Read on Twitter
And building out from this work with @djwatts, here’s a set of slides and a couple of hours of recorded lectures from my course @pocsvox.
Slides (megabyteful, acknowledged):
“Contagion-at-large and biological contagion—The spreading of bad things”
http://www.uvm.edu/pdodds/teaching/courses/pocs/slides/biological-contagion/ https://twitter.com/duncanjwatts/status/1232673602006241281
Slides (megabyteful, acknowledged):
“Contagion-at-large and biological contagion—The spreading of bad things”
http://www.uvm.edu/pdodds/teaching/courses/pocs/slides/biological-contagion/ https://twitter.com/duncanjwatts/status/1232673602006241281
The tail of this thread has a mercifully brief synopsis.
Iceland is involved, as an accidental Petri dish.
Iceland is involved, as an accidental Petri dish.
Episode 24a: Contagion—the unpredictability of pandemics
We contend with the confusing usage of contagion for disease and social spreading, walk through the SIR model, wonder why it's still used, and examine the unpredictability of real pandemics. http://www.uvm.edu/pdodds/teaching/courses/pocs/episodes/24a/
We contend with the confusing usage of contagion for disease and social spreading, walk through the SIR model, wonder why it's still used, and examine the unpredictability of real pandemics. http://www.uvm.edu/pdodds/teaching/courses/pocs/episodes/24a/
Episode 25a: Modeling pandemics, then on to social contagion
We show how a toy metapopulation model produces extremely unpredictable pandemic sizes, spurn the reproduction number, point to full-scale models, and more. http://www.uvm.edu/pdodds/teaching/courses/pocs/episodes/25a/
We show how a toy metapopulation model produces extremely unpredictable pandemic sizes, spurn the reproduction number, point to full-scale models, and more. http://www.uvm.edu/pdodds/teaching/courses/pocs/episodes/25a/
The start of these slides and lectures try to address the horror that is the pandemic throughout history.
We examined over 100 years of data from Iceland on contagious diseases, and found that once things took off, the final size was unpredictable. Very different to other kinds of disasters.
Spoiler: R0 (R-zero or R-nought) by itself is of no great use at all in prediction.
Spoiler: R0 (R-zero or R-nought) by itself is of no great use at all in prediction.
The timeline for SARS was not simple (below is from our paper; numbers may have been revised).
Apparent resurgence of diseases results from movements of infected individuals to new susceptible populations.
Modern travel makes this possible globally.
We’re seeing it happen.
Apparent resurgence of diseases results from movements of infected individuals to new susceptible populations.
Modern travel makes this possible globally.
We’re seeing it happen.
We made a simple model which added group-group movement to now age-old, simple disease spreading models.
We were able to generate unpredictability reminiscent of Iceland from uncommon movement of people.
(Very different R0’s didn’t change the story.)
We were able to generate unpredictability reminiscent of Iceland from uncommon movement of people.
(Very different R0’s didn’t change the story.)
Like SARS, disease caseloads in our model showed resurgence on top of unpredictability:
The main takeaway, which will seem unsurprising:
Restrict travel.
Restrict travel.
And of course we know this is not a simple thing.
Limiting travel is going to be hard, there are potentially massive economic impacts, and, all along, it’s a wicked Person Who Cried Wolf situation.
Limiting travel is going to be hard, there are potentially massive economic impacts, and, all along, it’s a wicked Person Who Cried Wolf situation.
And for a further non-Sherlockian insight, governments and people need to always be prepared.
But perhaps to be less Watsonian: Be prepared to be prepared.
But perhaps to be less Watsonian: Be prepared to be prepared.
