Jennifer Badham

The impact of network clustering and assortativity on epidemic behaviour

Epidemic models have successfully included many aspects of the complex contact structure apparent in real-world populations. However, it is difficult to accommodate variations in the number of contacts, clustering coefficient and assortativity. Investigations of the relationship between these properties and epidemic behaviour have led to inconsistent conclusions and have not accounted for their interrelationship. In this study, simulation is used to estimate the impact of social network structure on the probability of an SIR (susceptible-infective-removed) epidemic occurring and, if it does, the final size. Increases in assortativity and clustering coefficient are associated with smaller epidemics and the impact is cumulative. Derived values of the basic reproduction ratio (R(0)) over networks with the highest property values are more than 20% lower than those derived from simulations with zero values of these network properties.

Parameterisation of Keeling's network generation algorithm

Simulation is increasingly being used to examine epidemic behaviour and assess potential management options. The utility of the simulations rely on the ability to replicate those aspects of the social structure that are relevant to epidemic transmission. One approach is to generate networks with desired social properties. Recent research by Keeling and his colleagues has generated simulated networks with a range of properties, and examined the impact of these properties on epidemic processes occurring over the network. However, published work has included only limited analysis of the algorithm itself and the way in which the network properties are related to the algorithm parameters. This paper identifies some relationships between the algorithm parameters and selected network properties (mean degree, degree variation, clustering coefficient and assortativity). Our approach enables users of the algorithm to efficiently generate a network with given properties, thereby allowing realistic social networks to be used as the basis of epidemic simulations. Alternatively, the algorithm could be used to generate social networks with a range of property values, enabling analysis of the impact of these properties on epidemic behaviour.

Commentary: Measuring the shape of degree distributions

Degree distribution is a fundamental property of networks. While mean degree provides a standard measure of scale, there are several commonly used shape measures. Widespread use of a single shape measure would enable comparisons between networks and facilitate investigations about the relationship between degree distribution properties and other network features. This paper describes five candidate measures of heterogeneity and recommends the Gini coefficient. It has theoretical advantages over many of the previously proposed measures, is meaningful for the broad range of distribution shapes seen in different types of networks, and has several accessible interpretations. While this paper focusses on degree, the distribution of other node based network properties could also be described with Gini coefficients.

The Extortion Relationship: A Computational Analysis

Systematic extortion involves a long term parasitic relationship between the criminal and the victim. Game theory analysis has provided insight into the choices of individual hypothetical criminal and victim pairs. In this paper we present an agent-based model so as to extend the analysis to the relationship between extorters and other potential victims. The model is developed in two stages, the first to be closest to game theory, the second one making the decision informed by the social environment of the victim. The agent-based model shows the importance of social aspects for the functioning of extortion rackets.

Network methods to support user involvement in qualitative data analyses: An introduction to Participatory Theme Elicitation

BackgroundWhile Patient and Public Involvement (PPI) is encouraged throughout the research process, engagement is typically limited to intervention design and post-analysis stages. There are few approaches to participatory data analyses within complex health interventions.MethodsUsing qualitative data from a feasibility randomised controlled trial (RCT), this proof-of-concept study tests the value of a new approach to participatory data analysis called Participatory Theme Elicitation (PTE). Forty excerpts were given to eight members of a youth advisory PPI panel to sort into piles based on their perception of related thematic content. Using algorithms to detect communities in networks, excerpts were then assigned to a thematic cluster that combined the panel members’ perspectives. Network analysis techniques were also used to identify key excerpts in each grouping that were then further explored qualitatively.ResultsWhile PTE analysis was, for the most part, consistent with the researcher-led analysis, young people also identified new emerging thematic content.ConclusionsPTE appears promising for encouraging user led identification of themes arising from qualitative data collected during complex interventions. Further work is required to validate and extend this method.Trial registrationClinicalTrials.gov, ID: NCT02455986. Retrospectively Registered on 21 May 2015.

Calibrating with Multiple Criteria: A Demonstration of Dominance

Pattern oriented modelling (POM) is an approach to calibration or validation that assesses a model using multiple weak patterns. We extend the concept of POM, using dominance to objectively identify the best parameter candidates. The TELL ME agent-based model is used to demonstrate the approach. This model simulates personal decisions to adopt protective behaviour during an influenza epidemic. The model fit is assessed by the size and timing of maximum behaviour adoption, as well as the more usual criterion of minimising mean squared error between actual and estimated behaviour. The rigorous approach to calibration supported explicit trading off between these criteria, and ultimately demonstrated that there were significant flaws in the model structure.

Uses of agent-based modeling for health communication: The TELL ME case study

ABSTRACT Government communication is an important management tool during a public health crisis, but understanding its impact is difficult. Strategies may be adjusted in reaction to developments on the ground and it is challenging to evaluate the impact of communication separately from other crisis management activities. Agent-based modeling is a well-established research tool in social science to respond to similar challenges. However, there have been few such models in public health. We use the example of the TELL ME agent-based model to consider ways in which a non-predictive policy model can assist policy makers. This model concerns individuals’ protective behaviors in response to an epidemic, and the communication that influences such behavior. Drawing on findings from stakeholder workshops and the results of the model itself, we suggest such a model can be useful: (i) as a teaching tool, (ii) to test theory, and (iii) to inform data collection. We also plot a path for development of similar models that could assist with communication planning for epidemics.

A standing ovation for Nigel: An informal study

This article analyses a series of emails thanking Nigel for his stewardship of JASSS and the characteristics of their authors. It identifies a correlation between two measures of author activity in social simulation research, but no pattern between these activity measures and the email timing. Instead, the sequence suggests a classic standing ovation effect.

Simulating network intervention strategies: Implications for adoption of behaviour

This study uses simulation over real and artificial networks to compare the eventual adoption outcomes of network interventions, operationalized as idealized contagion processes with different sets of seeds. While the performance depends on the details of both the network and behaviour adoption mechanisms, interventions with seeds that are central to the network are more effective than random selection in the majority of simulations, with faster or more complete adoption throughout the network. These results provide additional theoretical justification for utilizing relevant network information in the design of public health behavior interventions.

Developing agent-based models of complex health behaviour.

&NA; Managing non‐communicable diseases requires policy makers to adopt a whole systems perspective that adequately represents the complex causal architecture of human behaviour. Agent‐based modelling is a computational method to understand the behaviour of complex systems by simulating the actions of entities within the system, including the way these individuals influence and are influenced by their physical and social environment. The potential benefits of this method have led to several calls for greater use in public health research. We discuss three challenges facing potential modellers: model specification, obtaining required data, and developing good practices. We also present steps to assist researchers to meet these challenges and implement their agent‐based model. HighlightsABM is effective for modelling health behaviour embedded within the environment.Key challenges are: formulating rules, obtaining process data, and acquiring skills.Experience from other disciplines may be adapted for public health research.The potential benefits of ABM warrant the effort required.