Social media interventions for precision public health: promises and risks

Abstract Social media data can be used with digital phenotyping tools to profile the attitudes, behaviours, and health outcomes of people. While there are a growing number of examples demonstrating the performance of digital phenotyping tools using social media data, little is known about their capacity to support the delivery of targeted and personalised behaviour change interventions to improve health. Similar tools are already used in marketing and politics, using individual profiling to manipulate purchasing and voting behaviours. The coupling of digital phenotyping tools and behaviour change interventions may play a more positive role in preventive medicine to improve health behaviours, but potential risks and unintended consequences may come from embedding behavioural interventions in social spaces.