Understanding the impact of socioeconomic status on Ebola transmission08 Jan 2016
I was excited to see this recent article by Mosoka Fallah and coauthors in PLOS Negected Tropical Diseases on the impact of socioeconomic status (SES) on transmission rates in the recent Ebola outbreak in West Africa.
Although there is an increasing appreciation of the role of social inequality in infectious disease dynamics, these have been slow to find their way into dynamic modeling applications like the one presented in the Fallah et al. PLOS NTD paper.
Unfortunately, the paper is a bit light on modeling specifics and figures that might make it a bit easier to understand the results. But what it shows is that transmission rates in low-SES areas of the county they studied in Liberia were as much as 3.5 higher than in high-SES areas.
(The model presented in the PLOS NTD paper is ostensibly based on the one presented in this earlier modeling analysis in the Annals of Internal Medicine)
A particularly interesting finding from this analysis is that transmission from low-SES into high-SES areas is much more likely to occur than is transmission from high-SES to low-SES areas. The authors suggest that this is likely the result of employment-related patterns of movement from poorer to wealthier areas.
These findings echo interesting work from David Dowdy & co on the potential regional impacts of hotspots of Tuberculosis transmission located in low-SES areas of Rio de Janeiro.
Although I appreciate the argument implied by these findings, that targeting prevention on the poorest individuals at greatest risk may also have population-level protective effects, and have used it in my own work, it’s also kind of depressing that this is a necessary adjunct to the basic argument that these interventions are necessary purely on the basis of their direct protective effects.
One final question about this work is whether it would have been more useful for the authors to predict infection rates as a function of specific social covariates (i.e., population density, household environmental quality) using a log-linear model rather than by using these factors to construct a discrete SES measure with low, medium, and high-SES categories. This may have provided a bit more insight into the specific SES-related factors that were most important in driving infection heterogeneity.
Nevertheless, this is a welcome and very useful addition to the slowly-growing set of sociologically-informed analyses of infectious disease transmission.