There has been much discussion recently about herd immunity, and the impact of heterogeneity on herd immunity. A thread. (1/)
First, I want to reiterate the point made so articulately by @nataliexdean and @CT_Bergstrom: Natural herd immunity is not a goal we should be striving for. Trying to reach the end of this pandemic via that path will lead to mass deaths. (2/)
nytimes.com/2020/05/01/opi…
nytimes.com/2020/05/01/opi…
But understanding heterogeneity in COVID-19 transmission is important as it'll help to optimize social distancing and (eventually) vacc strategies. A recent thread by @mlipsitch also discusses how heterogeneity may reduce herd immunity thresholds. (3/)
It's true that if there is variation in the # of contacts in a popn, fewer will need to be infected before the epidemic slows down. For example, in a population with a geometric degree distribution, fewer than 40% would have to be infected before herd immunity is reached. (4/) 
However, if there is little variation in a population's contact structure, the herd immunity threshold will be HIGHER than (1-1/R0). This is true for networks with a Poisson degree distribution, small-world networks, or where everyone is equally connected (regular network). (5/)
These ideas are discussed in our past work: bansallab.com/publications/f… (with @TheFerrariLab), bansallab.com/publications/i…, and bansallab.com/publications/f…, all with @meyerslab. I want to highlight two points from this work relevant to our current situation... (6/)
A) When it comes to respiratory infections, we don't expect networks to be particularly heterogeneous (unlike networks relevant to STDs). Respiratory diseases spread via shared space and air and require fewer specialized behaviors to spread. (7/)
bansallab.com/publications/i…
bansallab.com/publications/i…
B)These results also do not mean we should aim to make our networks more variable! Transmission over heterogeneous nets is explosive. In popns with variability, pathogens preferentially seek out those of highest infection risk, spread fast + can severely strain health systems (8/
Infection spread is indeed not random, and heterogeneity should be on our mind. But let's not take our eyes off the prize of reducing overall connectivity and transmission with social distancing, handwashing and masks.
Just found a great thread by @nataliexdean making similar points:
