09-30-2020, 12:36 PM
I think such numbers are silly when applied to very diverse conditions. For instance, is 4.5 appropriate for agrarian areas where there is little contact, and equally appropriate in the mega-dense, pack-them-into-mass-transit, cities? Coming up with a number for the whole world is more dependent upon the mixture of population densities & interpersonal distances than of the disease. (Is the elephant more like tree trunk or a snake?)
To me, a better number to estimate would be something like the probability of droplet/aerosol spread to someone with whom an infected person spends 15 minutes within 6' in an indoor situation without mask on each day of the illness. Then get the same number for when one of them is wearing a mask and when both are. Then the same number for outdoor conditions..
Call such a number E (exposures) for an individual per day. E_0, E_1, E_2 (for 0,1,2 masks). If you're within 6' of 3 people for 5 minutes (or 45 people for 20 seconds), that's the same as being within 6' of 1 person for 15 minutes. (I don't think this is the best possible measure, but these are the units that are being used for "close contact") Compute the average E_n for a population, using surveys to tell you what percentage of exposures are 0 mask, 1 mask, or 2 mask. Now calculate the number of new infections, and then the number of infections per 1K exposures for infected persons. Shutdowns drive exposures toward 0. Mask usage drives E_0 down without driving down the total exposures.
Suppose E_0 was 10 for some population. If 100 of that population was sick on any day, that's a total of 1000 infected exposures.
Something like the contact tracing apps that detect the nearby presence of others could help in estimating the values of the sum of E (but don't tell you about masks).
I would expect that infections per 1K E_0 (or E_1 or E_2) infected exposures should be roughly constant across populations and over time. Recognize that the number of exposures would be dependent upon the weather. (I could be wrong about it being constant. Things like temperature, sunlight (both on the exposure, and as something that impacts Vit. D levels) could certainly impact such a measure.)
Then you can integrate that with estimates of exposures in different societies and living conditions to estimate the rate of spread in rural Utah as well as New York City and Starkville (i.e., college town). You could also then judge whether somewhere like SCC is really doing a good job at reducing transmission compared to other counties, or is it the per-capita infection rate more a function of the combination of the different lifestyles in the county vs the combinations in other counties.
You can also estimate the number of additional exposures from opening up sectors (gyms, indoor restaurants, whatever). If one had a good estimate of new infections daily per daily infected exposures, one could estimate the number of new cases resulting from additional exposures from opening that sector for whatever the current number of infections.
To me, a better number to estimate would be something like the probability of droplet/aerosol spread to someone with whom an infected person spends 15 minutes within 6' in an indoor situation without mask on each day of the illness. Then get the same number for when one of them is wearing a mask and when both are. Then the same number for outdoor conditions..
Call such a number E (exposures) for an individual per day. E_0, E_1, E_2 (for 0,1,2 masks). If you're within 6' of 3 people for 5 minutes (or 45 people for 20 seconds), that's the same as being within 6' of 1 person for 15 minutes. (I don't think this is the best possible measure, but these are the units that are being used for "close contact") Compute the average E_n for a population, using surveys to tell you what percentage of exposures are 0 mask, 1 mask, or 2 mask. Now calculate the number of new infections, and then the number of infections per 1K exposures for infected persons. Shutdowns drive exposures toward 0. Mask usage drives E_0 down without driving down the total exposures.
Suppose E_0 was 10 for some population. If 100 of that population was sick on any day, that's a total of 1000 infected exposures.
Something like the contact tracing apps that detect the nearby presence of others could help in estimating the values of the sum of E (but don't tell you about masks).
I would expect that infections per 1K E_0 (or E_1 or E_2) infected exposures should be roughly constant across populations and over time. Recognize that the number of exposures would be dependent upon the weather. (I could be wrong about it being constant. Things like temperature, sunlight (both on the exposure, and as something that impacts Vit. D levels) could certainly impact such a measure.)
Then you can integrate that with estimates of exposures in different societies and living conditions to estimate the rate of spread in rural Utah as well as New York City and Starkville (i.e., college town). You could also then judge whether somewhere like SCC is really doing a good job at reducing transmission compared to other counties, or is it the per-capita infection rate more a function of the combination of the different lifestyles in the county vs the combinations in other counties.
You can also estimate the number of additional exposures from opening up sectors (gyms, indoor restaurants, whatever). If one had a good estimate of new infections daily per daily infected exposures, one could estimate the number of new cases resulting from additional exposures from opening that sector for whatever the current number of infections.