Mobile phone

A mobile phone app developed by Cambridge researchers that tracks how people behave during an epidemic could be used to limit disease spread.

Data collected over a few months demonstrated the enormous potential of mobile phones as scientific instruments with which to measure the social activity of a population in real-time.

Influenza is sweeping the UK, schools are closing, workplaces are short-staffed, and the race for a vaccine has begun: all the elements of a scenario that is not so very hard to imagine.

Localised epidemics and global pandemics periodically pose a significant national and global threat. When this happens, it becomes vital for public health and government authorities to track how far and how fast the disease is spreading.

But, to minimise the health, social and economic impact as far as possible, it’s also important to know how people are likely to change their behaviour during the course of an epidemic.

“How people behave could limit or exacerbate their risk of infection,” explained Professor Jon Crowcroft. “Patterns of social interaction that worsen the spread of disease pose a significant risk. On the other hand, if people stay at home rather than work, the cost to the economy may be greater than the cost incurred through actual illness.”

Understanding how people change their activities during an epidemic might not only help authorities to devise strategies to reduce the scale or length of the crisis but also help them to tailor healthcare messages effectively.

Meet, connect, communicate

The FluPhone study at the Computer Laboratory, led by co-principal investigators Professor Crowcroft and Dr Eiko Yoneki, has developed a new tool to study behaviour and its consequences using mobile phones.

Dr Yoneki explained: “The application in the mobile phone monitors influenza-like symptoms by prompting questions for the mobile phone owner. It also captures physical proximity information between individuals by recording other devices nearby via Bluetooth communication.”

In a pilot study, volunteers (mainly University of Cambridge employees and students) downloaded a Java-based application to their mobile phones using an intuitive interface. Ethical issues were handled carefully during participant registration to the FluPhone study, and the study did not record the location information using GPS because of ethical considerations.

“A post-facto analysis of these data will yield valuable insight into how human communities are formed, how much time people spend together, and how frequently they meet. Such data show complex network-like structures, which is very useful for understanding the spread of diseases,” said Dr Yoneki.

Data collected over a few months – ironically coinciding with an outbreak of swine flu – demonstrated the enormous potential of mobile phones as scientific instruments with which to measure the social activity of a population in real time. “There are more cell phones than people,” said Professor Crowcroft, “and, in most urban areas, network coverage is close to 100%, hence we can get very accurate measurement and sampling of the population.”

In fact, the study highlighted how mobile phones can provide data that would be otherwise unavailable, as Professor Crowcroft explained: “In this particular outbreak, it’s now known that some people carried the disease yet were asymptomatic. Our system is capable of identifying these asymptomatic ‘superspreaders’ because they show up by virtue of the contacts who develop the disease.”

To develop the model further, a virtual-disease epidemic application has been prototyped. Dr Yoneki explained: “A specific disease infection model can be programmed, and the fake ‘pathogens’ can be transmitted via Bluetooth radio communication when two individuals are in proximity range.”

“This has proved to be a fantastic tool – you can run a ‘what if’ experiment on the live population based on their contacts, simply by randomly choosing some of the mobile phones to be infectious,” said Professor Crowcroft. “We can then model the effect of behaviour on disease spread.”

Given that technology such as this raises ethical issues concerned with privacy, and approval from the relevant ethics committees was required before the project commenced, the team are now developing a series of guidelines. They hope that this resource will smooth future deployment of a communication protocol via a mobile phone app, particularly in the event of a major disease outbreak.

Spreading information

FluPhone is part of a wider project involving seven academic institutions and government agencies led by the University of Liverpool and funded by the Economic and Social Sciences Research Council. The Cambridge team is planning a larger project using the FluPhone technology together with one of the project partners, the London School of Hygiene and Tropical Medicine, to work in several African countries at village population level in Malawi, Kenya, Tanzania, Uganda, Gambia, Ethiopia and Ghana.

Dr Yoneki also leads a new five-year project funded by the Engineering and Physical Sciences Research Council to extend and develop the analysis and modelling approaches used in FluPhone. “Specific individuals can be identified who act as coalescing hubs at different points in space and time and who influence data flow. I want to investigate these spatial and social clusters to determine what impact they have on the spread of viruses.”

Together, the multi-institution network of epidemiologists, psychologists, economists and computer scientists is developing new ideas about the different ways individuals and organisations might respond to an outbreak of infectious disease.

These include how people might behave when they or their dependants become ill, why they might avoid populated areas or workplaces for fear of infection, how they might respond if their workplace or school was closed, and what might be driving their attitude to risk and infection. Crucially, they are also considering ways that could be used to counteract unhealthy or risky behaviours, or those that threaten economic stability.

The researchers see potential use for the approach in targeting and measuring the effectiveness of public health messages – something that is traditionally very difficult to do because of the time it takes to gather data using traditional methods that involve surveys through primary care providers. Tellingly, their findings indicate that people are more likely to intend to heed health advice if it is underpinned with an understanding of why people may not be able to conduct a recommended behaviour, such as carrying on working during a pandemic.

For more information, please contact Professor Jon Crowcroft (jon.crowcroft@cl.cam.ac.uk) and Dr Eiko Yoneki (eiko.yoneki@cl.cam.ac.uk) at the Computer Laboratory or visit http://www.cl.cam.ac.uk/research/srg/netos/fluphone/


Report on mHealth

The use of mobile communication devices to gather data on how pandemics spread is one of the many innovative applications examined in a wide-ranging study of mobile communications for healthcare (so called mHealth) recently completed by the University.

The report – which is available for download – describes the mHealth applications that have the potential to transform healthcare worldwide, from remote diagnostics and management of long-term conditions to clinical information systems, targeted public health messaging and supply chain management.

The report uses primary research in China to illustrate the applications being developed by the world’s largest mobile telecoms operator, China Mobile (who sponsored the project), together with an extensive literature review of applications in the rest of the world. It also examines the benefits, costs and business cases for deploying mHealth services in developing and developed economies, and outlines the business strategies for telecoms operators.

"In the next phase of deployment, generic service platforms are going to play a crucial role," commented Professor Ian Leslie, lead author of the report. "Until recently, advanced applications required dedicated servers, but the emergence of cloud computing is obviating this need, and the upfront investment required to deploy new applications is coming down. The actors best placed to drive a move toward publicly available platforms are the large network operators."

For more information, please contact project manager Dr Nick Gray (ng338 AT cam DOT ac DOT uk).


This work is licensed under a Creative Commons Licence. If you use this content on your site please link back to this page.