Social sensing of health and wellbeing impacts from pollen and air pollution

Air pollution and airborne pollen can have significant impacts on health and wellbeing. Health effects can be severe; for example, people with hayfever may also suffer from asthma, with complications sometimes leading to hospitalisation. However, most cases - while unpleasant and disruptive to the affected individual - are not severe enough to require formal medical treatment. These milder cases are not currently recorded or identified. This masks the true extent of these health conditions and their impacts on quality of life, making them hard to study and to manage.

Meanwhile, the digital revolution is creating huge datasets that can provide rich information. We have successfully pioneered a "social sensing" methodology that uses social media data to detect and locate environmental hazards. Social sensing involves several stages: data harvesting, data cleaning (e.g. filtering for relevance and removing spam accounts), event detection, and data visualisation. One of the most exciting opportunities for social sensing is provision of real-time information, allowing accurate monitoring and early warning of emerging hazards.

Social sensing has great potential for tracking pollen, air pollution and associated health impacts such as asthma and hayfever. This new source of data can fill an important information gap and bring benefits to public sector organisations and charities working to improve public health. Outputs can be easily provided to the general public through the web, to help asthma and hayfever sufferers manage their conditions more effectively. However, social sensing also raises ethical issues. Collection and analysis of social media data without user consent may be seen as a breach of privacy, especially where the data concern a sensitive topic such as health and/or wellbeing. Furthermore, since social media users are not necessarily representative of wider society, it is possible that decisions or policies based on social media data will unfairly benefit some sections of society at the expense of others.

This project at University of Exeter will evaluate a social sensing prototype focused on pollen, air pollution, asthma and hayfever. The prototype will be assessed for use by several partner organisations working on environment and public health issues: Met Office, Public Health England and AsthmaUK. An integrated ethical investigation will directly address the privacy and fairness concerns raised by social sensing.

The project has three aims: (1) Create a prototype social sensing platform to track health and wellbeing impacts of pollen and air pollution; (2) Work with partners to evaluate the usefulness of social sensing in a variety of real-world scenarios; and (3) Investigate ethical concerns around social sensing, in particular, fairness and privacy. These aims will be delivered by achieving four specific objectives: (1) Engage end-users and stakeholders to co-design project goals; (2) Develop a prototype social sensing tool focused on pollen, air pollution, asthma and hayfever; (3) Evaluate the prototype in multiple real-world scenarios identified by partner organisations; and (4) Use academic literature, public engagement and surveys to assess ethical concerns around privacy and fairness.

The project team brings together an interdisciplinary mix of academics from University of Exeter, policy partners from Public Health England (PHE), Met Office and AsthmaUK, and members of the public from the Health and Environment Public Engagement (HEPE) group. Expertise in computer science, environment and human health will be combined to solve real-world problems. Outputs will include social sensing software (free for anyone to use) and a comprehensive Case Study Report summarising all project findings. Overall this project will create an essential evidence base to guide future use of social sensing in the context of environment and public health.

The "social sensing" methodology demonstrated in this project has potential benefits for policy in many sectors and organisations, arising from better provision of information to support policy development and operational decision-making.

Public sector organisations concerned with public health and environmental forecasting will benefit from real-time data about several poorly measured factors: pollen, asthma, hayfever. Our project has been co-created with PHE and Met Office and they will evaluate the social sensing prototype within their organisations.
The UK Met Office and equivalent international bodies will benefit from better monitoring of pollen levels, at lower cost and higher spatial/temporal resolution than the existing pollen observing network. This will enable them to provide better public information and warnings, to improve hazard models, and to offer improved pollen forecasts.
Public Health England (PHE), and equivalent bodies in Wales, Scotland and Northern Ireland, will benefit from improved reporting on incidence of asthma and hayfever, which affect a large fraction of the UK population. Asthma and hayfever cases are rarely recorded, since most cases do not present for medical treatment.
Real-time information will allow healthcare practitioners and managers to predict and manage demand for healthcare services, and over longer time periods may improve organisational policies for dealing with asthma, hayfever and related health complications.
Public bodies with an interest in air quality, including Defra, PHE, Met Office and Local Authorities, will have better evidence for public concern and awareness, helping to support policy-making and legislation.

Third sector organisations with a focus on asthma and hayfever (e.g. AsthmaUK, AllergyUK, British Lung Foundation) will benefit from real-time mapping of asthma and hayfever cases, which are currently poorly recorded, and pollen, which is a common trigger for both conditions. This will enable them to offer better information to affected populations and to compile better datasets to support research into interactions between the environment and human health, speaking directly to their strategic missions. Our project has been co-created by AsthmaUK and they will evaluate the social sensing prototype for use by asthma sufferers.

Businesses and the UK economy will benefit from improved knowledge of pollen levels. Manufacturers and retailers of hayfever medication (with total over-the-counter sales of around £107.2m in 2015) will benefit from real-time mapping of pollen levels and hayfever incidence, allowing better planning and supply chain management. Employers will benefit by a reduction in the loss of staff productivity associated with hayfever if staff members can avoid exposure and more effectively manage their condition.

Sufferers from asthma (around 12% of UK population) and hayfever (around 20% of UK population) will benefit from better information allowing them to avoid exposure and manage their conditions more effectively. This will improve general health and wellbeing, as well as avoiding the loss of economic productivity and increased demand for healthcare services that are associated with these conditions. Our project will involve the public through the Health and Environment Public Engagement group and will target asthma sufferers directly through AsthmaUK.

The ethical investigation performed in this project will examine the potential privacy and fairness concerns surrounding the use of social media data as a source of information. Findings from this study will inform the wider public debate in this area and could have long term impacts on policy-making at many levels in UK society.