I've been following the Ebola outbreak since last spring, when news first hit of a West African outbreak. In both math modeling and microbiology courses, I got a good primer of some of the dynamics and infection stages of such diseases, and their unique incubation, infective, hospitalized, and recovered stages. Initially, it appeared that this Ebola outbreak was just another normal outbreak, as seen in prior decades in other parts of Sub-Saharan Africa.
At Williams, there is a requirement for a colloquium by all senior math majors. Given the hype around Ebola virus, my interest in math modeling, and the idea of doing something current rather than presenting mathematical proofs from the 1800s, I decided to design a presentation around Ebola.
A few days later, I sat on the plane next to a woman whose husband had just returned from the front lines of the disease in Liberia (I had to ask very quietly when she said her husband worked for the government in Atlanta). He was not under official quarantine as a CDC employee, but there was definitely that what-if factor that could have thrown someone on a packed 737 into panic mode. Needless to say, our conversation about Ebola left out the story for the rest of the plane ride. And though 21 days later, I show no symptoms of the disease, it presents an interesting case for modeling-- how to best model contacts and communication networks, especially given that Ebola is generally spread by direct contact with an infected person or body. Having an outbreak reach into West Africa is arguably the more interesting problem, and presents a difficult vector-based modeling problem.
Interestingly, the current Ebola virus strain does not appear much different from past strains. And Bayesian and other models of the disease based on past outbreaks place the current outbreak size within a reasonable 99% confidence interval of what could be expected. The lack of foreign mobilization behind Ebola in West Africa makes it difficult to confirm or deny statistics about the current situation as well, presenting difficulties and significant variability in estimating R0 and other infection and mortality rates. This complicates modeling of when the outbreak may end and how large it can get. Then we can add another element of virus research and gene modeling, especially given that primates were successfully cured of closely-related Marburg virus in the last few months.
All of this is really just to get you to think about Ebola, what shortcomings we have in current data, and how those shortcomings may be reflected in actual care and the future of the outbreak. If you care, consider a donation to Medicins Sans Frontieres or another organization with a verified field presence, and perhaps write a quick letter petitioning your legislators to allow US healthcare personnel to travel to West Africa and return home without excessive criticism.
At Williams, there is a requirement for a colloquium by all senior math majors. Given the hype around Ebola virus, my interest in math modeling, and the idea of doing something current rather than presenting mathematical proofs from the 1800s, I decided to design a presentation around Ebola.
A few days later, I sat on the plane next to a woman whose husband had just returned from the front lines of the disease in Liberia (I had to ask very quietly when she said her husband worked for the government in Atlanta). He was not under official quarantine as a CDC employee, but there was definitely that what-if factor that could have thrown someone on a packed 737 into panic mode. Needless to say, our conversation about Ebola left out the story for the rest of the plane ride. And though 21 days later, I show no symptoms of the disease, it presents an interesting case for modeling-- how to best model contacts and communication networks, especially given that Ebola is generally spread by direct contact with an infected person or body. Having an outbreak reach into West Africa is arguably the more interesting problem, and presents a difficult vector-based modeling problem.
Interestingly, the current Ebola virus strain does not appear much different from past strains. And Bayesian and other models of the disease based on past outbreaks place the current outbreak size within a reasonable 99% confidence interval of what could be expected. The lack of foreign mobilization behind Ebola in West Africa makes it difficult to confirm or deny statistics about the current situation as well, presenting difficulties and significant variability in estimating R0 and other infection and mortality rates. This complicates modeling of when the outbreak may end and how large it can get. Then we can add another element of virus research and gene modeling, especially given that primates were successfully cured of closely-related Marburg virus in the last few months.
All of this is really just to get you to think about Ebola, what shortcomings we have in current data, and how those shortcomings may be reflected in actual care and the future of the outbreak. If you care, consider a donation to Medicins Sans Frontieres or another organization with a verified field presence, and perhaps write a quick letter petitioning your legislators to allow US healthcare personnel to travel to West Africa and return home without excessive criticism.
