Malaria control is increasingly being tailored to local needs, this is especially necessary in humanitarian settings where resources are poor. Pyrethroids are the most widely used class of insecticides for mosquito control. Using them effectively requires measuring their epidemiological impact and understanding how this is reduced by the emergence of pyrethroid resistance in mosquitoes. In the first two chapters of this thesis, I consider how we could measure the impact of pyrethroids using the prevalence of infection in pregnant women, a potentially cheaper and more reliable alternative to clinical incidence or prevalence in children. In Chapter 2, I fit a Bayesian regression model to show that the malaria burden measured at hospitals near internally displaced populations is higher than the regional average. In Chapter 3, I demonstrate that the prevalence of infection in pregnant women and the clinical incidence in children change together over time. Collecting routine data from pregnant women seems promising as a measure for assessing malaria burden trends. In the second half of the thesis I explore the impact of different pyrethroid-based interventions in a variety of contexts. In Chapter 4, I expand an existing mathematical model of malaria transmission to predict the impact of distributing emanators (a type of spatial repellent) where insecticidal nets are not commonplace. In Chapter 5, I establish how outdoor evening biting could sustain transmission in places where insecticidal nets are used but residual transmission remains. In Chapter 6, I investigate a sub-lethal effect of pyrethroid bed nets that I call temporary feeding inhibition, mosquitoes that are exposed to pyrethroids do not die but are unable to bite humans for a short while afterwards. Together the work shows how statistical and transmission dynamics models can be used to understand the efficacy of vector control interventions and measure their effectiveness in the field.