Networks of infection: Online respondent-driven detection for studying infectious disease transmission and case finding

A broad range of infectious diseases such as influenza, measles and Ebola are transmitted through direct or close human contact. These pathogens do not spread randomly through a population, but follow the structure of human contact networks. Making use of these networks may help to understand and control the spread of infectious diseases. In this thesis, we introduce online respondent-driven detection (RDD). With RDD, individuals are asked to recruit contact persons from their network via the Internet. Contact persons are then asked to do the same, resulting in successive waves of contact persons as respondents. The aims of this thesis were to investigate the feasibility of using online RDD (I) to study infectious disease transmission within contact networks and (II) to enhance case finding during outbreaks of emerging or re-emerging pathogens. We conducted empirical studies in Thailand and the Netherlands to pilot the use of online RDD for sampling contacts of contacts of individuals in the general population. Participants were asked to record their numbers of contact persons at different settings and self-reported influenza-like-illness symptoms, and to invite four persons whom they had met in the preceding weeks. In the first part of thesis, we investigated mixing patterns by analysing correlations between recruiters and their recruitees. Such information is needed for mathematical models that provide evidence-based support for outbreak preparedness and interventions. We found that participants had a tendency to invite contact persons with similar socio-demographic characteristics, but they did not preferably invite persons who also reported similar numbers of contact persons. In the second part of this thesis, we analysed self-reported symptoms of individuals linked in recruitment chains and observed that symptomatic participants recruited other symptomatic participants. This suggests that RDD can enhance identification of hitherto ‘hidden’ cases that go unnoticed by traditional surveillance systems, especially for infectious disease outbreaks in which the majority of symptomatic patients do not seek health care. Online RDD also has large geographical coverage and can provide timely, detailed information about individuals and their contacts. However, RDD has important limitations concerning peer recruitment. During all our studies, sample sizes of recruitment remained small. To increase our understanding of factors important for the success of the recruitment process, we developed a stochastic simulation model where parameters were suggested by our empirically collected data. This model is useful for preparing future RDD surveys, e.g., by providing input on the required mean number of successfully sent invitations to reach large recruitment trees. In practice, however, it remains challenging to motivate participants to invite others. More empirical research is needed to understand the reasons why people choose to invite others, why they choose to invite specific contact persons, which recruitment options are most convenient for them and why or why not their contact persons decide to participate. If peer recruitment can be further increased, our research shows that online RDD has great potential for enhancing case finding and improving epidemiological knowledge on the transmission of pathogens within contact networks.