Mick Roberts - Research

Recent and current research

The evolution and transmission of a virus
A simplified model for the evolution and transmission of a mutating virus is described. The essential features of the model are: within host evolution; host heterogeneity; between host transmission and dynamic viral diversity. The virus replicates within the host, generating variants through random mutation. A virus with a mutation in an epitope that would be otherwise recognised by the host has an advantage in being able to escape at least that component of the immune response. On the other hand, a virus with a mutation in a non-recognised epitope bears a cost. Replication of the virus is modulated by the cytotoxic T lymphocyte (CTL) response of the host. The strength of the response depends on the human leukocyte alleles (HLA) of class I expressed by the host. Hence, for the purpose of the model we divide the human host population into types, where those of the same type are HLA-similar with regard to their response to the virus types represented in the model. We model the transmission of the virus using an SIR model, with no recovery resulting in immunity and infectious individuals having an increased death rate. Hence we are motivated by the transmission of the HIV virus. The final complication in the model gives rise to dynamic viral diversity via selection by CTLs at the time of transmission between HLA-disparate individuals. An inoculum of virus transmitted to a host expressing different HLA class I alleles to the donor will give rise to an outgrowth of the quasispecies that is most fit in the recipient host. This is on a much shorter time-scale (measured in days) than the evolution of the virus within the host resulting from random mutation. Development and analysis of the model is part of an ongoing collaboration with Prof. Angela McLean, University of Oxford.

Strategies for minimising the risk of importing exotic diseases
The danger of a global influenza pandemic is now greater than ever before. The global epidemic of severe acute respiratory syndrome (SARS) in 2003, and concerns that terrorists may use an infectious agent such as smallpox as a weapon, had previously demonstrated the need to determine control strategies for exotic infections. The prior determination of such strategies, and the use of mathematical models to assist this, are hampered by the obvious lack of data. We propose an integral equation model of Kermack-McKendrick type that may be used to compare strategies based on the isolation of infectious individuals and targeted vaccination. The model structures the incidence of infection according to the location of exposure, and requires some knowledge of the infectivity kernel and the initial rate of exponential increase of cases. The model’s use in the design of strategies to minimize the risks of pandemic influenza, SARS and smallpox in a previously unexposed community is demonstrated.

Roberts, M. G., Baker, M., Jennings, L. C., Sertsou, G. & Wilson, N. A model for the spread and control of pandemic influenza in an isolated geographical region. Journal of the Royal Society Interface In press.
Aldis, G. K. & Roberts, M. G. 2005 An integral equation model for the control of a smallpox outbreak. Mathematical Biosciences 195, 1-22.
Roberts, M. G. 2004 Modelling strategies for minimizing the impact of an imported exotic infection. Proceedings of the Royal Society, Series B 271, 2411-2415.

Vaccination strategies for the control of infectious diseases
A model, developed for the Ministry of Health in 1996, successfully predicted the 1997 measles epidemic. A control policy, initiated in response to the model, swiftly contained it. It was determined that further epidemics would occur unless the timing of the second measles vaccination was changed from 11 years to 5 years of age. The vaccination schedule was changed for 2001. A similar model was used to investigate the vaccination strategies for pertussis.
Roberts, M. G. & Tobias, M. I. 2000 Predicting and preventing measles epidemics in New Zealand: Application of a mathematical model. Epidemiology and Infection 124, 279-287.

The type reproduction number and the characterisation of reservoirs of infection
We propose a new threshold quantity for the analysis of the epidemiology of infectious diseases. The quantity is similar in concept to the familiar basic reproduction number, R0, but it singles out particular host types instead of providing a criterion that is uniform for all host types. Using this methodology we are able to identify the long-term effects of disease-control strategies for particular subgroups of the population, to estimate the level of control necessary when targeting control effort at a subset of host types, and to identify host types that constitute a reservoir of infection. These insights cannot be obtained by using R0 alone.

Roberts, M. G. & Heesterbeek, J. A. P. 2003 A new method for estimating the effort required to control infectious diseases. Proceedings of the Royal Society, Series B 270, 1359-1364.

