‘Open data’ has become a buzzword in the development community. The excitement rests largely on an assumption that our ability to collect increasingly large amounts of data will automatically improve the quality of aid programs. But as the well-worn quote from Einstein goes: Information is not knowledge. To use data to strengthen development efforts, we need to ensure that data are actively and visibly incorporated into development practice and policies, thereby closing the feedback loop between development data and development action.
As data is increasingly controlled by the public, it is often seen by technology optimists as the democratization of knowledge. Yet we are still far from knowing just how accessible the bulk of information is to development practitioners and communities. Almost a third of reports published online by the World Bank, for instance, are never downloaded, raising concerns for the integration of the data into aid responses by showing that available development data is not consistently used by practitioners or policy-makers.
The aim of development data policies should be to strengthen the feedback loop between data collection and its application, switching from an emphasis on ‘open data’ to one on ‘opening data’.
In public health, good data are, quite literally, a matter of life and death—the Ebola outbreak in West Africa has made this explicit. Being able to trace the movements and human connections of an Ebola patient in the 24 hours before their diagnosis via tracing calls and geolocation, could make the difference between one case and one hundred. Data could range from predictive surveillance of virus outbreaks, to medical supply management, to citizen reporting of suspect cases. Yet with access to all of this data, an outstanding challenge for practitioners and policy-makers is how to make increasingly large datasets have a tangible impact on the day-to-day lives of health systems, workers, and patients.
How open data can be useful
There are at least three possible ways the aid community can promote this, which are illustrated by examples emerging from the medical field.
The first use should be to ensure data are used to improve aid agency responsiveness. The data platform AidData, has been applied in Uganda to strengthen the feedback loop between public investment and community opinion, allowing gaps in aid supply and aid demand to be identified, and responded to, in a targeted, effective manner. Real-time information exchange, if used properly, should facilitate the adaption of aid efforts to create the best outcomes in a changing environment.
The mHero mobile tool currently being used by the Ministry of Health and Social Welfare in Liberia is a good example of how real-time data management can be used to respond to specific local needs within a larger disease program, thus reducing waste as well as improving the timely provision of medical equipment and resources. The next step could be to extend the service from government to patient knowledge demand, allowing patients to present real-time health queries on Ebola symptoms to trained health workers who can advise them remotely, avoiding the health risk to the patient or public posed by requiring them to physically travel to the nearest health care centre.
When the Ebola outbreak first started in Guinea, a free hotline set up by the Ministry of Health in Guinea for questions on Ebola received 200–300 calls daily, highlighting the population’s need for this kind of a service. By transferring this to an SMS service, an aggregator platform could analyze geographical trends between requests for information, which could reveal additional needs and gaps in local health service provision. Through identifying and matching health demand and supply, real-time data could be used to play an important role in improving health care access.
The second is to ensure that data strengthen the trust relationship between data providers and data users, by highlighting the use of community data to the communities themselves. The lack of a tangible response to feedback strongly disincentivizes citizen reporting. An analogy made in a study on community health in a rural village in Karnataka, India, where data was collected but never returned to the community through improved health policy, was that they felt like “students who sat an exam but never received the results”. This type of gap undermines any attempt by a public health provider to use citizen data to improve health care impact and outcomes.
Nor should making data available for public consultation be confused with the effective use of data. Citizens rarely want data itself. What they want are the resources which better data can bring by clearly identifying their local needs to authorities, such as improvements in service provision and accountability. If governments, donors and NGOs can show that they are responding to citizen need by actively responding to publically provided information, a virtuous cycle will be created where more data lead to better services: closing the feedback loop. This will in turn create greater trust in public systems and greater local community willingness to identify needs, making aid interventions more targeted by switching focus from the ‘information’ to the ‘communications’ element of ICT in development.
A third and final way to strengthen data usage would be to emphasise it as an escape route from vertical program approaches. If data could centralize and publicize the geographic and thematic areas where agencies are working, this could help identify opportunities for inter-agency collaboration by avoiding duplicated content and offering new opportunities. Some groups, notably the World Bank, have begun using this to track their own activities. If this were opened up to create a global geomap for aid programs it could increase collaboration, effectiveness, and impact across organizations and sectors.
Data is only a tool
The lesson drawn from the aid community’s approach to data is that it is not a polypill which can cure all ailments in public health. It is a tool which, if properly used, can work to strengthen health systems and improve their identification and response to specific health challenges. However these gains are not automatic, and development actors need to understand the data as a means to an end, as opposed to the end in itself, by mending the loop between data collection and development policy – from database to data-based policy. This includes planning how to incorporate data into programmes before it is collected; establishing a continuous feed between data gatherers and policy-makers; ensuring data from different sources can be centralised quickly and efficiently; and delivering any response based on the data in a way which allows citizens to see and feel the impact of their contribution.
Napoleon Bonaparte once remarked that winning a war was 90 percent based on information. In reality, the difference between winning and losing a war also rests on how leaders respond to timely information, not just on the information’s existence. If we don’t understand how to properly use data in development, we might as well not have it.
The opinions expressed in this blog post are those of the author and do not necessarily reflect the official position of GDN.