Exploring the epidemiology of malaria and the impact of malaria control interventions in malaria-endemic and Ebola-epidemic West Africa

Despite significant advances in the number and type of control measures available, malaria remains one of the leading causes of death worldwide, with the majority of burden concentrated in sub-Saharan Africa. Long-lasting insecticide treated bed nets, antimalarial treatment using artemisinin combination therapies, mass drug administrations, indoor residual spraying and seasonal malaria chemoprevention, are used individually and in combination, supported by community education programs and early detection and treatment protocols. Despite these advances, little evidence exists on how to utilize these interventions effectively in hyperendemic settings or during emergencies. This thesis focuses on malaria control and surveillance in West Africa, particularly Guéckédou Prefecture, the Republic of Guinea and Monrovia, Liberia. Issues that malaria control programs face and the impact that such programs can have in hyperendemic settings and other challenging environments, specifically, the West Africa Ebola epidemic are explored. The evidence presented here builds a case for placing a stronger emphasis on implementing and sustaining control measures in areas of hyperendemicity. Additionally, the need to develop alternative strategies for managing the burden of malaria in both hyperendemic settings and during outbreaks is emphasized. A multi-component malaria control intervention that was implemented in program conditions from 2011-2014 in Guéckédou Prefecture. In Chapter 4, both intervention coverage and the impact of the malaria control intervention on malaria parasite prevalence are quantified using data from biannual population based cross-sectional surveys. Over time, intervention coverage increased while rapid diagnostic test confirmed malaria parasite prevalence decreased in areas where the control activities were implemented. Yet, in the comparison area where activities were not implemented there was no significant change. Nevertheless, while the decrease in malaria parasite prevalence measured during the period of intervention was encouraging, the overall decline was relatively small and suggests a need to develop new or modify currently available control strategies in order to have a greater impact on malaria burden in similar areas. During the study period, the area of intervention became the initial epicenter of the 2013-2016 Ebola Virus Disease (EVD) outbreak. Malaria control activities were reduced because of the outbreak in Guéckédou and all activities related to the multi-component malaria control intervention ceased. In Chapter 5 data collected from the cross-sectional surveys was used to parameterize a stochastic malaria transmission model to assess the impact of the sudden cessation of malaria activities in this context. The model predicted the monthly incidence of malaria cases according to two scenarios, i) a counterfactual scenario that assumes no reinforced malaria interventions occurred between 2011 and 2014, and ii) a scenario with reinforced malaria interventions that ceased at the start of the Ebola outbreak. Interrupted time series analysis was used to assess the impact of malaria control activity cessation on malaria incidence from April 2014. The incidence of uncomplicated malaria was estimated to have resurged to levels higher than that predicted to have occurred in the counterfactual scenario within 8 months of activity cessation in Guéckédou (April 2014). The models show that gains made in malaria control are not sustained and resurgence becomes a significant risk. In areas where malaria is highly endemic, advances made during control activity implementation are quickly negated when activities stop. Upon arrival in Guéckédou in 2010, data from health facility based surveillance was deemed unreliable due to underreporting. Consequently, concurrent with the cross-sectional surveys, community-based sentinel site mortality surveillance was implemented in the same areas in order to monitor malaria attributable mortality. As described in Chapter 6, data on mortality from 43,000 individuals under surveillance was collected for 36 months. The ability of the surveillance system to capture mortality, health-seeking behavior and quantify malaria attributable mortality is described. No early warning system was in place prior to the Ebola outbreak, consequently it was detected 4 months after it began. Data collected through the community-based mortality surveillance system was evaluated retrospectively for its ability to detect outbreaks, specifically of Ebola, when adapted to syndromes. Indeed, two of the suspect Ebola deaths captured through the surveillance system were among the first laboratory confirmed cases from the 2013-2016 outbreak. Although challenging, this demonstrates that prospective community-based mortality surveillance using sentinel sites can provide a means to document mortality and facilitate outbreak detection in low resource settings. As the Ebola outbreak evolved into a multi-country epidemic, the response moved from vertical programming to a more holistic response that incorporated the innovative use of classic malaria control strategies. One of these strategies included the first mass drug administration (MDA) of malaria chemoprevention carried out during an Ebola outbreak. In post distribution surveys of individuals attending the distributions reported in Chapter 7, the incidence of self-reported fever decreased from 4.2% in the month prior to the first distribution to 1.5% after the first distribution. Yet, only 52% of household members initiated treatment after round 1 and only 22% after round 2. While the reduction in self-reported fever cases suggests that MDA may be effective in reducing cases of fever during Ebola outbreaks, the low treatment initiation suggests the need for longer-term interventions to prevent malaria and to improve access to healthcare. All of these different aspects combined provide a unique perspective on malaria control in normal and emergency settings in malaria endemic areas of West Africa. Malaria control programs implemented in hyperendemic settings in program conditions can result in a decrease in malaria parasite prevalence and malaria attributable mortality. Yet, malaria morbidity can be expected to resurge quite quickly if activities are not sustained. In areas where health facility surveillance is weak, community-based surveillance can be implemented to both capture malaria related mortality and detect outbreaks. Finally, in order to mitigate the mortality that is inherent with both malaria and Ebola infections, particularly during outbreaks, the ability to differentiate between the two (in light of their common features) need to be improved. These lessons need to be translated into improved surveillance and response strategies in order to detect and respond to both diseases, potentially resulting in a synergistic decrease in mortality