The spread of healthy behaviors in social networks through megatrend - a promising strategy for potentiating public health interventions

Abstract

This work aims to design system for describing, analyzing and implementing physical activity (PA) interventions by using information and communication technologies from a complex systems perspective. In the first chapter, i present an observation of PA interventions as a system. In the second chapter, I group PA inteventions by categories and I provide a review of reviews of the reported effects around the globe. I conclude that the effect size of those interventions is, in most of the cases, low or moderate. In the third chapter I propose a model for measuring the potential effect of PA interventions through megatrends, specifically mobile phones and internet. I conclude that there is a potential effect for PA promotion through internet and mobile phones around the globe given that the internet and mobile phone access is high even if the countries income is low. Given the megatrends are behavioral changes observed in population which are driven by social relations, understanding the spread of behaviors through social relations can be crucial for improving the use of megatrends for the spread of healthy behaviors during PA intervention. The social network analysis provides a natural framework for modeling behavioral spreading processes, and allows exploring strategies for designing an intervention in order to transform a social system. For that reason in the fourth chapter I explore models for representing spreading processes in social networks where the spreading drivers are interactions or social pressure. Then I propose a stochastic variation of the deterministic threshold model for representing behavioral spreading processes that could simulate the expected adoption of a behavioral during the implementation of a community-based intervention

A main challenge in the implementation of a public health intervention is the selection of the seed individuals for initializing the intervention. It is necessary to plan a sampling strategy in order to target the majority of the community and potentiate the effectiveness of the intervention at a low cost. For addressing that problem, in chapter five I compare different strategies for identifying seed spreaders in a social networks which are the key actors that have to be targeted for initializing the intervention. I propose two new strategies that are based on detecting key individuals in each of the communities detected in the network and I compare the spreading efficiency of those strategies with an intervention initialized in dividuals that are in center of core of the network. I discuss in which cases one strategy should be considered over the others, depending on the social network topology. Finally, I present a case study of a real field experiment for determining the effect of PA intevention through mobile phones that we developed whit the EpiAndes research group in three schools of Bogotá (Colombia)