How Ebola spread during the first few months of the outbreak. R = radial diffusion pattern; H = hierarchical pattern or behavior.
Using the mapping of Ebola provided by this NY Times overview of what has happened, we can see how radial and hierarchical diffusion behaviors effect disease diffusion patterns. This method of modeling also demonstrate how two other historical medical geography teachings–the sequent occupancy interpretation of a place, and transportation modeling–provide insights into predicting the disease pattern.
These events barely integrate any natural elements linked to Ebola diffusion once the disease erupts. These natural features may however remain present and active, although not predominant due to the epidemic outbreak process.
There is also the classical modeling of epidemics in Iceland produced about 50 years ago by British Geographers, that these notes can be related to. We see the Iceland outbreak events re-occurring on the maps provided in this NY Times video.
In medical geography, a important lesson I learned nearly 10 years ago from ESRI’s medical geography statistician was the role of area size in understanding a diffusion pattern.
Spatial analysis requires we repeat certain formulas over different levels of spatial resolution. We then determine at what are size the most sensible and useful prediction model can be developed. That critical spatial feature is defined by the transportation history and patterns of the people involved with an outbreak. Africa’s transportation process defines 30 miles as a limiter, noted in this NY Times video. For Ebola, this relates to when radial diffusion patterns (‘R’ in the above figures) end up being replaced by hierarchical patterns.
The transportation of the Ebola to towns and villages distanced from the nidus or nest is a product of these features, and represent early stage hierarchical diffusion behavior (‘H’ in the above figures). Like cholera and other diseases that erupt in communities due to poor sanitation, lower SES, poor living conditions, there can be this reversed hierarchical diffusion worth noting. The transportation of Ebola to local, regional hospitals is an example of this stage in "gearing up" the epidemic for a major outbreak. Transportation is the link in the change from radial to hierarchical diffusion patterns.
The first outbreak was expected to peak in March or April and then reduce to little or no more disease outbreaks. This didn’t happen however. Instead the early stage hierarchical diffusion processes had already taken place, setting up new regions for perpetuating the outbreak and keeping its organism alive and making the likelihood for continued diffusion possible. Small radial outbreaks ensued, involving neighbors and immediate family, and once again, hierarchical diffusion commenced, this time striking larger towns and cities (next step in the reversed hierarchical diffusion pattern).
As soon as larger cities are hit, and the rural/wilderness setting of Ebola has been left behind, this sets the stage for international travel. Due to the different forms of transportation and their potential speeds and distance of travel, introduction of any disease to a large urban setting makes it possible for hierarchical diffusion to commence to other countries on other continents. Thus Europe was stricken as well as North America, and larger economic centers were impacted.
Some other useful ways to look at this spatial behavior include the application of Christaller’s model, which focuses on economics patterns around major urban centers, and Theissen’s Polygon modeling, which when combined with transportation time-distance spatial relationships can be used to define the most susceptible regions to be attacked once a major urban center is impacted, such as Dallas, New York, Chicago, London, and Hong Kong. (I have some GIS pages posted on each of these two major spatial epidemiology topics.)
Again, it is also important to note that the US medical geographer and economist Gerald F. Pyle defined radial, hierarchical and mixed disease diffusion patterns based on this interpretation of disease spread as a by-product of commercial patterns and economic geography. I added to his model the notion of the reversed hierarchical pattern about 14 years ago, noting the role of SES, social inequity, poverty, and sanitation on disease diffusion patterns (the introduction of cholera into Irish settings in NY and IL, and into Kingston, Jamaica in 1849/50 were my lead examples).
Sequential Occupancy, a late 19th C geographic interpretation of how landuse changes when a region economically develops, relates to this in how it provides insight into transportation related impacts. (Sequential Occupancy is also my recommended replacement for epidemiological transition theory, since it is more precise in defining cause and effect.)
Earlier stages in regional economic development define the local technology, and as a result also impact how a disease can and will diffuse, and in which directions, at what speed. At the tricountry border of Guinea-Liberia-Sierra Leone, the use of early stage transportation (waterways) helped spread the Ebola, as noted in this video (‘R-H transition’ in the figures).
Parts of modeling disease spread are much like modeling water flow patterns along a river or traffic flow along the roadways heading towards urban regions. Spatial analytics makes it possible for specific routes to be predicted and then tested as soon as an outbreak happens.
Traditional epidemiological research focuses primarily on the natural human ecological events, but fails to fully take into account the economic geography like behaviors of disease outbreaks.
Every major transportation of this Ebola to a new location occurred due to hierarchical modelling human behaviors. The Reversed hierarchical diffusion behavior prevailed all throughout this past year of activities. Radial patterns defined mostly the local disease spread. Sequent Occupancy modeling shows us that diseases can follow a 125 year old economic geography way of modeling man and disease due to the lack of much change in human behavioral patterns, spatially or temporally.
This Ebola also behaved much like the Cholera of 1800 to 1873. With each new outbreak, it traveled much further in its diffusion process around the world, following Gerald F. Pyle’s diffusion model.