Page 1 of 1

Social Distancing Effectiveness and Improved Therapies

Posted: Thu Jun 11, 2020 1:09 am
by erc4003
I'm preparing to do a summer program with high school students, and look forward to using SimPandemic with them. I'm trying to explore lots of scenarios in preparation. So far, I've gotten two unexpected results, and I was hoping to ask if these are due to simplifications in the simulation parameters (I fully understand that these are needed for computational efficiency!) or if I'm simply failing to understand the results.

First, I was exploring the impact of very effective social distancing for the elderly, and I expected that this would cause reduced infection in the elderly and therefore a drop in the death rate. However, I did not see that. To ensure that this was not due to chance with small values, I compared social distancing effectiveness of 50 vs. 99% for the elderly bucket (75% effectiveness for all others), with a 90% death rate for that bucket (all other parameters set at default values). In both cases, death rate was very high (about 2300). The only explanation I could figure out was that there is not a direct, bucket-specific link between infection probability and social distancing effectiveness.

Second, the impact of improved therapies appears to be independent of hospitalization (i.e., improved therapies have the same impact whether the hospitals are above capacity or not). Is that correct?

This is a really great tool--thank you for making it available!

Re: Social Distancing Effectiveness and Improved Therapies

Posted: Fri Jun 12, 2020 2:51 pm
by KenHess
Hi erc4003,

1) I have verified that Social Distancing and the infection rate are working properly, bucket by bucket.

If you do a Save/Share in SimPandemic, copy the link and send it to (put my name in the note you send), I can take a look and give you a better answer. Plus, we would be interested in learning how you plan to use SimPandemic with your students.

Without seeing your simulations, I think that you may have two things going on.

With the default settings for the U.S., only 13.1% of simulants are in the Elderly bucket. If you isolate an intervention strategy on them alone (a logical thing to do given the nature of the disease), by definition you are working with a small population and you will see a lot of variation between runs.

If you are seeing thousands of deaths and changes to settings don’t seem to be making a difference, then there is probably something set such that “the killing power” of the pandemic is simply swamping your other changes. For example, if the duration of Social Distancing is too short, you can have a situation where most simulations result in a full-blown pandemic.

2) You are correct, improved therapies reduce the death rate in the same amount, regardless of whether hospitals are above capacity or not. Obviously, this is a simplification of the real world.