The death toll of armed conflict and food insecurity in north-east Nigeria, 2016-2019: a statistical study

BackgroundArmed conflict, compounded by displacement and food insecurity, has affected Adamawa, Borno and Yobe states of northeast Nigeria (population {approx} 12 million) since 2009. The crisis escalated in 2013-2014 and featured a delayed humanitarian response. We wished to estimate the death rate and toll attributable to the crisis. Data availability constraints restricted estimation to the period 2016 to 2019. MethodsWe did a small-area estimation statistical analysis of previously collected data, stratified by local government area and month. We fitted a mixed effects model to household mortality data collected as part of 70 ground surveys implemented by humanitarian actors within the study area and period. Model predictors, selected from a wider candidate list, included livelihood typology, the price of staple cereal, vaccination geo-coverage and the presence of humanitarian actors. To project accurate death tolls, we adjusted population denominators based on internal and refugee displacements. We used the model and population estimates to project mortality under observed conditions and under different assumptions of counterfactual conditions, had there been no crisis, with the difference between observed and counterfactual providing excess mortality. ResultsCrude and under 5 years death rates were highly elevated across most ground surveys, with net negative household migration. Between April 2016 and December 2019, we projected that 490,000 deaths (230,000 children under 5y) died in excess of the counterfactual non-crisis level; this death toll ranged from medians of 90,000 to 550,000 depending on the counterfactual assumptions chosen, specifically whether to consider a Nigeria-wide price inflation phase as inherently part of the crisis in the north-east. Crude and under 5 years death rates were two to three times higher than counterfactual levels, and highest in 2016-2017. DiscussionThis may be the first crisis-wide estimate of mortality attributable to the crisis in north-east Nigeria. Our findings do not reflect acute emergency periods before 2016 and the situation in neighbouring countries affected by the conflict. Sensitivity analysis suggests that results for Borno state, where conflict has been most intense, are subject to under-estimation due to data not representing hard-to-access areas. Further studies to document mortality in this and other crises are needed to guide decision-making and memorialise their human toll.