It is commonly understood that there is an inverse relation between the wealth of a nation and disaster risk.Reference Kellenberg and Mobarak1 Richer nations are better able than poorer nations to protect themselves and mitigate the effects of disasters. They experience fewer deaths and damage, although the number of catastrophic events in wealthy countries is about the same as in poorer countries.Reference Kahn2 The technology and resource-rich solutions available to the developed world may not be completely applicable to or replicable in the developing world. Poorer countries suffer disproportionately more because disasters overwhelm local health resources and response more easily and further damage the economies of poor nations.Reference Toya and Skidmore3 Not all nations are affected equally and after a major disaster, some of the better-off developing-world economies may improve themselves economically by reinventing themselves.Reference Crespo Cuaresma, Hlouskova and Obersteiner4
The body of research available on the subject of disaster medicine is growing. Considering, however, that 85% of disastersReference Quarantelli5 and 95% of disaster-related deathsReference Alcántara-Ayala6 occur in the developing world, the overwhelming number of casualties has contributed insignificantly to the world's peer-reviewed literature.Reference Kahn2 This evidence, or the lack of it, is of concern because it forms the basis of practice and advocacy for addressing future disasters. With the looming hazards of pandemics, climate change, global terrorism, and armed conflicts, policy makers and governments will need high-quality scientific data to make informed decisions for preparedness and mitigation.Reference King7 We examined the available evidence on disasters in the developing world in peer-reviewed journals for quality and quantity in this systematic review.
METHODS
Using the available literature about disasters in the developing world in medical and public health journals, we searched PubMed using the Medical Subject Heading (MeSH) search terms disasters, disaster medicine, rescue work, relief work, and conflict. There were no limits set for the time period or the type of articles. All foreign-language articles were included. The search was refined using the MeSH terms developing country, including developing nation, under-developed countries, under-developed nation, Third-World countries, less-developed countries, and less-developed nations. Developing nation was defined using the International Monetary Fund criteria of emerging and developing economies in the analysis.8 Duplicate articles were removed. The final list of selected manuscripts was analyzed by type of article, level of evidence, theme of the manuscript and topic, author affiliation, and region of the study. All of the articles were classified according to the type of publication as original article, review article, case report, comment, editorial, interview, letter to the editor, news article, or report. They were further classified into various levels of evidence using a standardized classification of levels of evidence, formulated by the Centre for Evidence Based Medicine.9 Briefly, the levels are level I: randomized controlled trials, prospective cohort studies; level II: retrospective cohort studies, outcomes research; level III: case-control studies; level IV: case-series studies; level V: case reports, expert opinion, or comment.
Two of the authors (P.T. and N.R.) reviewed and coded all of the records independently. All of the coding discrepancies were discussed and reconciled through consensus. Authors of articles with no clear affiliations were subjected to a broad Internet search using Google. Author affiliations that were unavailable after this process were classified as “missing information,” and 20 articles were excluded from the author analysis. The origin of collaborative studies between developing- and developed-world authors was classified using the country of the first author's institute or organization. Each of these articles was also classified by the topic or theme of the manuscript. Fourteen articles were excluded because they did not belong to the domains of disasters, conflicts, or disaster medicine, even though they were selected by PubMed. Level III articles were evaluated in detail for methodology.
RESULTS
Citations using the MeSH search terms disasters, disaster medicine, rescue work, relief work, and conflict yielded 63 196 results. After these results were refined using the MeSH term developing country, 438 articles were retained. Less than 1% (0.69%) of the citations in the disaster-related literature on PubMed addressed developing-country disasters.
Publication type and level of evidence were analyzed, and about half of these manuscripts (46.5%) were found to be original research articles (36.1%) or reviews (10.4%), and more than one-fourth (29.5%) were commentaries. Case reports, editorials, interviews, letters to editors, news articles, or reports made up the balance of all of the publications. A total of 97.4% (149/153) of all of the original research articles were level IV or V evidence. All of the reviews were either level IV or V evidence. No publication was categorized as level I or II evidence. These findings were tabulated (Table 1). The percentage of agreement between the 2 authors, while classifying each of the articles for the level of evidence using the Centre for Evidence Based Medicine criteria, was 93.8%.
