Epidemic dynamics of infectious disease in metropolitan area and its optimal intervention strategy

Identification of high-risk community is essential for a policymaking of efficient and effective disease intervention strategies. Under the limited resources for countermeasures (e.g. stocks of vaccine and anti-viral drugs) and need to minimize the economic effect of quarantine (e.g. shut down of factories and offices), it is inevitable to apply intervention strategies from higher risk areas, in order to prevent a disease invasion and to reduce the prevalence in case of pandemic. In this study, we show that the sensitivity analysis of basic reproductive ratio R0 is useful for determining the degree of risk for each location within a metropolitan area. Here we have used the actual commuter flow of Tokyo metropolitan area as an example (commuting data is acquired from the Urban Transportation Census).The model of contagious disease spread in the Tokyo metropolitan area is formulated as a meta-population model, where each node corresponds to each local station (about 1,500 stations) and recurrent movements of commuters (about 150,000 individuals in census sample) interconnect the nodes. For an infection dynamics we assume influenza like diseases and set epidemiological state of each individual to one of Susceptible, Infectious and Recovered state (SIR model). Using this model we calculate the basic reproductive ratio Ro using the next generation matrix method.By applying a sensitivity analysis to R0 and calculate the effect of changing the susceptible population structure (i.e. corresponding to vaccination and/or isolation), we are able to rank all residential stations and commuting pathways according to its epidemiological risk. From this we compare two strategies: applying countermeasures using this ranking from the top and randomly regardless of the rank. The results show that this epidemiological risk ranking according to the sensitivity analysis is very effective in preventing the spread of infectious disease.