Contacts in context: large-scale setting-specific social mixing matrices from the BBC Pandemic project

Petra Klepac; Adam J Kucharski; Andrew JK Conlan; Stephen Kissler; Maria Tang; Hannah Fry; Julia R Gog

As the contact matrices are related to next generation matrices, the differences in structure between the POLYMOD and BBC matrices have direct consequences for disease dynamics. Here, we are particularly interested in the age-groups that are most responsible for transmission, which is described by the dominant eigenvector of the next generation matrix. We compare POLYMOD matrices for Great Britain consisting of the average number of contacts recorded per day per survey participant that are available from Mossong et al. [22] Table S5 . Even though the raw data from POLYMOD is available, these matrices are most commonly used, which is why we choose them for the comparison with BBC matrices. Ages are grouped in 5-year bins until the age of 70, and a single 70+ age group (total of 15 age-groups). For the comparison of matrices we follow the same grouping (Fig 7) , and make both BBC and POLYMOD matrices reciprocal using the 2018 population vector for England and Wales [1] .