In humanitarian crises, quantifying the number of victims contributes to estimating the needs for assistance, advocating for additional resources, promoting diplomatic actions, supporting transnational justice, and informing political decisions. It also provides a clearer understanding of the severity of a crisis within its historical, geographical, political, and social contexts. However, data collection in conflict settings is frequently hindered by insecurity and political barriers.Relevant studies of war mortality have used epidemiological methods, such as prospective community surveillance, body count, retrospective household surveys, capture-recapture analysis, and key informant interviews. The present paper describes the advantages and limitations of each of these approaches. Subsequently, it summarises and tabulates estimates of mortality indicators in 19 humanitarian crises taking place in African countries, Syria, Iraq, and Yemen over the last twenty years. Most estimates were based on household surveys. Attention is brought to the difficulties in the assessment of indirect war-related mortality.Mortality indicators are politically sensitive. As a result, politicisation of numbers in armed conflicts has not been uncommon: cases of data manipulation to serve political goals are described.Improvements in humanitarian practice and public health interventions have contributed to the decline of reported mortality rates in recent years. However, challenges in measurement and data standardisation still need to be addressed. Public health information remains an underfunded and under-prioritised part of humanitarian assistance, despite the growing emphasis by donor governments on the importance of quantitative evidence to inform decision-making and justify aid budgets.