Colonization and transmission of Staphylococcus aureus in schools: a citizen science project

Abstract

Aggregation of children in schools has been established to be a key driver of transmission of infectious diseases. Mathematical models of transmission used to predict the impact of control measures, such as vaccination and testing, commonly depend on self-reported contact data. However, the link between self-reported social contacts and pathogen transmission has not been well described. To address this, we used Staphylococcus aureus as a model organism to track transmission within two secondary schools in England and test for associations between self-reported social contacts, test positivity and the bacterial strain collected from the same students. Students filled out a social contact survey and their S. aureus colonization status was ascertained through self-administered swabs from which isolates were sequenced. Isolates from the local community were also sequenced to assess the representativeness of school isolates. A low frequency of genome-linked transmission precluded a formal analysis of links between genomic and social networks, suggesting that S. aureus transmission within schools is too rare to make it a viable tool for this purpose. Whilst we found no evidence that schools are an important route of transmission, increased colonization rates found within schools imply that school-age children may be an important source of community transmission.

Keywords: MRSA; WGS; schools; social network analysis; transmission.

Fig. 1.

Exemplar network from School 1 (Year 1) illustrating how network structure relates to ST, school year (age of student), gender and whether students report sharing drinks with their friends. Each edge corresponds to a reported contact from a student within the study population. Nodes with no shape/colour represent named contacts within the school who did not participate in the study. Clustering with respect to school year, gender and preference to share drinks is visually evident in contrast to strain type, which demonstrates no clear pattern.

Fig. 2.

Assortativity with respect to key variables in School 1. Assortativity measures the preference of pupils to report contacts with individuals within the same school year, of the same gender, same preference for sharing drinks [which we also stratify by gender: male (M) and female (F)] and positivity for S. aureus (swab 1, swab 2). Lines represent 95 % CIs based on 10 000 bootstrap networks.

Fig. 3.

Population structure of S. aureus isolates. (a) Maximum-likelihood phylogenetic tree of 400 S. aureus isolates from schools and a supplemental data set of 384 isolates from the CARRIAGE study. In silico predicted antimicrobial-resistance determinants are shown for 9 antimicrobials; isolates positive for the determinant are highlighted in red. (b) Distribution of STs per year in School 1 and School 2 and distribution of STs in the CARRIAGE isolates.

Fig. 4.

Integrated social and genomic networks for School 1. (a) Maximum-likelihood phylogenetic tree of 75 isolates collected from the first swab in Year 1. (b) Integrated social and genomic network for Year 1. (c) Maximum-likelihood phylogenetic tree of 108 isolates collected from the first swab in Year 2. (d) Integrated social and genomic network for Year 2. Putative genomic links are coloured in grey; genomic links confirmed by social links are coloured red.