Estimates of Mortality in Conflict Situations Global Health Fellowship, St. Lukes- Roosevelt Hospital Center.

Estimates of

Mortality in Conflict

SituationsGlobal Health Fellowship, St.

Lukes-Roosevelt Hospital Center

Changing Profiles of Conflict

Changing Profile of Conflicts

Most conflicts within nat’l borders

Poor countries

Civil wars with poorly trained armies

Asymmetric wars

Civilians and soldiers affected

Disease/malnutrition > violence

Deaths in Major Recent Wars

Country Number of deaths*

‘66-’70 Biafra 1.0-3.1 million

‘71 Bangladesh 300,000-3 million

‘70-’75 Cambodia 2-600,000

‘75-’80 Pol Pot 1.2-3.3 million

‘74-’94 Ethiopia 1-2 million

‘75-’93 Mozambique 600,000-1 million

‘79 Afghanistan 1-2 million

‘83-’98 Sudan 1.3-2.0 million

‘80-’88 Iran/Iraq 1 million

’91-’98 Somalia 4-500,000

‘91 Gulf War 150-200,000

‘92-’95 Bosnia 160-250,000

‘98-’02 DRC 3.3 million

Deaths in Major Recent Wars: Epidemiological Estimates

Country Number of deaths*

‘66-’70 Biafra 1.0-3.1 million

‘71 Bangladesh 300,000-3 million

‘70-’75 Cambodia 2-600,000

‘75-’80 Pol Pot 1.2-3.3 million

‘74-’94 Ethiopia 1-2 million

‘75-’93 Mozambique 600,000-1 million

‘79 Afghanistan 1-2 million

‘83-’98 Sudan 1.3-2.0 million

‘80-’88 Iran/Iraq 1 million

’91-’98 Somalia 4-500,000

‘91 Gulf War 150-200,000

‘92-’95 Bosnia 160-250,000

‘98-’02 DRC 3.3 million

Biafra

Dr Karl Western from CDC conducted a convenience sample

Ran around countryside where security would permit

Estimated fraction of children in enclave with small pox vaccination scar

CDC estimated # living, know small pox doses administered Estimated 1 million deaths Estimated 30% of population suffering

from famine edema, 2/3rds with dangerous loss of weight

Bangladesh

Monitored 120,000 people in Matlab

Birth rate went down to 7% in year of war

CMR went from 15 to 21/1000/year

Assuming population of 70 million, approx 500,000 excess deaths from war (Curlin GT, et al)

Liberia through ‘92

MSF did survey in part of country

Applied rate to entire country

Estimated 150,000 excess deaths

DRC

1999 IRC survey in Katana, estimates 15,000 deaths

2000 NYTimes estimates 50,000 excess deaths

IRC does 5 surveys, estimates 1.7 million deaths

2001 IRC 6 surveys estimating 2.5 million

2002 IRC 20 surveys estimating 3.3 million

2004 IRC nationwide survey 3.8 million

Estimates of Mortality

Crude mortality rate (CMR)-total # of deaths that occurred in a population of known size that is at risk of death in a certain period of time

Under 5 mortality rate (U5MR)-age-specific mortality rate with above definition

Crude Mortality Rate

CMR=(number of deaths)/(midterm population at risk X duration of time period) x 10,000 persons

Under 5 Mortality

U5M=(number of deaths)/(midterm population of U5s at risk X duration of time period) X 10,000 persons

Estimates of Mortality

The numerical difference between a “crisis CMR” and the “baseline CMR” is termed the excess mortality

Used for estimating the magnitude of the emergency and monitoring the humanitarian response

Estimates of Mortality

Direct deaths—those caused by war-related injuries and attacks

Indirect deaths—those caused by the worsening of social, economic, and health conditions in the conflict-affected region

Estimating Mortality: direct deaths

Mortality after the 2003 invasion of Iraq: a cross-sectional cluster sample survey

