BACKGROUND                                                                                    Sinka et al., 
                                                                                                           (2020)
             Provides  a  framework  to  investigate  what  factors  will 
             influence public health impact
             Builds  on  Sinka  et  al.,  (2020)  which  found  high  suitability 
             across Africa
             Evidence  that  An.  stephensi  is  playing  a  role  in 
                                 Occurrence of Anopheles        Annual confirmed malaria cases 
             malaria transmission in Djibouti                                                              Sinka et al., 
                                 stephensi in Djibouti City             in Djibouti, MoH
                                                                                                           (2020)
    Seyfarth et al., (2019)
             Can  we  attempt  to  quantify  what  has  potentially 
             happened in Djibouti in order to project what could 
             happen in Ethiopia? (not a forecast or prediction of 
             what will happen)
           METHOD                                                                  Uncertainty in vector bionomics
           Use deterministic malaria model to 
           estimate An.stephensi vector density using 
           Djibouti malaria incidence data
             https://github.com/mrc-ide/deterministic-malaria-model
             Multiple runs to account for uncertainty in mosquito          Current situation varies across the country
             bionomics (A).
                                                                                B) IRS           C) ITN               D) EIP (extrinsic 
           Extrapolate increase in An. stephensi                                coverage         coverage             incubation)
           vector density from Djibouti to Ethiopia to 
           produce predictions of how malaria 
           incidence may change
             Account for pre-existing (B) IRS coverage (C) ITN coverage 
                                                                    Uncertainty in where malaria and species could invade
             (D) temperature dependent EIP and malaria prevalence
           Predicted population at risk (E and F)                                   E) Altitude                F) An. stephensi suitability
           Scale up interventions and apply these to 
           new predictions of malaria transmission
             ITNs (increase to 80% use) for pyrethorid and pyrethroid-
             PBO nets
             IRS (increase to 80% use) long-lasting which mosquitoes 
             susceptible
             Larval Source Management (LSM, 40% reduction in 
             emergence)
                                                                         Prevalence increase by 
         RESULTS                                                         administrative 
                                                                         grouping
        Huge uncertainty around 
        results
        Substantial increases in 
        prevalence across Ethiopia 
        with large amounts of 
        subnational heterogeneity
        Large increases in cases 
        seen in areas with low                                          National increase in malaria cases
        existing transmission
         Low population immunity to malaria and 
         existing vector control
        Increase in incidence 
        depends on population 
        expected to be at risk
         541,000 additional malaria cases 
         per year (95% CI 134,000 – 
         979,000)
         Currently ~740,000 reported cases of 
         malaria per year in Ethiopia (World 
         Malaria report 2020)
       RESULTSSubnational heterogeneity in the impact of interventions
       Considered 
      different 
      combinations of 
      interventions at 
      different 
      coverages 
      (0/80%)
       Pre-existing 
      intervention and 
      transmission 
      important to 
      consider
       Without the use 
      of PBO nets, 
      reduction to pre-
      existing levels of 
      transmission very 
      difficult even with 
      scaleup of 
      ITN/IRS/implemen
      tation of larvicide
               CONCLUSION
               Large parts of Ethiopia are vulnerable to substantial 
               increases in malaria if An. stephensi establishes itself 
               across the country
               Huge uncertainty in estimated impact
               Large scale up of interventions needed following 
               estimated increases
               Additional surveillance and data needed
                 This study is a first step of a long process of estimating the impact of An.                               Acknowledgements
                  stephensi
                 Work estimates large increases, but huge uncertainty around this, in order                                 Seth Irish, Dereje Dengela, Aklilu 
                  to improve estimates more data on vector bionomics                                                         Seyoum, Eric Tongren & Jennifer 
                                                                                                                             Armistead
               Several limitations
                 Many assumptions due to limited data                                                                        Funders/collaborators
                 Single vector considered, no accounting for inter-species competition                                        PMI/Vector-link/Ethiopian National Malaria 
                                                                                                                                Programme
                 Do not consider Plasmodium vivax malaria
                                                                                                                               CEASE group
                 Invasion dynamics are very simple
                                                                                                                                 control the spread of Anopheles stephensi in 
                    No accounting of geographic spread or differential suitability in administrative units, no                   Ethiopia and Sudan
                     seasonality
                    Invasion and establishment in Ethiopia is likely to be very different to Djibouti, but the absence 
                     of data has necessitated assuming it will be the same