Paul B. Spiegel

During the past decade, indicators for the assessment, monitoring, and evaluation of services provided by humanitarian organizations to populations affected by complex emergencies (CEs) were developed to improve the effectiveness and accountability of humanitarian response. The quality of data used to develop individual indicators and their relationship to positive health outcomes varies greatly. This article states the essential characteristics necessary for the development and implementation of effective indicators in CE response and proposes the establishment of an evidence-based grading system. The importance of trend analysis and the modification or addition of various indicators and their thresholds, according to phase and location of CEs, are stressed. Limitations in the development, implementation, and interpretation of these indicators, including those outside of the organizations' control are discussed. More evidence-based research is needed as to the type and thresholds of indicators that lead to improved health outcomes in populations affected by CEs. The use of indicators by non-governmental organizations, and how they affect their program's decision-making in different phases and settings within CEs need further study. Finally, the establishment of a regulating body with the authority to enforce the attainment of standards by use of these indicators is necessary to avoid inappropriate humanitarian assistance causing loss of life in the future.

For humanitarian organisations, accurate data are essential to identify emerging health problems and determine programme needs. We visited 45 post‐emergency phase displaced persons camps and collected three months' mortality data which we compared with organisations' routine mortality reports. Organisations reported 612 deaths and we identified 741 deaths, for a mortality‐reporting ratio, defined as the number of organisation‐reported deaths divided by the number of investigator‐identified deaths, of 83 per cent. For the majority of camps which under‐reported deaths, mortality reporting ratios were significantly higher for women than men, and for camps with central mortality registers rather than those without. In the few camps which over‐reported deaths, these occurred primarily among children younger than five years of age, probably due to the inclusion of abortions and stillbirths. Despite the overall under‐reporting of deaths by humanitarian organisations, the existing health information systems appear to estimate mortality rates adequately in these post‐emergency camps. However, organisations should improve the precision and completeness with which they report the characteristics of deaths in order to provide valuable data to target their programmes at the most vulnerable people.