Spatial statistics can be used to monitor tropical disease spread and this session featured three experts in this area.
Emanuele Giorgi (pictured) has been studying ways in which to track lymphatic filariasis, a disease that affects the limbs in populations across Africa and Asia, mainly transmitted by mosquito. There are two ways of counting the microfilariae in individuals; the first is more accurate but can only be done at night as they are nocturnal; the second diagnostic does not need to be done at night but it does not give an overall count. Emanuele explained how a bivariate geostatistical model could be used to get a more accurate count using the second diagnostic given the the count found in a sample of individuals.
Dr Ewan Cameron from the University of Oxford talked about his work mapping progress in malaria control with spatial statistics. He explained that because malaria is essentially a disease of poverty, the health systems where it is still prevalent often do not have good surveillance data. So national and small-scale prevalence surveys are used to map estimates of disease burden. However, as malaria declines, more sensitive diagnostics are required which can be more expensive. Dr Cameron described how he developed a model to embed serological data within a geostatistical framework on a national scale to produce maps to inform policy.
Victor Alegana of the University of Southampton talked about advances in mapping malaria for elimination by fine resolution modeling of plasmodium falciparum, the parasite that causes malaria in humans. He spoke of the challenges identifying ‘hotspots’ in countries almost at the point of eliminating malaria. Many countries hover around the boundaries of elimination but because low prevalence is not detected they rise up again. Robust cartography is needed to map the prevalence of malaria, as hand-drawn maps used previously have been found to be unreliable.
Alegana described his work developing ways of moving between data sets to get a reliable fine-scale map. This can now be used to target interventions in detected ‘hotspot’ areas, improve estimates of clinical burdens in low malaria areas, providing a more cost effective solution, especially in sparsely populated areas.