Can Twitter sentiment help us design better mass communication interventions for public health?A feasibility study focussing on skin cancer prevention

Research has shown that social networks affect the transmission of healthy and unhealthy behaviours and Social Media (SM) have been shown to transmit moods and feelings However, few studies have, by design, harnessed the power of social networks within bespoke public health "e-interventions".

Whether such social media could be harnessed to support a successful public health intervention would depend on a range of practical and theoretical issues that have yet to be studied. The proposed study will address the following knowledge gaps to support the feasibility assessment of a Social Media-enabled intervention to promote "Care in the Sun", for the prevention of skin cancer, including: (i) would SM constitute an acceptable means for delivering such interventions in the population?; (ii) would people be willing to share across a SM platform like Twitter, personal issues (e.g. health behaviour/attitudes)?; (iii) what types of SM communication and health issues would attract the attention of the target population, e.g. would the use of certain types of visual images be acceptable/effective?; (iv) what behaviour change theories might plausibly underpin the intervention design. Twitter and Facebook are quite distinct SM platforms, the latter offering more potential to identify a user's characteristics and their expressed friendship networks.

We will develop a SM-enabled mass communication campaign designed to change attitudes and behaviours of adults related to reducing harmful Ultra Violet light exposure. This will be delivered using Twitter and in tandem we will explore the potential of Facebook for similar purposes.

The outcomes of this feasibility study will include: (i) changes in reported attitudes and behaviours related to UV exposure and skin cancer obtained from "before" and "after" population surveys; (ii) pre and post intervention usage of Twitter, social network architecture of Twitter messages, analyses of sentiments expressed within tweets towards the skin cancer prevention SM campaign and patterns of sentiment diffusion across the social network.

The study will help us learn lessons for using social media like Twitter in future public health campaigns.

Previous "care in the sun" campaigns have had mixed success but most pre-date the social media (SM) era. Social media-enabled campaigns have the advantage of targeted messaging, at low cost and large reach but we know little about what in practice would determine their feasibility. Aim: To test the feasibility of designing, implementing and evaluating a bespoke SM-enabled intervention for skin cancer prevention. Stage 1: To undertake a review of public health guidance and interventions for skin cancer prevention to ascertain effective intervention components. Focus groups with stakeholders (n=3) and users (n=3) will consult on key design and implementation aspects. Following this a co-design workshop will lead to the development of a logic model. Stage 2: To develop, implement and evaluate the feasibility of a SM-enabled campaign using a quasi-experimental, interrupted time series with comparison study. A "cross-over" design permitting 30-days with one message "frame", an 8-week "washout" period, and a further 30-days of the SM campaign with a different message frame will be used. Social media responses will be tracked in two geographical areas within a subset of tweets (and Facebook posts), one area exposed to the NI "Care in the Sun" campaign, and Wales (which will not receive the specific SM campaign). Stage 3: A final workshop will identify lessons learned about how future SM campaigns and their evaluation might be mounted. Outcomes: (i) changes in reported attitudes and behaviours related to UV exposure and skin cancer obtained from "before" and "after" population surveys (n=1000); (ii) pre and post intervention usage of Twitter, social network topology of Twitter messages, analyses of sentiments expressed within tweets towards the skin cancer prevention SM campaign and patterns of sentiment diffusion across the social network architecture. If deemed feasible, the next stage will be to undertake a Phase II pilot trial.