The Ebola emergency: perspectives on information management and mapping responses

As for many organisations, the Ebola outbreak was a new experience for MapAction, which had never previously responded to a large-scale health emergency. The gravity of the outbreak and evident applicability of geospatial analysis to inform the humanitarian response led the organisation to mobilise a large proportion of its resources in support. Over a period of four months, MapAction deployed 11 volunteer Geospatial Information professionals and three paid personnel to Liberia, Sierra Leone, Ghana and Mali. In addition, remote support was provided prior to, during and since those field deployments to a variety of agencies, including the UN Office for the Coordination of Humanitarian Affairs (OCHA), the UN Disaster Assessment and Coordination System (UNDAC) and the UN Mission for Ebola Emergency Response (UNMEER), as well as the governments of each country, other UN agencies, NGOs, clusters and remote technical communities. As a member of the Digital Humanitarian Network, MapAction was involved in liaising with a wide network of online data providers, and remains the channel for UK Department for International Development (DFID) funding of an Ebola Coordinator for the Network.

Initially, MapAction offered remote support to the regional OCHA office in Dakar, Senegal, assisting with mapping core services remotely (treatment centres, laboratories, safe burial locations and social mobilisation). Once in-country MapAction established the administrative mapping framework for the government-led Ebola Response Command Centres in Liberia and Sierra Leone, set out an information management structure for Mali in preparation for a possible outbreak there and worked with UNMEER’s information section in Ghana to develop a data management system for use across the region.

The Ebola emergency was characterised by its inherently geographic nature – location was critical. Locating people with the virus was essential for treatment and contact tracing. Identifying clusters of transmissions helped operational prioritisation locally, while knowledge of broader patterns and trends helped to shape strategic responses at a national level. Applying geographical analysis to the problem therefore seemed an obvious course of action.

GIS analyses data and turns this analysis into visual maps – or ‘pictures’ – that tell the story of what’s happened. Such story-telling should aim to deliver useful analysis about the allocation of resources, the scale of response and locations of greatest unmet need. Ultimately it is only useful if it results in aid delivered to the right people at the right time in the right way. Our teams, as first responders, are used to the rapidly changing story that evolves during an emergency, but the Ebola outbreak presented a much more dynamic picture as the number of patients grew, geographical patterns shifted and urban contexts became more significant, requiring detailed analysis on a daily basis.

This highly dynamic environment placed new demands on the data and information management community. It raised questions about the provenance and accuracy of the data being shared, and whether data could be collected and transferred in a timely way. The close link between accurate information and effective response made us realise that assumptions and approximations that would be acceptable in other contexts could not be made here. Bed availability at a particular treatment centre, for example, not only changes from day to day, but is a very specific number that is only useful if precise and accurate. This meant that constant, meticulous effort was needed to maintain datasets and their corresponding map products, while emphasising the importance of recording the source and date of every detail. The fact that this was an emergency experienced at a village/neighbourhood, chiefdom, district, country, regional and international level further complicated information flows and the telling of the story. In short, in information terms, significant effort and resources were required.

Lessons

Four core lessons emerged for MapAction during the emergency. First, data is a critical element of robust information management services. Second, data is only useful if it informs decisions relevant to the emergency. Third, being on the ground alongside service users is essential in understanding their needs and evaluating the relevance and use of emerging data. Fourth, in an escalating and ongoing emergency, handing over information management systems that could be readily maintained was essential.

The role of data in information services is one of the key lessons MapAction took away from this emergency. This lesson was not new. We have dealt with large volumes of data before: in Haiti, an overwhelming amount of data was generated by social media and mobile technology, and in the Philippines MapAction produced over 2,000 unique maps in a six-week period. In the Ebola response, however, we realised the importance of being selective about which data was significant. The complexity of the situation, with different information providers in different locations, meant that it was difficult to finalise which data should be used as the master data set, and it was unclear who had the authority to decide this. This governance question is key in complex emergencies; although guidelines exist in relation to some data sets, IASC Guidelines Common Operational Datasets (CODs) in Disaster Preparedness and Response As Requested by the 77th IASC Working Group Meeting IASC Endorsed November 1 2010.  there was a lack of clarity about this in relation to the health treatment unit dataset. In this sense, Ebola was more of a data-driven emergency than MapAction had experienced previously, but one where judgement needed to be exercised in terms of which data was useful.

In Sierra Leone MapAction initially found map and data information that was contradictory, without clearly defined products or validation. This served as a useful reminder that it is always important, when collecting data, even in a data-driven environment, to ask what the data is for. The most critical questions facing decision-makers included how to stop the spread of the infection and how and where to care for those already infected. This led MapAction to adopt a deliberate strategy of focusing information management on a few key datasets: case data and Ebola care facilities. These datasets were quite simple to identify but demanding to maintain because of the rapidly changing situation. Any delay in collecting data or inconsistencies in the way data was collected meant checking and cleaning data took more time, delaying a rapid turnaround of analysis each day. Thus, MapAction deliberately chose key datasets, such as case data and Ebola care facilities, that both had a defined purpose and could realistically be maintained by the organisation’s volunteers.

MapAction’s proximity to information providers and service users enabled a contribution to the response that would not have been possible remotely. For example, it became apparent in Liberia that there were particular sensitivities related to identifying settlements below a certain size. This was an invisible but important issue highlighting the imperative of understanding the context in which information management is delivered. In Sierra Leone, MapAction focused on Ministry of Health statistics, which provide more authoritative and complete data than other datasets. Much of this was not visible prior to arrival in the country, but provided essential context to map production, and for relationship-building and coordination.

It was a priority to identify a focal point within the first two weeks to hand over the mapping function to. However, recognising that this had been a particularly demanding emergency for producing meaningful and useful products led the field teams to strip back the function to essential map products and data management processes to deliver these maps, and the necessary skills in emergency mapping. In the case of Sierra Leone, the datasets and maps established by MapAction during its mission are now maintained and published by UNMEER.

Conclusion

MapAction was not the only GIS provider in this emergency: many others with whom we work closely, as well as new colleagues, were active in-country and remotely. MapAction benefited from the work of organisations including the Humanitarian Open Street Map team and the British Red Cross. In Liberia MapAction and the Liberian Institute of Statistics and GIS convened a GIS coordination group for all providers. This proved a useful starting-point for sharing GIS information further in-country. MapAction also supported the coordination of online volunteer communities, through the establishment of a Skype Ebola group and by channelling funding from DFID to the Digital Humanitarian Network to hire a coordinator for the Ebola group. The growth and scope of this group of volunteer networks illustrate the value that participants have gained from it and the value of the Digital Humanitarian Network in providing support more widely to the humanitarian response.

Although data sharing was not always straightforward, there is no doubt that there was a desire to do it as well if not better than in previous emergencies as the value of GIS was well recognised and the tools were available. However, it is also clear that governments need to maintain and make available accurate demographic, reference and health data with clear sign-off procedures on new data emerging, to give clear visibility of the problem, and the basic framework within which to respond. The place of ‘command and control’ – albeit gently applied and received – was as relevant in this emergency as in any other. This of course is not a new lesson, but one that is amplified when all the goalposts of a ‘standard’ humanitarian emergency have moved.

Liz Hughes is Chief Executive of MapAction. Nick McWilliam and Anne Frankland are GIS professionals and volunteer members of MapAction.