Strategies for the control of bovine tuberculosis in possums
A model was developed that described the epidemiology and control of tuberculosis in possum populations. The model was extended to reflect the spatial interaction of possums and Tb on interconnected habitat patches, and new threshold quantities for diseases in metapopulations were derived. A similar model was employed to describe the dynamics and control of brucellosis in American bison herds. A model, motivated by the search for a possum biocontrol agent, was used to describe the interaction of multiple parasite species in wild animal populations, including threshold quantities determining when a species can invade or when an existing species is ejected from a parasite community.

Fulford, G. R., Roberts, M. G. & Heesterbeek, J. A. P. 2002 The metapopulation dynamics of an infectious disease: tuberculosis in possums. Theoretical Population Biology 61, 15-29.
Kao, R. R. & Roberts, M. G. 1999 A comparison of wildlife control and cattle vaccination as methods for the control of bovine tuberculosis. Epidemiology and Infection 122, 505-519.
Kao, R. R., Roberts, M. G. & Ryan, T. J. 1997 A model of bovine tuberculosis control in domesticated cattle herds. Proceedings of the Royal Society, Series B 264, 1069-1076.
Roberts, M. G. 1999 A Kermack-McKendrick model applied to an infectious disease in a natural population. IMA Journal of Mathematics Applied in Medicine and Biology 16, 319-332.
Roberts, M. G. & Kao, R. R. 1998 The dynamics of an infectious disease in a population with birth pulses. Mathematical Biosciences 149, 23-36.

Allocation of resources for the management of diabetes, physical inactivity and obesity
Multistate life-tables may be used to describe the age-dependent epidemiology of chronic conditions in a population. The flows between the states are described by systems of differential equations known as IPRM models, because they may be solved to express the incidence of a condition in terms of prevalences, remission rates and relative risks of mortality, all of which may be functions of age and obtained from population surveys. These models may be used to predict the effects of proposed interventions on measures of population health. They have been applied to a range of conditions of importance in New Zealand: diabetes, physical inactivity and obesity; and used to explore the interaction between diabetes and obesity.

Roberts, M. G. & Tobias, M. I. 2001 The use of multistate life-table models for improving population health. IMA Journal of Mathematics Applied in Medicine and Biology 18, 119-130.
Tobias, M. I. & Roberts, M. G. 2001 Modelling physical activity: a multi-state life-table approach. Australian and New Zealand Journal of Public Health 25, 141-148.

The sustainable control of intestinal nematode parasites of sheep and the management of anthelmintic resistance
A model, devised in collaboration with the Zoology Department, University of Cambridge, explained the relative influences of climate and acquired immunity on the observed temporal patterns of the parasite population. A new threshold quantity (Q0) for helminth population dynamics was devised. An analysis of the discrete/continuous dynamical system that describes parasite epidemiology found it to exhibit chaotic behaviour in some regions of parameter space. The methodology was used to describe the epidemiology of sheep nematodes in New Zealand, Australia and the United Kingdom, and extended to provide criteria for the development of anthelmintic resistance under different control strategies.

Heesterbeek, J. A. P. & Roberts, M. G. 1995 Threshold quantities for helminth infections. Journal of Mathematical Biology 33, 415-434.
Kao, R. R., Leathwick, D. M., Roberts, M. G. & Sutherland, I. A. 2000 Nematode parasites of sheep: a survey of epidemiological parameters and their application in a simple model. Parasitology 121, 85-103.

Roberts, M. G. 1999 The immunoepidemiology of nematode parasites of farmed animals: A mathematical approach. Parasitology Today 15, 246-251.
Roberts, M. G. & Heesterbeek, J. A. P. 1995 The dynamics of nematode infections of farmed ruminants. Parasitology 110, 493-502.
Roberts, M. G. & Heesterbeek, J. A. P. 1998 A simple parasite model with complicated dynamics. Journal of Mathematical Biology 37, 272-290.