Five challenges for spatial epidemic models

Riley, Steven, et al. “Five challenges for spatial epidemic models.” Epidemics 10 (2015): 68-71.

There is an increasing awareness of the spatial component in transmission dynamics. Three different mechanistic models are discussed and how they are used in modeling epidemics: individuals-based models, metapopulation models, and network models. In the paper, the authors outline five challenges with incorporating spatial information in theoretical infectious disease dynamics.

1. How can network models best be constructed to reflect spatial population structure?

Previously mentioned mechanistic models do not allow for disjoint sets. We can think of all the models in a network formulation. There still needs network models to be developed to reflect the spatial structure and how then these are not fully connected (disjoint). Also, how these these properties of an epidemic occurring on the network approximate the full spatial dynamics.

2. How should we model contact structure in spatially heterogeneous populations?

Movement of individuals (humans in particular) do not move uniformly in space. Individuals move around daily driven by the distribution of the density of the population around them. Moving forward, spatial models of infectious diseases’ assumption of movement should be compared to the spatially resolved social contact data.

3. How do we define a threshold parameter for spatial models?

The basic reproduction number (R0) is a common quantity determined which represents the average number of secondary infections caused by a single infectious individual in a fully susceptible population. This is often used when determining the threshold condition for an outbreak to occur. However, determining this value in a spatially explicit model is difficult and unclear.

4. How should we analyze models with long distance interactions?

We are often interested in how contact structure and influence the duration of an outbreak. There needs to be methods to investigate the proper scaling for networks based on empirical data.

5. On what scale is intervention most effective?

The scale for transmission, that the data is available for, and intervention are not the same. We should recognize that spatially heterogeneous interventions may influence the transmission between individuals in unintended ways. Models should include the same granularity that is used for intervention strategies.