TABLE 1 Publication Type With Level of Evidence
One-fifth (20.3%) of the authors of all of the articles on disasters in the developing world were from the developing world (82/404), the rest being from the developed world (Table 2). Three publications (3/404) with level III evidence were written by developing-world authors and 1 was by developed-world authors. Author affiliations of 20 of 404 manuscripts could not be determined and were labeled as missing (4.9%). A total of 2.7% (11/404) of the articles were collaborative studies.
TABLE 2 Author From Developed or Developing Country
The predominant theme (29.1%) was missions, health care provision, and humanitarian aid during the acute phase of disasters in the developing world. Commentaries on policies, vulnerable populations, and food, water, and nutrition were the next most common topics. Tools, mental health, specific diseases, armed conflicts, ethics, and epidemiology were addressed by a smaller proportion of articles. The breakdown of topics and themes is found in Table 3. The 4 evidence level III articles were evaluated and tabulated in Table 4.
TABLE 3 Topics of Developing-World Disaster Publications
TABLE 4 Summary of Level II and Level III Publications
COMMENT
Number of Citations
The paucity of articles written about developing-world disasters is evident. The reasons for the publishing silence may be multifactorial.Reference Sumathipala, Siribaddana and Patel14 Poor funding for research, competing clinical commitments, pressure in developing-world hospitals to provide services rather than publish research, poor methodology leading to rejection by peer-reviewed journals, and the lack of ability in English are possible barriers, which may explain this disparity. There may be a publication bias in medical journals against diseases that ravage the least-developed regions of the world.Reference Horton15 Our dismal finding that <1% of articles focused on developing world disasters was in stark contrast to the fact that 85% of disasters do occur in the developing world.Reference Quarantelli5
Levels of Evidence
Besides the number of publications, the other cause for concern is the lower level of evidence that dominates the existing disaster medicine literature. The subgroup of developing-world research publications must be elevated first to be on par with the disaster publications from the developed world. Unfortunately, level IV and V publications, which dominate developing-world disaster publications, add little to the evidence base. There is a body of literature, however, that provides the knowledge base for relief during disasters in the developing world. These are internal documents published by international relief agencies such as the World Health Organization, the United Nations, the Red Cross, Doctors Without Borders, and Oxfam, which are usually available in the public domain with free full-text access. Although these documents constitute expert opinion reports (level V evidence), which have not been subject to the rigor of the peer-review process, these documents are sometimes used as best practices guidelines by most disaster relief workers.
Generating level I data (randomized controlled trials) is unlikely in the disaster setting, given the nature of the event and interventions; therefore, the strategy would be to generate more level II and III publications. Cohort studies, audit, or outcomes research, which would qualify as level IIb/IIc or Level IIIa/IIIb, require follow-up studies of the victims for an extended period of time. Long-term studies, with more robust methodology, are what are sometimes needed to improve the scientific base of studies in disasters. To aid in this objective, the developed world could aid researchers from developing nations by building capacity for good research methodology and commissioning level II or III studies. Some of the examples of better developing-world research are seen in Table 4, in which the authors have used epidemiological methods for evaluating trends, validated testing tools, case-control or case-audit methodology to research developing-world problems. Finally, learning from developed-world research, the direction for the future should be systematic surveys with primary data collection, appropriate sampling methods, prospective cohort studies with >80% follow-up rates, and analysis based on clinically sensible costs or alternatives.9 These would constitute high-quality (level Ib/Ic) studies.
