Civilian family members' deaths and perceived causes in war-torn Tigray, Ethiopia: modified poisson regression

Civilians suffer greatly from mortality as a result of armed conflicts.The magnitude of family members' deaths and their perceived causes in households in war-affected settings of Tigray region of Ethiopia has not been investigated.

This study was part of a large integrated cross-sectional survey conducted in selected 13,915 households from 19 districts.The data analysis was done using STATA 17 versions software packages.We have fitted all the possible count regression models to discover the best fitted model. Log-likelihood, Akaike Information Criteria (AIC), and Bayesian Information Criteria (BIC) were also used to compare various candidate models. We found that the robust Poisson regression model better fits the mortality data. Adjusted Incidence Rate Ratio (AIRR) with 95% confidence intervals (CI) at P < 0.05 was used to establish the statistical significance in multivariable robust Poisson regression to assess the association between independent variables and the out come variable.

In this study, from the total 13,915 participants, 5% (n = 699) of them experienced civilian family members' war-associated death.Households with a disabled member (AIRR = 2.46; 95%CI: 1.80, 3.39), displaced from home (AIRR = 0.60; 95%CI: 0.50, 0.74,five and more family size (AIRR = 1.09; 95%CI: 1.04-1.14) and widowed marital status (AIRR = 3.72; 95%CI: 2.89-4.81),, and semi-urban place of residence (AIRR = 0.49; 95CI: 0.30, 0.82) were associated with the outcome variable .

Civilian family members' death report in this study is high. Large family size, widowed marital status, not displacing from home, having a disabled member of the family, and rural residences were the identified factors associated with civilian's death. In the context of the cessation of hostilities following the Pretoria Peace Agreement, sustained efforts from the government, humanitarian organizations, and the international community are essential .