A Novel Data Science Approach to Personas' Creation to Study Willingness to Receive Vaccination in the General Population

Vaccines have proved to be effective in reducing mortality of the COVID-19 pandemic. However, a part of the population is still reluctant to be vaccinated. Thus, the aim of this work was to apply a framework to create Personas, fictional representations of real people, to assess the characteristics of the population willing to be vaccinated in order to develop personalized eHealth-based interventions to increase compliance to vaccinations. Data was collected through an online survey at the beginning of 2021. Multiple dimensionality reduction methods were used as input for K-Medoids clustering with PAM algorithm to create Personas. The optimal number of Personas and dimensionality reduction method to be used were evaluated through the average silhouette graph and the percentage of statistically different attributes between Personas. From 1070 respondents, three Personas were identified. Persona 3 showed statistically significant lower trust in institutions, lower education and lower willingness of being vaccinated when compared to the other two Personas. The developed approach to create Personas was deemed able to identify the main characteristics of those more prone not willing to be vaccinated, suggesting that behavioral change techniques should focus on taking advantage of the closer social circle of those reluctant to vaccines.