eAppendix Hand , Foot and Mouth Disease in China : Patterns of Spread and Transmissibility during 2008-2009

Here we derive a statistical transmission model for the surveillance data of HFMD. For stratum k of prefecture i, let Sk i (t) be the number of susceptible people at the beginning of week t, and Y k i (t) the number of cases infected during week t who were identified by symptom onsets in week t + 1, assuming a one-week incubation period. Let Xi (t) be the vector of covariates associated with stratum k of prefecture i in week t. Define Ai as the collection of prefectures that are adjacent to prefecture i. Let γ0, γ1 and γ2 be the baseline rates of disease transmission from an infectious source to a susceptible person for the three types of sources: the unknown reservoir, an infectious person within the same prefecture, and an infectious person in an adjacent prefecture. The exact interpretation of these parameters depends on how covariates are formulated. For example, γ1 may refer to the transmission rate when the susceptible person and the infectious person are both males in the age group of ≤ 1 year in the same prefecture. Let γkl ij (t) be the effective transmission rate, adjusted for covariates, from an infectious person in stratum l of prefecture j to a susceptible