Gilbert Burnham, Riyadh Lafta, Shannon Doocy, Les Roberts

Mortality before and after the 2003 invasion of Iraq: cluster sample survey

Les Roberts, Riyadh Lafta, Richard Garfield, Jamal Khudhairi, Gilbert Burnham

**100,000 excess deaths, violent

Estimating Mortality: indirect deaths

Mortality in the Democratic Republic of Congo: a nationwide survey

Benjamin Coghlan, Richard J Brennan, Pascal Ngoy, David Dofara, Brad Otto, Mark Clements, Tony Stewart

**600,000 excess deaths, 38,000/month, easily preventable and treatable illnesses

Indirect deaths

Potentially holding political and military leaders responsible

Collaboration between epidemiologists, statisticians, human rights organizations

Bolstering of public health system

Challenges of collecting data

Security

Breakdown of health infrastructure Relapsing fever southern Sudan 1998 Marburg hemorrhagic fever in Angola

2005 Hospital records and death certificates

to verify death often absent; alternative verbal autopsy

No baseline data Especially in chronic emergency settings

How to attribute indirect deaths to indirect impacts of conflict

How to estimate mortality in a conflict situation

Retrospective mortality surveys

Prospective surveillance

Analysis of multiple data sources

**Combination of above

Retrospective Mortality Surveys

The HH is the sampling unit; many of the methods we’ve discussed Simple random sampling Systemic random sampling Cluster sampling—geographic

similarity

Usually try to differentiate violent vs non-violent

As with all surveys, bias; survivor and recall

Retrospective Mortality Surveys: Examples

War and mortality in Kosovo, 1998–99: an epidemiological testimony

Paul B Spiegel, Peter Salama

Two stage cluster survey of 1197 HHs, 8605 people

67 (64%) of 105 deaths from war-related trauma corresponding to 12,000 deaths in total population

CMR increased 2.3 times from pre-conflict

Highest age-specific mortality men 15-49, and older than 50

Retrospective Mortality Surveys: Examples

Mortality after the 2003 invasion of Iraq: a cross-sectional cluster sample survey

Gilbert Burnham, Riyadh Lafta, Shannon Doocy, Les Roberts

Cross-sectional cluster survey of 1849 HHs and 12, 801 people

601,027 (92%) of 654,965 deaths violent

CMR 2.6 times pre-invasion levels

Most violent deaths men aged 15-59

Most non-violent persons > 60 yo, children, women 15-59 yo

Retrospective Mortality Surveys: Examples

Mortality in the Democratic Republic of Congo: a nationwide survey

Benjamin Coghlan, Richard J Brennan, Pascal Ngoy, David Dofara, Brad Otto, Mark Clements, Tony Stewart

Stratified three-stage cluster survey of 19,500 HHs

CMR of 2.1 deaths/1000 40% higher than the sub-Saharan region

38,000 excess deaths/month

Most deaths due to preventable causes, malnutrition and IDs

Prospective Mortality Surveillance

Through HIS, targeting health facilities and death registries

Superior, but usually HIS is weak

Ad-hoc systems can be put in place by NGOs, etc, but often not a priority in humanitarian response

Refugee settings—UNHCRs passive HIS

Cause of death, like RMS, difficult

INDEPTH: sub-national demographic surveillance systems (19 developing countries)

Analysis of Multiple Data Sources

Reconstruction of mortality profiles using sources of statistics before, during, and after conflict by statisticians and demographers Prospective surveillance and

previously-conducted RMS Database of human rights violations

and interviews Census of public graves Exhumations Missing persons from HHs in surveys Health surveys and census data

Focus on the quality of the data sources

Method Setting Advantage Disadvantage

RMS During and post -Useful for rapid assessment-Doesn’t require pop denominator-Practical in disorganized settings

-Logistics/security-Bias-Not real time-Statistically complicated

Prosp Surveil During and post -Real time-Simple analytical procedures

-Weak HISRequires regular updating of data and pop size-Camps, stable populations

Analysis of multiple data sources

Mainly post -Used to assess quality/strengths of multiple data sources-Statistical tech to employ best aspects of data sources

-Dep on quality, availability, timeliness of primary data sources

Politically controversial?