infection severity

C. difficile is a serious infection that can result in death, particularly in the elderly. A significant number of people are admitted to hospital with C. difficile infection symptoms. This project builds on a previously developed scorecard of combinations of symptoms and thresholds (e.g. respiratory rate > 17/min)  for clinical staff to follow to determine the severity of the disease. Since creating the scorecard, the RD&E hospital in Exeter have been using it to prioritise treatment for their patients and we propose to analyse this new dataset to determine how effective it has been in determining disease severity. Additionally, the project aims to create a new computational tool that will automatically mine clinical data to create new scorecards that will be more accurate than the version developed previously. The developed tool could then also be applicable to any disease where suitable clinical dataset exists.

The project is led by Professor Ed Keedwell from the University's Computer Science department, and colleagues Dr Steve Michell from the Biosciences department, and Dr Ray Sheridan, a clinician at the Royal Devon and Exeter Hospital. They will team up with the Centre's Scientific Programmer Dr Diane Fraser and Research Fellow Dr Hossein Mohammadi.

Optimising clinical data for predicting the outcome of epilepsy surgery

Epilepsy is one of the most common brain disorders. It causes sudden, unpredictable seizures and can therefore be quite disabling. One third of patients cannot be treated with drugs, but they could benefit enormously from brain surgery that targets brain areas causing seizures – if those areas can be clearly identified. Current methods, however, are limited in their accuracy, meaning that multiple (and costly) investigations are needed to find the target, and that epilepsy surgery still fails in a considerable number of patients.

To improve this, this project proposes a radically new way of looking at brain activity recordings from epilepsy patients (so-called electroencephalogram, EEG). This method uses the patient’s own EEG combined with advanced mathematics to simulate seizures, and to test what would happen if certain brain areas were operated on. This 'computational operation theatre' could be used to safely test different operation strategies, determine which tests are further needed to refine them, and predict how successful an operation might be.

If successful, this approach could dramatically increase the information gained from standard clinical EEG studies, and lead to highly personalised treatment decisions for people with difficult to treat epilepsies.

The project is led by Centre co-investigator Dr Marc Goodfellow, and clinicians Professor Mark Richardson and Dr Eugenio Abela from Kings College London. They will team up with Centre Research Fellows Dr Leandro Junges and Dr Jennifer Creaser, with assistance from Rhys Lloyd, an intern through the Nuffield Research Placement Programme.

 

The next round of seed corn funding will open in autumn 2017, with an application deadline of January 2018. Please check the seed corn webpages for updates.

Google+