Sharing evidence on humanitarian reliefBMJ 2005; 331 doi: http://dx.doi.org/10.1136/bmj.331.7531.1485 (Published 22 December 2005) Cite this as: BMJ 2005;331:1485
- Edward J Mills, fellow ()
- Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, ON, Canada L8N 3Z5
Needs a publicly accessible, searchable, and comprehensive database
One year ago the Asian tsunami struck, resulting in the largest humanitarian efforts of our generation. This year's hurricane Katrina and earthquake in Kashmir also showed that both developed and developing nations are ill prepared for major disasters. Rapidly sharing relevant information from relief agencies and academic and non-government organisations (NGOs) at such critical times can make an important difference to tens of thousands of people.
Relief agencies conduct fact finding expeditions in emergencies, as well as important public health measurements such as water testing, measles surveillance, and conflict surveillance. Their reports often provide the most up to date and relevant evidence on relief situations,1 but are too often shared only internally. For agencies and field coordinators to make informed decisions, access to this information is vital.
We must, therefore, consider how to create and disseminate evidence regarding humanitarian interventions.2 One absolute necessity is a publicly accessible, searchable, and comprehensive database on humanitarian disasters and approaches to relief. The lack of systematically documented or disseminated information leads to unnecessary duplication of efforts and ill informed decisions. Given the inadequacy of funding for relief aid, resources must be used wisely.
Some relief databases are already accessible to the public and NGOs. The largest is Relief Web (http://www.reliefweb.int/), established in 1996 by the United Nations, but it has been hindered by a lack of submissions from agencies and a reticence by academics to submit reports that may be under review at journals. Other resources include the SHARED Global Database, ELDIS (the Electronic Development and Environment information System), and ID21 (Information for Development in the 21st Century), but these have the same drawbacks as Relief Web and their reporting styles vary widely. Large NGOs and international agencies have, at times, maintained publicly accessible databases; smaller agencies sometimes post reports on their websites. None of these resources is sufficiently comprehensive.
A comprehensive database would have many aims but would also have to overcome certain challenges (box). Furthermore, the quality of evidence needs to be considered. The thresholds for acceptable evidence on humanitarian situations may be different from those for therapeutic interventions,3 and a formal hierarchy for it has not yet been established.1 4 Access to reports may allow evaluation of the effects of interventions through before and after analyses and systematic reviews. Many reports remain unpublished or inaccessible, however, making interpretation of single reports potentially misleading and interpretation of systematic reviews unnecessarily difficult.
Lack of access to reports from humanitarian agencies can reduce the quality of aid provided, just as inadequate evidence can hamper health care.5 After the tsunami, for example, several agencies made poorly informed decisions, such as using resources for mass burials.6 In areas of Banda-Aceh, Indonesia, health agencies conducted overwrought measles campaigns, resulting in children receiving as many as four measles vaccinations.7
On the other hand, access to evidence on psychological debriefing for survivors changed practice during the tsunami crisis. The best available evidence, a Cochrane review, showed that the intervention was of little use (odds ratio for post-traumatic stress disorder 1.22, 95% confidence interval 0.60 to 2.46) and may be harmful (2.88, 1.11 to 7.53),8 as well as wasting resources that could be applied to beneficial ends.9 The Cochrane Collaboration's work in the after-math of the tsunami led to Evidence Aid, a growing resource of summaries of best evidence on the effects of health care in disasters (www.cochrane.org/evidenceaid/project.htm).
Aims of a comprehensive database
Help the people who are making decisions by giving them access to the best current information
Facilitate systematic reviews to summarise and synthesise information
Avoid unwarranted duplication of efforts
Encourage collaboration across agencies
Provide ready access to the public directly and through the media
Improve before and after evaluations of conflicts, disasters, and interventions
Identify gaps in knowledge
Facilitate the development of measures and methods to evaluate relief and development
Creating a culture of responsible participation
Minimising threats to agencies on issues of contention and threats to staff from host nations
Encouraging academics to release findings before journal publication
Responsible participation can make these initiatives work. Academics should submit their relevant manuscripts to databases such as Relief Web. Moreover, we urge journals to submit the full text of all of their public health related articles to Relief Web, a policy which BioMed Central, an open access publisher, has pioneered, and has recently been joined by PLoS Medicine. The mass media could report more accurately on humanitarian situations. And funding agencies should look more favourably on evaluations of relief efforts and of the impact of their own responses.
We do not seek to place blame upon the many agencies and NGOs that provide selfless and important care to the most vulnerable people in humanitarian crises. We understand that our proposal may be viewed as a challenge, and we recognise that it is impossible to make all reports available, particularly those about relief in political disasters. In exceptional circumstances, publishing a report that seemed to be censorious of a host country could place an organisation's staff or the population in danger or risk the expulsion of the agency. Indeed the head of the Sudan mission of Médecins Sans Frontières (MSF) was arrested in May 2005 and detained for releasing a report reporting sexual violence observed at MSF clinics.10
We share a vision that everyone involved in making decisions about relief will be able to use evidence and knowledge generated by agencies and others. Archived evidence is a potent form of witness and testament for historical accountability and memory, and to achieve such an archive we have to collaborate. It is only a matter of time before another disaster will find us in disarray.
Collaborating authors on this paper, which was conceived at the 2nd McMaster-Lancet Peace Through Health conference in May 2005, are: Janet Robinson, USA; Amir Attaran, Canada; Mike Clarke, UK; Sonal Singh, USA; Ross E Upshur, Canada; Kenneth J Herrmann Jr, Vietnam; Salim Yusuf, Canada. We thank Kumanan Wilson, Lexi Bambas, Charles Tauber, Anthony Zwi for suggestions or critical revision. EM is supported by the Ontario HIV Treatment Network.
Competing interests None declared.