malariacontrol.net: Science update part I: till January 2013

It has been a busy year at malariacontrol.net. Much too busy to tell you about the good work you’ve contributed to in just one post. Therefore, this science update comes in three parts, to be published over the next few days. This first post we will talk about some work that was published last fall, looking at how best to estimate the best way to eliminate malaria in low transmission settings. Malaria transmission is governed by many things, but when scientists are talking about transmission, they are generally thinking of the entomological inoculation rate [EIR], that is, the average number of infected mosquito bites a person receives in a year. In some of the worst malarial areas, this number can easily be in the hundreds of bites per year. EIR is generally measured by trapping mosquitos and seeing what percentage of them are infected with malaria and then factoring in the number of bites they give a night. For example, a catch of 20 biting Anopheles per person per night, where 16 are human-fed and 2 of those are infected with malaria sporozoites would correspond to an EIR for that day of 20 x 16/20 x (2/16) 1 = 1.68. Each individual in that area receives an average of 1.68 infective bites per night or an annual EIR 613 - an indication of very high malaria transmission. But when the transmission rate is very low (which is, in and of itself, a good thing), perhaps an EIR of 1 or 2 per year, you would need to trap many more mosquitos to get a reliable estimate of the percentage of them carrying malaria. Further, one should not assume that the overall dynamics of transmission would be the same in these low transmission areas compared to the higher, better studied ones. Erin Stuckey at the Swiss TPH used malariacontrol.net to explore transmission dynamics in a low-transmission setting, the Rachuonyo South highlands above the shores of Lake Victoria in Kenya. One of the reasons we run models is to try to understand which factors have the most impact on outcome of interest (in this case, malaria control). She found that key issues for Rachuonyo were vector biting behaviour, their susceptibility to indoor residual spraying (IRS), and the detection method used for human surveys – all of these affect the impact of interventions in areas with low and/or unstable P. falciparum transmission. Erin also looked at the influence detection method used for surveys on the final estimate of prevalence. To address model sensitivity to the ability of a given test to detect a P. falciparum infection, an experiment was created to mimic the detection limits of a rapid diagnostic test (RDT), polymerase chain reaction (PCR), skilled microscopy, and a low-quality diagnostic such as a poor-quality RDT or unskilled microscopy. The prevalence estimate decreases with higher detection limits, as does the stochasticity of the predictions. This graphic from Erin's paper shows this effect of changing the detection limit (number of parasites per microliter) at which the survey is able to detect P. falciparum infection on the simulated number of P. falciparum infections in a population of 10,000 individuals for

• a) baseline model with a detection limit of 200, equivalent to RDT; b) detection limit of 40, equivalent to PCR; c) detection limit of 100, equivalent to skilled microscopy; and d) detection limit of 500, equivalent to a poor quality diagnostic.

The implication is that if RDTs used in surveys perform poorly, whether the result of low quality manufacturing or improper storage conditions or use, according to simulation results up to half of infected individuals would be misclassified. Decision makers need some kind of guidance on where to best put their efforts at malaria control. We need simulations such as these especially when the field data are sparse. In this case, measuring EIR through mosquito collection may not be the optimal way to define transmission in areas with low, unstable transmission, but simulation results from models such as OpenMalaria can help fill the gap between what we can realistically measure in the field and what we need to know about a given area for malaria control.

More...