The Role of Social Contacts on Children's Health: Statistical Inference on Infectious Disease Data

Elisabetta De Cao, Tor Vergata University and Bocconi University
Piero Manfredi, Università di Pisa
Alessia Melegaro, Università Bocconi

The impact of public health interventions depends on how individuals mix. Social contacts are relevant in explaining children's wellbeing, and in particular, the spread of infectious diseases. Many countries lack of data on social mixing patterns, and rely on theoretical assumptions on population mixing to evaluate interventions. The aim of this work is to understand how different social interactions affect close-contact childhood infection processes. We propose a model which, by integrating different data sources (time use data and contact surveys), obtains mixing matrices that describe the social structure and reproduce seroprevalence profiles. We assume that potentially infectious contacts are proportional to self-reported number of social contacts and/or time of exposure in social activities. To evaluate the uncertainty of model outputs, we use the Bayesian Melding approach. We empirically analyze Italian data, where contact survey, time use data from early ages, and data on close-contact childhood infections are available.