Bayesian Non-Parametric Methodology for Stochastic Epidemic Models
Lead Research Organisation:
University of Nottingham
Department Name: Sch of Mathematical Sciences
Abstract
Simulating from and making inference for stochastic epidemic models are key strategies for understanding and controlling the spread of infectious diseases. Despite the enormous attention given to methods for parameter estimation, there has been relatively little activity in the area of nonparametric inference. That is, drawing inference for the infection rate without making specific modelling assumptions about its functional form. In this project we are concerned with heterogeneously mixing models in which the infection rate between two individuals is a function of their characteristics, for example location or type. We develop a novel method for inferring the function nonparametrically, removing the need to make questionable parametric assumptions. We adopt a Bayesian approach by assigning a Gaussian Process prior distribution to the infection rate function and then develop an efficient data augmentation Markov Chain Monte Carlo methodology to estimate the infection rate function, the GP hyperparameters and the unobserved infection times. We will illustrate our methodology using simulated data and as well as data on an Avian Influenza outbreak from the Netherlands.
Organisations
Studentship Projects
Project Reference | Relationship | Related To | Start | End | Student Name |
---|---|---|---|---|---|
EP/N50970X/1 | 01/10/2016 | 30/09/2021 | |||
1800036 | Studentship | EP/N50970X/1 | 01/10/2016 | 29/12/2019 | Rowland Seymour |
Description | We have been developing new nonparametric methods to model how infectious diseases spread. Using our method you do not need to assume the functional form of the infection rate between any two individuals, only the nature of the relationship between them. We are using a likelihood based inference approach, which is challenging as there is a often a large amount of missing data in epidemic data sets, for example the time individuals were infected. We have developed an efficient MCMC framework using Gaussian Processes (GPs) to infer the infection rate, and we can also infer the GP hyperparameters, and the times individuals were infected. We are also developing multi-output GPs to model infection rates where the probability of being infection also depends on the type of individual, e.g. an individual's sex. The method allows us to make the most efficient use of the data as information can be shared among the GPs when learning the infection rates. We have investigated an outbreak of Highly Pathogenic Avian Influenza (H7N7) in the Netherlands, where the infection rate between any two farms depends on the distance between them. We have also analysed an outbreak of Foot and Mouth Disease in the UK where the infection rates depends on the distance between any two farms and the type of animals on the susceptible farm. Our findings show that farms with more than one type on animal present were more susceptible to the disease, than those where only one type of animal was present. |
Exploitation Route | My findings may be used by government authorities to develop better control strategies for outbreaks of infectious diseases among farm animals. For example, if we can better understand the probability of one farm infecting another and how this depends on their locations, and the number or type of animals on them, we can develop and implement better vaccination and culling strategies. |
Sectors | Agriculture, Food and Drink,Environment |
Description | We have collaborated with members of the Wageningen Biovetinary Research Unit, the Netherlands, and used data from the Netherlands Food and Consumer Product Authority and the Dutch Ministry of Agriculture, Nature and Food Quality. This was to analyse an outbreak of Avian Influenza. We met with a member of the research unit to discuss the findings of our research and how this could affect culling strategies in future outbreaks of Avian Influenza. |
First Year Of Impact | 2019 |
Sector | Agriculture, Food and Drink |
Impact Types | Policy & public services |
Description | Royal Society RAMP Rapid Review Group Member |
Geographic Reach | National |
Policy Influence Type | Implementation circular/rapid advice/letter to e.g. Ministry of Health |
Description | Submitted paper to the Dutch Ministry of Agriculture, Nature and Food Quality |
Geographic Reach | National |
Policy Influence Type | Implementation circular/rapid advice/letter to e.g. Ministry of Health |
Description | EPSRC Doctoral Prize |
Amount | £58,000 (GBP) |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 07/2020 |
End | 06/2022 |
Description | Graduate School Travel Prize |
Amount | £300 (GBP) |
Organisation | University of Nottingham |
Sector | Academic/University |
Country | United Kingdom |
Start | 06/2019 |
End | 06/2019 |
Description | Junior Travel Award |
Amount | £400 (GBP) |
Organisation | International Society for Bayesian Analysis (ISBA) |
Sector | Charity/Non Profit |
Country | United States |
Start | 06/2019 |
End | 06/2019 |
Title | BNP 4 HMSEM |
Description | The repositroy contains code for simulating outbreaks of heterogneously mixing infectious diseases and fitting models to outbreak data. In the model, the infection rate from individual i to j is beta{ij}. The methodology is Bayesian nonparametric meaning the functional form of beta{ij} does not need to be specified. The form is inferred using a Gaussian process. The model allows for the population to be split into a number of types, e.g. cattle farms and pig farms, and the infection rate to depend on the type of infectious farm. In cases where the times individuals are infected are unknown, these can be inferred using the code provided. |
Type Of Material | Computer model/algorithm |
Year Produced | 2021 |
Provided To Others? | Yes |
Impact | The has provided analysis for an outbreak of foot and mouth disease in Cumbria. |
URL | https://github.com/rowlandseymour/BNP_4_HMSEM |
Description | University of Washington Summer Institute for Modelling Infectious Diseases |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | Around 50 clinicians and epidemiologist took part in a workshop on MCMC for epidemic models on which I was a teaching associate. |
Year(s) Of Engagement Activity | 2021 |
URL | https://si.biostat.washington.edu/suminst/archives/SISMID2021/MD2111 |
Description | University of Washington Summer Institute for Modelling Infectious Diseases |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | Around 50 clinicians and epidemiologist took part in a workshop on MCMC for epidemic models on which I was a teaching associate. |
Year(s) Of Engagement Activity | 2020 |
URL | https://si.biostat.washington.edu/suminst/archives/SISMID2020/MD2012 |