Modelling disease persistence and elimination in low-transmission settings

Malaria and human African trypanosomiasis are two vector-borne diseases caused by protozoa. In Zanzibar, Tanzania, malaria persists at a low prevalence despite the implementation of vector control, passive surveillance, and reactive surveillance. Human movement, and subsequent case importation, is hypothesised to be a key driver of persistence. Human African trypanosomiasis prevalence is below 1000 globally and is targeted for elimination. However, treatment options were previously limited and the diagnosis process is invasive and painful. The introduction of a new oral treatment improves options for treatment but may have an impact on transmission, as compliance levels may be lower than with previous treatments. Stochastic metapopulation models of disease transmission were used to explore a range of questions regarding malaria and human African trypanosomiasis transmission and elimination. A metapopulation model of malaria transmission was developed and parameterised to data from Zanzibar, Tanzania. It incorporated human movement and reactive case detection, and was used to investigate the impact of improvements to reactive case detection or treatment of imported cases on prevalence levels. The model was then expanded to include separate categories for imported, introduced, and indigenous cases, which allowed us to apply the WHO definition of malaria elimination (three years with no indigenous cases). An already established model of human African trypanosomiasis transmission was adapted to incorporate treatment by fexinidazole and potential non-compliance. This was used to test the potential impact of widespread versus limited access to fexinidazole under a range of compliance scenarios, and the potential impact of increased treatment seeking rates. The controlled reproduction number for malaria was estimated to be below the threshold value of 1 on both major islands of Zanzibar, confirming that importation is driving disease persistence. Reactive case detection is estimated to reduce malaria incidence by approximately 10% on Zanzibar. To achieve non-zero probabilities of elimination, infections in travellers need to be targeted, and onward transmission from imported cases needs to be reduced. Considering human African trypanosomiasis transmission, an increase of 20% in the passive detection rate is expected to counter a small negative impact of non-compliance to fexinidazole. While reactive case detection is useful for surveillance and does reduce malaria incidence, the large number of low parasite density infections prevents reactive case detection from removing large parts of the parasite reservoir. The controlled reproduction number needs to be kept well below 1 in order to minimise the chances of any imported cases leading to chains of transmission that lead to indigenous cases. A better understanding of treatment compliance with fexinidazole and changes in treatment seeking behaviour is necessary to better estimate the potential impact of fexinidazole on human African trypanosomiasis transmission in the Democratic Republic of the Congo.