IJE Advance Access originally published online on March 15, 2006
International Journal of Epidemiology 2006 35(2):446-447; doi:10.1093/ije/dyl018
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Published by Oxford University Press 2006
Commentary |
Commentary: When should we monitor mortality in humanitarian crises?
Columbia University, 617 West 168th Street, New York, NY 10032, USA. E-mail: garfier{at}gmail.com
Accepted 24 January 2006
This paper on mortality patterns in Guinea-Bissau1 contributes to the current extensive discussions on monitoring mortality rates during humanitarian crises. Mortality is the most sincere expression of vulnerability and, when well monitored, should direct efforts for relief, protection, and humanitarian intervention. Unfortunately, areas with humanitarian crises in the world today seldom enjoy an active case finding system like that set up by the authors prior to conflict in Guinea-Bissau. One lesson from this research is the importance of maintaining, and if possible enhancing, any such system in areas of crisis. Failing this, at best we get occasional special mortality studies that are easily biased by rapid population movements and difficult to interpret owing to their cross-sectional nature. This is just the situation in most of the major humanitarian crises in the last decade, including the Democratic Republic of Congo, Afghanistan, and Iraq. Without a system of on-going monitoring, even if special studies are accurate, the information is frequently not available to act on in a timely basis. In Guinea-Bissau, monitoring helped minimize all these problems.
Humanitarian interventions are supposed to reduce harm to the most vulnerable. But if we measure vulnerability only by counting events that cause harm, how can we know if vulnerability has been attenuated or if assessment of populations at greatest risk were flawed? I would say that we cannot know this, and do not need to know. This research showed that mortality rate inequalities declined overall among affected groups. That is enough! Where humanitarian interventions are most successful, vulnerability to many groups should decline; precise metrics of relative risk can only be examined later. Notably, but not surprisingly, the historically most disadvantaged groups—children, girls, and members of the Pepel ethnic group, continued to have the poorest mortality outcomes. Larger-scale and longer-term interventions would have been needed to change these patterns.
What we still do not know is what led to, or attenuated, the risk of death. Information on mortality rates alone cannot give us the answer; we would need information on diseases and causes of death as well. In essence, active case finding gave us a series of cross-sectional studies on risk of death. Yet each person's life was a longitudinal condition, rich with information that cannot be captured by current methods. We urgently need validation of a simple approach to verbal autopsy assessment to do this.
It is remarkable that mortality rates declined to pre-war levels during the last 6 .months of the war when people returned to their homes, where pre-war infrastructure was still intact. People vote with their feet, and usually are loath to leave the relative safety of home for the risks of the wild or crowded camps unless death at home seems certain. In this case, residents of Guinea-Bissau rightly determined that their chances of surviving the war were greater in the war zone at a certain point than in camps where social and physical infrastructure was weak. We will be able to do much more to protect vulnerable groups when our methods of assessment catch up to their sense of relative risk and safety and we can use that information to enhance their subjective collective decision-making.
 |
Reference |
|---|
 Top Reference |
|---|
1 Nielsen J, Jensen H, Andersen PK, Aaby P. Mortality patterns during a war in Guinea-Bissau 1998–99: changes in risk factors? Int J Epidemiol 2006;35:438–46.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||