Authors
The underrepresentation of the developing world in peer-reviewed disaster literature may be a reflection of the publishing world in general. In a retrospective survey of articles published in the British Medical Journal, The Lancet, the New England Journal of Medicine, Annals of Internal Medicine, and the Journal of the American Medical Association, Sumathipala and coworkers found that 6.5% of the publications in these journals are written by authors from countries in which 90% of the world's population live.Reference Sumathipala, Siribaddana and Patel14 In our study, four-fifths of the authors were from the developed world writing about the developing world. Therefore, it was not unusual that many publications focused on the acute phases of response and short-term recovery. These phases of disasters are convenient for visiting rescue teams to write about. The short temporal relation of health care personnel and disaster victims is reflected in the cross-sectional nature of the publications. This developed-world view of developing-world disasters represents the predominant view in the indexed PubMed literature. In our study, only 2.7% (11/404) articles were collaborative studies. This is not uncommon in other medical research works, in which there are a significant number of articles using developing-world data without a single author from these countries.Reference Sumathipala, Siribaddana and Patel14 However, collaborative studies that build capacity in countries with low-science intensity by developed-world scientists with greater experience in disaster research may provide the push needed for high-quality evidence in disaster medicine.Reference Horton16
Topics and Themes
The most common topics (28.2%) were missions, health care provision, and humanitarian aid during developing-world disasters. Commentaries about policies, vulnerable populations, and food, water, and nutrition were the next most commonly found in citations. These commentaries and letters to the editor usually lamented the poor working conditions and lack of interagency coordination in the developing-world disaster zones. Furthermore, there was a strong reporting bias about the good work done by missions and field hospitals for developing-world victims, although there is evidence of their limitations.Reference von Schreeb, Riddez, Samnegård and Rosling17Reference Roy, Shah, Patel and Coughlin18Reference Roy19
There are definite gaps in the themes covered by these publications. Tools, mental health, specific diseases, conflicts, ethics, and epidemiology were addressed in a small proportion of articles. Mental health, in particular, lends itself to systematic research, using premeasures and comparison groups between exposed and the unexposed people.Reference Stoddard20 The future priority areas in research include long-term economic outcomes, health system recovery, occupational rehabilitation of victims, community-based disaster preparedness, resilience of communities in low-resource settings, public health interventions, monitoring and evaluation of interventions, and research tools validated for the developing world. Alongside these, the developing world needs to be concerned about the ethical issues involved in research on vulnerable populations.Reference Chung, Jones, Campbell, Glover, Gelberg and Chen21Reference Perlman22 Not all nations are affected equally and some of the more well-off developing-world economies will emerge economically more robust after a major disaster by reinventing its economic strategy.Reference Crespo Cuaresma, Hlouskova and Obersteiner4 These existing social and economic risk differences between the nations, within the group of emerging and developing economies, is another unexplored area in the current disaster-related literature.
Developing World Research Priorities
Proposed solutions for the developing world are missing in the literature. These solutions are primarily opinions or exploratory studies. To develop indigenous research, a more systematic approach and a local research agenda for developing-world nations will be required, rather than transplanting developed-world tools and interventions. The focus of the knowledge base will need to shift away from the acute phase of disasters and to prevention, mitigation, and preparedness. Long-term recovery studies will be the immediate next priority, along with standardizing definitions and research methodology for future studies. Finally, funding will determine the pace of research. Humanitarian aid is directed at the acute phase of disasters and millions of dollars have been spent on humanitarian missions. It would be cost-effective to channel some of these funds toward developing-world research, as sustained aid to promote research in disaster medicine. These investments in developing-world preparedness planning, would pay dividends in mitigating future disasters.
Although there may be lessons learned from disasters in the developing world, these are not visible in the disaster-related literature. This may be a publication bias or a genuine lack of submissions addressing developing-world disasters. Research drives advocacy and that in turn drives policy; however, the developing world does not yet contribute to it.
Limitations
The limitation of this study is that it uses only PubMed. Because PubMed is the only free database available to researchers in the developing world, it was selected for the searches. Although the searches use the robust MeSH classification system developed by the National Library of Medicine, the exact proportion of misclassification errors in PubMed is not known. Furthermore, disaster medicine is an emerging science and was only recognized as a MeSH term by the National Library of Medicine in 2008. There are differences that need to be acknowledged within the capacities of developing nations, although they have been pooled together as emerging economies. China and India are notable examples, as are a few countries in Africa, such as South Africa.Reference King7 Lastly, all of the authors of the present study are from the developing world, which may have contributed to some of the classification bias about the levels of evidence and author affiliations of the articles.
CONCLUSIONS
Most disasters and deaths from disasters occur in the developing world; however, <1% of all disaster-related articles are about disasters in the developing world. The developed world authors four-fifths of the articles about developing-world disasters and contributes the predominant perspective. It is important that the developing world contributes to our learning about disasters through better-quality systematic research. Aid for sustaining long-term disaster research may be a more useful investment in mitigating future disasters than short-term humanitarian aid missions to the developing world.
Author Disclosures: The authors report no conflicts of interest.
Acknowledgments: The authors thank Meenakshi D’cruz for help in editing the manuscript.
