Bayesian Nonparametric Inference for Stochastic Epidemic Models

Understanding the spread of communicable infectious diseases is of great importance in order to prevent major future outbreaks and therefore it remains high on the global scientific agenda. It has been widely recognised that mathematical and statistical modelling has become a valuable tool in the analysis of infectious disease dynamics by supporting the development of control strategies, informing policy-making at the highest levels, and in general playing a fundamental role in the fight against the spread of disease.

Despite the enormous attention given to the development of methods for efficient parameter estimation, there has been relatively little activity in the area of non-parametric inference. That is, drawing inference for the quantities which govern transmission, i) the force of infection and ii) the period during which an individual remains infectious, without making certain modelling assumptions about its (parametric) functional form or that it belongs to a certain family of parametric distributions.

The proposed research is concerned with the development of new methodology which will enable non-parametric estimation of the parameters which govern transmission within a Bayesian framework and the application of the proposed methods to large disease outbreak datasets.

Expected beneficiaries of the proposed research are not only statisticians but also clinicians who are interested in the application of efficient statistical modelling tools to analyse the data that have been acquired either via surveillance or by specific studies. In addition, likely 'users' and beneficiaries of the proposed research are policy makers who need to base their decisions based on robust estimates of important epidemiological quantities both at national and international level. It is highly likely that various agencies, departments and organisations which are concerned with disease outbreaks will be interested in the developments of the proposed research; examples include researchers (e.g. modellers) working in the UK's Heath Protection Agency (HPA), the Department of Environment, Food and Rural Affairs (DEFRA), the European Centre for Disease Prevention (ECDC) and World Health Organization (WHO). It is anticipated that the open-source software which will be made available during the course of the project will ensure that the impact of the proposed research has a more immediate effect.

Infectious diseases account for more than 17 million deaths worldwide each year. Understanding the spread of communicable infectious diseases is of great importance in order to prevent major future outbreaks. It has been widely recognised that mathematical modelling has become a valuable tool in the analysis of infectious disease dynamics and in general, is playing a fundamental role in the fight against disease spread. Therefore, the wider public is another beneficiary since the proposed work is within the area of modelling communicable infectious diseases providing new efficient tools to combat infectious diseases. Analysing past and future disease outbreak datasets effectively using the proposed methodology will shed more light as to the key parameters that govern transmission. In turn, this will allow for efficient control-strategies. Therefore, the proposed research has the potential to contribute to the the public's health, both nationally and internationally in short and long term.

The PDRA to be employed will receive the necessary training to achieve familiarity with the project. This will include, for example, the review of the relevant literature and developing computer code to analyse simple epidemic models. The PRRA during the course of the proposed research will acquire skills both in advanced mathematical modelling as well as computationally intensive methods (e.g MCMC). These skills are highly transferable and not only can be applied within the academic sector, but also within the industry. Mathematical modelling and programming skills are essential to any problem of practical significance. Moreover, apart from developing skills in independent working, time management and project organisation, the PDRA will also develop presentation skills by being member of two active research groups (i.e. the Epidemic Modelling and the Computational Statistics Groups). All these, combined with the wide range of supplementary courses on offer within the University, e.g. in business and networking skills, will positively impact on the future career prospects of the PDRA.

The award of this First Grant would enable the PI to establish an independent research career and to build a research group centred around his research interests and methodologies. Through the proposed work he will fully capitalise on his promising collaboration with Prof Ghahramani's and Dr Griffin's groups, along with his existing strong links with Prof O'Neill at the University of Nottingham. It would also allow him to continue to foster his effective collaborations with members from the health sector (for example, Dr Jonathan Edgweworth, Head of Pathology and Associate Medical Director at Guy's and St Thomas' Hospital Foundation Trust) which bring fundamental science solutions to the fight against infectious diseases.