Abstract
BACKGROUND: Estimating mortality in conflict settings is essential for effective humanitarian response yet remains fraught with methodological, political, and ethical challenges. Conventional techniques such as retrospective surveys, prospective surveillance, and multiple-source analysis, often collapse under insecurity, displacement, or censorship. Emerging approaches, including key-informant reporting, small-area estimation (SAE), satellite-based remote sensing, and social-media analytics, offer new possibilities but raise questions about bias, data ownership, and the ethics of digital surveillance. This study examined expert perspectives on how these methods are applied, adapted, and constrained in conflict-affected contexts. METHODS: This qualitative study explored expert perspectives on mortality estimation in conflict-affected contexts. Twelve semi-structured interviews were conducted with humanitarian workers, public health specialists, and officials with direct experience in mortality estimation. The data were analysed thematically, following Braun and Clarke’s six-phase framework, with codes reviewed collaboratively for consistency. Reporting adheres to the COREQ checklist. RESULTS: Participants agreed that no single method is universally reliable. Ground-based approaches such as key-informant reporting were valued for their speed and community access but lost validity during prolonged displacement. SAE was regarded as statistically robust yet dependent on high-quality baseline data and specialised expertise. Non-ground-based approaches, including satellite imagery and social-media analytics, expanded visibility in inaccessible areas but introduced ethical and technical limitations related to surveillance, data ownership, and digital exclusion. Political interference, displacement-related logistical constraints, and community trust emerged as prominent cross-cutting considerations across participant accounts. These operational barriers reflect systemic weaknesses in public-health information and coordination structures during emergencies, underscoring the need for integrated surveillance frameworks across humanitarian and national health sectors. Differences also reflected conflict typology: protracted conflicts were dominated by political manipulation, whereas acute emergencies faced infrastructure collapse. Trust, coordination, and community engagement consistently enabled higher-quality data. CONCLUSIONS: Mortality estimation in armed conflict is not purely a technical task but also a political and ethical process. Multiple methods and context-sensitive strategies, combined with the transparent governance of digital tools, are needed to generate credible data that can inform humanitarian decision-making and advocacy. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12889-026-26831-5.