Mapping the risk and distribution of epidemics in the WHO African Region: a technical report.

Disease epidemics result in substantial ill health and loss of lives and therefore pose a threat to global health security, undermine socio-economic lives and destabilize societies. Disease surveillance is a critical component in detecting and effectively responding to epidemics to minimize loss of live and their disruptive consequences. Carefully assembled surveillance data at the highest possible spatial resolutions also permit the understanding of the burden of epidemics, their co-occurrence and the key biological, ecological, economic, health system and governance determinants. It is for this purpose that the WHO-AFRO has commissioned this report. The overarching objective was to develop a comprehensive spatially defined database of outbreaks and epidemics and delineate the ecological zones of diseases that are classified as Public Health Emergency of International Concern (PHEIC) according the International Health Regulations (IHR) 2005 and malaria. The main tasks included the assembly of an inventory of all epidemics reported in Africa from 1970-2016 characterised by date of occurrence, length of epidemic and magnitude and district of occurrence; definition of ecological zones of PHEIC diseases and malaria; the assembly of data on important socio-economic, health systems and environmental correlate and an basic statistical analysis of their relationship with occurrence and frequency of epidemics and outbreaks. Through this study, several products have been developed including time series graphs of outbreak and epidemic occurrence by country, maps of the ecologies of the PHEIC diseases and epidemics, the distribution of these diseases by district. Overall, over 1730 outbreaks/epidemics have been reported in the WHO African region in the period 1970 to 2016. Because the outbreak/epidemic thresholds of the different diseases vary and the actual case data is incomplete, it is difficult to compare which diseases are most prevalent or pervasive. However, in terms of frequency of events, cholera, the arboviruses, measles and meningitis rank the highest. Of the nearly 5250 administrative 2 units analysed in this study, almost each one has reported some form of a disease outbreak in the period 1970-2016, with cholera being the most geographically widespread. The resulting databases are spatially defined and should serve as the basis of subnational inventory of disease outbreaks and epidemics in the region. Several challenges were encountered in the process of implementing this exercise: There is limited information on the definitions and thresholds of outbreaks used over time and how this may have been affected by changing diagnostics and case definitions. This results in uncertainties in the temporal comparison of data. There were inconsistencies between the major databases used in this report in terms of occurrence and magnitude of outbreaks. Agreement across more than two databases supported with literature review and consultation with the WHO AFRO and HQ teams helped with data verification. Poor access to national surveillance bulletins, which are an important source of original outbreak and epidemics data, was a challenge to the data verification process and the development of online portals of national surveillance reports is critical. There is limited data before 1980, most likely as a consequence of poor reporting or archiving of outbreak and epidemic data in libraries outside of Africa. In contrast, there is a greater frequency of reported outbreaks in the last 15 years as diagnosis and surveillance improved. For these reasons, extreme caution must be exercised when interpreting the trends in disease outbreaks and epidemics in Africa. The description of the location of outbreak and epidemic events were variable. In some cases the name of village, town or district were reported. In others, the spatial definition of data was at regional or country levels. The aim of this project was to harmonize these data was to define events by district and where this was not possible, the data were reflected in the time series graphs but are not shown on the district maps. Most of the potential correlates of epidemics, particularly those on socio-economic and health system development, are rarely available by district restricting the analysis to national level. The databases and maps produced in this report should be considered as the foundation for tracking epidemics sub-nationally within the WHO African region. However, they require continuous verification, improvements in spatial resolution and regular updating. Finally, these data can also serve as the inputs for disease specific risk and vulnerability analysis. For example, the simple analysis in this report of the correlates shows that urbanization, gross domestic product (GDP) per capita, percentage GDP expenditure on health, human development index, global hunger index, conflicts, El Nino occurrence and forest cover seemed to correlate with patterns of epidemics. Combining such correlates with the detailed district level outbreak and epidemic data developed under this report could be the basis of further analysis of the assessment of risk of and vulnerability to the PHEICs.