Ebola: How Big Data Failed To Help Curb Its Spread

When cholera hit the crippled island of Haiti in 2010 it wasn’t only antibiotics that halted the killer disease. It was mobile phone records.

So when the haemorrhagic killer Ebola struck West Africa last year researchers were sure they could use the same tools to limit the death toll.

Cell records could have helped track the movement of people and help stop the spread of disease

Picture: Ebola treatment unit (ETU) in N’zerekore by UNMEER. Released under CC4.0

In the case of Ebola, one of the world’s deadliest and most infectious diseases, researchers were working against the clock. The first Ebola infections struck in March 2014, in the West African country of Guinea. Historically, with Ebola, one infected person typically infects two others. By October the death toll had risen to more than 10,000.

‘We Needed The Data’

But instead of up-to-the-minute information on who was moving where, in order to potentially identify carriers of the deadly disease that can kill people within one to two weeks of infection, the mobile operators handed over only limited data, much of it historical. In some cases operators flatly denied access at all.

“We needed the data,” says Dr. Linus Bengtsson, a co-founder and executive director at Flowminder, a Stockholm-based non-profit organization staffed by epidemiologists from Karolinska Institute that is helping pioneer the use of mobile data analytics to address global health.

The missed opportunity underscores how and why the promise of Big Data is still not being fully realized.

Researchers with Flowminder, academics at Harvard University and data scientists at Telefónica have pioneered the applied use of call data records (CDR). Studies in Zanzibar and Kenya on the spread of malaria showed that mobile data can reveal how populations move about the city in their daily commutes; that data can help researchers understand the spread of infectious diseases.

Measuring how active a cell tower is over a 24-hour period shows how many people are moving in and out of the cell’s area.

So spotting any changes in activity in the towers in areas hit by Ebola could have made a vital difference in curtailing its spread, says Nuria Oliver, scientific director of Telefónica Research.

It was not to be.

The lesson from the Ebola crisis is that unless governments, regulatory bodies, NGOs, mobile operators and other businesses come up with an agreed way to share and control use of the data — and get buy-in from the public — this kind of mobile data can not be fully used for the public good.

Fears Data Could Be Misused

“It is important to note that unlike other companies that may provide Internet-based mobile communication services, mobile operators are subject to express rules that require them to protect the confidentiality and security of customer data and the privacy of customers,” says Pat Walshe, the GSMA’s director of data privacy. “It is natural, and proper, therefore, that operators seek to ensure they comply with relevant laws and to protect not only their customers’ privacy but other risks that may arise from new and uncontrolled uses of data.”

Walshe says operators were flooded by requests from research and response communities and that these were “not always accompanied by clear plans about how and why that data was needed and going to be used.” What’s more, he says, “there didn’t seem to be clarity around how any potential analysis using this data would actually be integrated in the response and be used by decision makers.”

Complicating matters is that “some of the affected countries have a recent history of civil conflict,” says Walshe, and there was fear that information about the movements of certain groups could be misused.

Why Network Operators Wouldn’t Share

In 2014 The World Economic Forum questioned mobile operators about why they were reluctant to share population movement data during the Ebola outbreak in West Africa. Reasons cited included:

Geopolitical: Concern that sharing detailed population movements with third parties could lead to that data being used for purposes unrelated to the disease outbreak, such as targeting specific groups in times of conflict.

Regulatory: In case of data breaches mobile operators could face loss of their operating licenses, criminal or civil investigations, damage to regulatory ratings or financial penalties.

Public Backlash: Sharing data in a context that was not originally intended could result in reputational risk that hurts mobile operators’ businesses.

Financial Fallout: Operators fear they could be held liable if inaccurate data or faulty analytic models end up negatively impacting the financial health of individuals, businesses and national governments.

Privacy and Security: Privacy breaches could include exposure of individuals’ identities, health status, activities and associations. For businesses this could mean the loss of intellectual property or physical assets.

Nonetheless some of the requested data was released, he says. And the good news is that Millicom, a global telco with over 56 million customers in Latin America and in African countries such as Chad, Democratic Republic of the Congo, Ghana, Rwanda, Senegal and Tanzania, has pledged to work more closely with Flowminder in the future.

Complex Issues Need To Be Resolved

In October the GSMA issued privacy guidelines for mobile operators to help them structure the way they could disseminate data related to the outbreak.

A fundamental principle of the detailed guidelines is that the raw CDR data should never leave the custody or control of the mobile operator. The mobile operator can then partner with expert organizations such as Flowminder to produce and provide the relevant non-sensitive analysis and statistics to approved third parties. Only the resulting aggregated analysis, maps and forecasts would leave the mobile operator’s custody.

The collaboration with Millicom is a first of its kind to fully align to the industry privacy guidelines. Flowminder says it hopes it can function as a template and encouragement for other similar mobile operator partnerships.

Still, numerous complex issues need to be resolved. “It is not as simple as allowing data to be used for purposes that were not originally envisaged nor assuming that because data is anonymized that risks disappear,” says the GSMA’s Walshe. “The question is more, what data is needed to support the eradication of Ebola and from this what are the risks to individuals, communities, unique groups or individuals and so on from the various contexts in which the data will be used. The necessary techniques and accountability frameworks are not agreed or in place.”

Finding a way forward will be high on the agenda at the Mobile World Congress in Barcelona this year. The GSMA is hosting a seminar during the Congress entitled “Mobile Big Data: Solving Real Life Challenges While Maintaining Users’ Privacy and Trust.”

The Ethics And Governance Of How Data Is Used

The GSMA has been working with a number of institutions focused on data ethics, including the World Economic Forum, which is planning its own workshop on the topic during Mobile World Congress.

“Using big data to drive transformative change for sustainable development is an unprecedented opportunity,” says William Hoffman, the World Economic Forum’s head of data-driven development. “One of the key priorities of addressing this opportunity will be the ethics and governance on how data is used.”

The World Economic Forum, in collaboration with A.T. Kearney, is planning to publish a report on this topic during the first quarter of 2015.

Many questions need to be answered if Big Data can effectively be used for the public good. Among them is who is going to foot the bill. “There is a cost to extracting this aggregated data. It could be non-negligible,” says Telefónica’s Oliver.

But Oliver is optimistic. “It is fantastic that the World Economic Forum, the UN and GSMA are interested in this. It is a multidisciplinary challenge but if we work together we can come together on a reasonable plan.”



Related posts