The ongoing Ebola epidemic in West Africa has drawn attention to the potential uses of big data in public health. The term big data is loosely defined and even more loosely used, but in current usage generally refers to naturally generated data (not the result of a survey or test) drawn from an extraordinarily wide set of points.

In regard to Ebola, a commonly cited source of big data is the vast call-data records (CDRs) from mobile phones. While mobile phones are less common in West Africa than in more prosperous areas, they are nonetheless owned widely enough to provide an accurate picture of the general population. Telecommunications companies records the caller, receiver, location, and time of each call placed, and also the location of phones at regular intervals when no calls are being made. Thus CDRs can be used to track frequency of calls to healthcare providers or hotlines, giving early warning of growing outbreaks. But the location data can also be used to track trends in movement of people in real-time with accuracy superior to that of the data from any other available technique.
CDRs have been used with great success in public health crises in recent years, as discussed in an Economist article: in 2009, the swine-flu in Mexico; in 2010, cholera in Haiti following the earthquake; and in 2012, malaria in Kenya (in this case, ordinary infections rather than an epidemic). In each case, researchers used CDRs to track movements of wide swaths of the population — 15 million phones were used in the Kenya study — to identify where people were contracting diseases and where they tended to travel once infected. In some cases, this enabled authorities to prepare earlier and more effectively to treat growing local outbreaks.

Yet despite a deadly epidemic and many pleas, researchers in West Africa have not been granted access to CDRs. At the center of the issue is privacy; the telecommunications companies appear ready to provide the information if regulators will permit (or in some cases, require) them to do so, but current privacy regulations restrict its release. CDRs are specific and identifiable, but the information can easily be made anonymous.
The problem is simply that the data, and concept of using it in such a context, is so new that most regulatory bodies have no precedent to guide their actions. New legislation takes time to draft, approve, and implement; the gravity of the situation has so far failed to accelerate that process sufficiently.
Ideally, regulators around the world will learn from this crisis and update policies now, before the next epidemic. With proper privacy measures in place, and the legal framework to facilitate and force the release of records, big data from mobile devices and other sources could play an enormous role in mitigating future health crises.

Comments are closed.