Background: The crude mortality rate of the whole population and the mortality of children <5 years of age are the common indicators of the severity of a complex emergency situation. However, these indicators rarely take account of differences in socioeconomic conditions and vulnerability.
Methods: We followed a population in Guinea-Bissau, which fled when fighting took place in the capital during the war in 1998-99. The population stayed close to the area of conflict and returned as soon as a cease-fire was negotiated. A peace treaty was signed after half-a-year. The following 6 months was a period of returning and re-settlement, even though two outbreaks of fighting occurred.
Results: In the first half-year the mortality rate was 78% [mortality ratio (MR) = 1.78; 95% CI 1.61-1.97] increased and mortality for children <5 years of age doubled (MR = 2.07; 95% CI 1.79-2.38). In the last 6 months of the war, mortality was slightly increased for children and not at all for the total population. In the first half-year, households living in better houses and having members with schooling were less affected. In the 're-settlement' period two inequalities emerged; the largest ethnic group, Pepel, continued to have high mortality when the mortality of other groups declined; likewise girls continued to have an elevated mortality whereas mortality of boys declined.
Conclusion: Whereas specific 'free' interventions reduced social inequalities for the groups affected, for the total population health-inequalities were slightly amplified during the war. Once the population returned to their urban homes, mortality fell to pre-war levels even though some fighting continued, limited humanitarian aid was available and the pre-war infra-structure had not been re-established.