Polio Vaccine Hesitancy in the Networks and Neighborhoods of Malegaon, India

This study was funded by the World Health Organization (WHO), one of the partners in the Global Polio Eradication Initiative (GPEI), in a quest to better understand vaccine-refusing behaviour as they strive to eradicate polio worldwide. To learn about social network factors associated with polio vaccine hesitancy, the researchers investigated social and spatial clustering of households by their vaccine acceptance status in Malegaon, India, an area known for vaccine refusal and repeated detection of polio cases. The goal was to investigate whether the resistance of households to having their young children vaccinated against polio might be related to similar resistance in households to which they are socially connected. The researchers also endeavoured to distinguish between network and neighborhood effects; in short, does your behaviour with respect to polio vaccines depend on whom you know, where you live, or both?

The researchers carried out a population-based study of 2,462 households in 25 contiguous high-risk (by polio planning definitions) neighbourhoods in Malegaon Municipal Corporation in the Nashik district of Maharashtra state in the western region of India. In July 2012, 25 teams collected data through in-person household surveys. Oral polio vaccine (OPV) vaccination status data were available from WHO, which had concluded a supplementary polio immunisation campaign just prior to the study. As part of this campaign, the organisation classified a household as: vaccine-accepting if it accepted the vaccine for eligible children in the household; vaccine-reluctant if it accepted the vaccine on the second attempt; and vaccine-refusing if it continued to refuse to have their vaccine-eligible children vaccinated even after the second attempt. Also, household heads were asked to identify up to 4 contacts with whom they discuss general issues (e.g., business, sports, personal matters, or issues that affect their community) and up to 4 contacts with whom they discuss health-related issues.

Drawing on a series of network concepts, which are defined and briefly described in the paper, the researchers explain how they constructed the polio vaccine network. They clarify how they constructed a simple statistical model to predict the existence of a tie in the vaccine network between 2 households based on the observed household attributes in order to detect which attributes are most highly predictive of ties.

Among participating households, 1,074 accepted the vaccine, 137 were reluctant, and 144 refused the vaccine. In brief, the researchers find that vaccine-refusing households had fewer outgoing ties than vaccine-accepting households and that they had fewer health ties than vaccine-accepting households. Moreover, the surveyed neighbourhoods varied significantly by their vaccine status composition, and, on average, 49% of network nominations were to households in the same neighbourhood. This suggests that social ties are spatially localised. Finally, vaccine-refusing households nominated on average 93% more vaccine-refusing households in the largest connected component (LCC) of the vaccine network than vaccine-accepting households did, meaning that vaccine refusers clustered in the social network and not just within neighbourhoods. Thus, there was both social and geographic clustering (similar to what has been found regarding the clustering of latrine use in India).

The researchers discuss limitations of the study, such as the fact that, if longitudinal data were available, it would be possible to use a model-based approach to get a sense of the extent to which each mechanism might be driving the behaviour. Furthermore, while networks have the potential to propagate behaviour change, convincing vaccine-refusing households of the benefits of the polio vaccine is likely to remain a challenge. However, they contend, "identification of these clusters is an important step in that it enables better targeting of subsequent interventions and, furthermore, establishes the relevant units of study, the clusters, for future investigations into different vaccine-adoption strategies."

They continue: "Moreover, taking into account people's social network position can enhance the effectiveness and efficiency of public health messaging and interventions: the effectiveness because people are more likely to change their perceptions and behavior if their friends do, and the efficiency because perceptions and behavioral changes can spread out from the primary targets of interventions, creating spillovers that may benefit whole population."