Social and spatial network analysis is an important approach for investigating infectious disease transmission, especially for pathogens transmitted directly between individuals or via environmental reservoirs. Given the diversity of ways to construct networks, however, it remains unclear how well networks constructed from different data types effectively capture transmission potential. We used empirical networks from a population in rural Madagascar to compare social network survey and spatial data-based networks of the same individuals. Close contact and environmental pathogen transmission pathways were modelled with the spatial data. We found that naming social partners during the surveys predicted higher close-contact rates and the proportion of environmental overlap on the spatial data-based networks. The spatial networks captured many strong and weak connections that were missed using social network surveys alone. Across networks, we found weak correlations among centrality measures (a proxy for superspreading potential). We conclude that social network surveys provide important scaffolding for understanding disease transmission pathways but miss contact-specific heterogeneities revealed by spatial data. Our analyses also highlight that the superspreading potential of individuals may vary across transmission modes. We provide detailed methods to construct networks for close-contact transmission pathogens when not all individuals simultaneously wear GPS trackers.

BackgroundSuperspreading infections play an important role in the SARS-CoV-2 pandemic. Superspreading is caused primarily by heterogeneity in social contact rates, and therefore represents an opportunity for targeting surveillance and control via consideration of social network topologies, particularly in resource-limited settings. Yet, it remains unclear how to implement such surveillance and control, espeically when network data is unavailable.MethodsWe evaluated the efficiency of a testing strategy that targeted potential superspreading individuals based on their degree centrality on a social network compared to a random testing strategy in the context of low testing capacity. We simulated SARS-CoV-2 dynamics on two contact networks from rural Madagascar and measured the epidemic duration, infection burden, and tests needed to end the epidemics. In addition, we examined the robustness of this approach when individuals’ true degree centralities were unknown and were instead estimated via readily-available socio-demographic variables.FindingsTargeted testing of potential superspreaders reduced the infection burden by 40-63% at low testing capacities, while requiring between 45-78% fewer tests compared to random testing. Further, targeted testing remained more efficient when the true network topology was unknown and prioritization was based on socio-demographic characteristics.InterpretationIncorporating social network topology into epidemic control strategies is an effective public health strategy for health systems suffering from low testing capacity and can be implemented via socio-demographic proxies when social networks are unknown.FundingThis research was funded by the Agence Nationale de la Recherche, a NIH-SSF-NIFA Ecology and Evolution of Infectious Disease Award (No. 1R01-TW011493-01), and a Duke University Provost’s Collaboratory grant.

Background Targeted surveillance allows public health authorities to implement testing and isolation strategies when diagnostic resources are limited, and can be implemented via the consideration of social network topologies. Yet, it remains unclear how to implement such surveillance and control when network data are unavailable. Methods We evaluated the ability of socio-demographic proxies of degree centrality to guide prioritized testing of infected individuals compared to known degree centrality. Proxies were estimated via readily-available socio-demographic variables (age, gender, marital status, educational attainment, and household size). We simulated SARS-CoV-2 epidemics via a SEIR individual-based model on two contact networks from rural Madagascar to further test the applicability of these findings to low-resource contexts. Results Targeted testing using socio-demographic proxies performed similarly to targeted testing using known degree centralities. At a low testing capacity, using the proxies reduced the infection burden by 22-33% while using 20% fewer tests, compared to random testing. By comparison, using known degree centrality reduced the infection burden by 31-44% while using 26-29% fewer tests. Conclusions We demonstrate that incorporating social network information into epidemic control strategies is an effective countermeasure to low testing capacity and can be implemented via socio-demographic proxies when social network data are unavailable.