Who died as a result of the tsunami? Risk factors of mortality among internally displaced persons in Sri Lanka: a retrospective cohort analysis

BMC Public Health. 2006 Mar 20;6:73. doi: 10.1186/1471-2458-6-73.


Background: Describing adverse health effects and identifying vulnerable populations during and after a disaster are important aspects of any disaster relief operation. This study aimed to describe the mortality and related risk factors which affected the displaced population over a period of two and a half months after the 2004 Indian Ocean tsunami in an eastern coastal district of Sri Lanka.

Methods: A cross-sectional household survey was conducted in 13 evacuation camps for internally displaced persons (IDP). Information on all pre-tsunami family members was collected from householders, and all deaths which occurred during the recall period (77 to 80 days starting from the day of the tsunami) were recorded. The distribution of mortality and associated risk factors were analysed. Logistic regression modelling using the generalized estimating equations method was applied in multivariate analysis.

Results: Overall mortality rate out of 3,533 individuals from 859 households was 12.9% (446 deaths and 11 missing persons). The majority of the deaths occurred during and immediately after the disaster. A higher mortality was observed among females (17.5% vs. 8.2% for males, p < 0.001), children and the elderly (31.8%, 23.7% and 15.3% for children aged less than 5 years, children aged 5 to 9 years and adults over 50 years, respectively, compared with 7.4% for adults aged 20 to 29 years, p < 0.001). Other risk factors, such as being indoors at the time of the tsunami (13.8% vs. 5.9% outdoors, p < 0.001), the house destruction level (4.6%, 5.5% and 14.2% in increasing order of destruction, p < 0.001) and fishing as an occupation (15.4% vs. 11.2% for other occupations, p < 0.001) were also significantly associated with increased mortality. These correlations remained significant after adjusting for the confounding effects by multivariate analysis.

Conclusion: A significantly high mortality was observed in women and children among the displaced population in the eastern coastal district of Sri Lanka who were examined by us. Reconstruction activities should take into consideration these changes in population structure.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adolescent
  • Adult
  • Age Distribution
  • Aged
  • Child
  • Child, Preschool
  • Cohort Studies
  • Disasters / statistics & numerical data*
  • Female
  • Geography
  • Humans
  • Indian Ocean
  • Infant
  • Infant, Newborn
  • Logistic Models
  • Male
  • Middle Aged
  • Mortality*
  • Population Dynamics
  • Population Surveillance / methods*
  • Refugees / statistics & numerical data*
  • Rescue Work
  • Retrospective Studies
  • Risk Factors
  • Sex Distribution
  • Sri Lanka / epidemiology