A method for small-area estimation of population mortality in settings affected by crises

Area of Science:

• Demography
• Public Health
• Epidemiology

Background:

• Demographic surveillance is inadequate in crisis-affected populations (conflict, disasters).
• This data gap hinders accurate estimation of population mortality during crises.
• Lack of evidence impedes effective humanitarian response and conflict resolution.

Purpose of the Study:

• To introduce a small-area estimation method for quantifying crisis mortality.
• To estimate both total and excess (crisis-attributable) death rates and tolls.
• To provide granular geographic and temporal mortality data.

Main Methods:

• Utilizes existing data: mortality surveys, displacement-adjusted population denominators, predictor variables.
• Employs a six-step sequential implementation process.
• Applies the method using generic analysis scripts, demonstrated in Somalia, South Sudan, and northeast Nigeria.

Main Results:

• Descriptive analysis reveals patterns in mortality causes (e.g., injuries) and migration.
• Predictive mixed-effects models for crude and under-5 death rates were developed and validated.
• Counterfactual and excess mortality estimates were derived using crisis absence predictor assumptions.

Conclusions:

• The method allows retrospective estimation of crisis-attributable mortality with geographic and temporal detail.
• Enhances understanding and historical record of public health impacts of crises.
• Highlights limitations and areas for future methodological development.