tag:blogger.com,1999:blog-582368152716771238.post7306836553494325377..comments2024-03-28T11:14:02.660-05:00Comments on The Grumpy Economist: An SIR model with behaviorJohn H. Cochranehttp://www.blogger.com/profile/04842601651429471525noreply@blogger.comBlogger12125tag:blogger.com,1999:blog-582368152716771238.post-21967628717364829012020-11-20T17:24:11.953-06:002020-11-20T17:24:11.953-06:00In my SEIR model I do a moving time average window...In my SEIR model I do a moving time average window to cases data with a single parameter in the fit which is Rt. I can then compare your betat value to my Rt. I get a nice match up until the recent surge in cases. In other words, when the number of cases is low but rising people can social distance, decrease betat (or Rt) and bring pandemic under control. With the recent surge your betat goes down as infections (or deaths) rise but the fit Rt continues to rise. I'm thinking your model is saying that social distancing works up to a point, but if the cases rise too fast then social distancing can't keep up and the "wave" will run it's course. Thoughts My problem now is I can't find a functional form that describes betat with social distancing at low infection but then grows if the infection is high.Scott Whittenburghttp://scottwhittenburg.comnoreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-63244512352143535752020-11-06T11:42:08.471-06:002020-11-06T11:42:08.471-06:00Finally a well data-driven analysis about this mod...Finally a well data-driven analysis about this model. Ty!Kelevra Keilahttps://kelevra.diagonalmedia.netnoreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-38952620280290323852020-08-03T05:11:19.289-05:002020-08-03T05:11:19.289-05:00Something I don't see here (and I don't th...Something I don't see here (and I don't think it's in Toxvaerd's paper, either) is the possibility that there is a non-linearity in costs of behaviour adjustment (social distaincing) such that it is infeasible (or too costly) to adjust behaviour sufficiently to get Rt to 1 or below. It strikes me that in that scenario quite a lot could change about the results. For example, if there are going to be a cycle of peaks where health systems are overwhelmed it could be advantageous to reduce one's efforts not to get the disease at points at which (with a lag) one expects health system pressures to be less at the time one became most ill. I presume there would be some relationship between the non-linearity of the costs and the non-linearity in the S-curve at peak infection rates. One could imagine properties such as that the first peak is infeasible to behave sufficiently to suppress below Rt=1 but later peaks might be (as S drops). Andrew Lilicohttps://www.blogger.com/profile/10609184492700616517noreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-80844115153759507142020-06-16T13:09:02.618-05:002020-06-16T13:09:02.618-05:00It seems like everyone addressing the issue of pan...It seems like everyone addressing the issue of pandemic spread has a pet model. Frankly, this is GOOD. Each model has strengths and weaknesses. This approach has some definite benefits. Economics succinctly captures the behavior of people. Taking a behavioral approach makes sense. After all, people behave. Your having to moderate as a result of abusers is a sad commentary on our times. People abuse the poster rather than discuss the content. Too bad. Society is worse off.Anonymoushttps://www.blogger.com/profile/09033013448200002816noreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-15314300109349983502020-06-02T08:32:42.048-05:002020-06-02T08:32:42.048-05:00So I have been thinking quite a bit about the R0 a...So I have been thinking quite a bit about the R0 asymptote to 1 result here. I think it is looks like that more because infections are small and so S(t) falls very slowly. If you look at the implied maximum level of infections from your result it is Log[S(t)\beta0/\gamma]/\alpha_I which means that as S(t) falls so should I(t). But your graphs don't show that which means you get R0 = 1. Instead, here R0 would be changing over time as S(t) changes. It won't look like that if S(t) moves slowly which it does under social distancing. In other words, the observed data is there because this is all much much slower to evolve than the standard SIR model.Joshua Gansnoreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-20930293157890782732020-05-23T00:15:26.629-05:002020-05-23T00:15:26.629-05:00That's a good question, I think.
It intuitive...That's a good question, I think.<br /><br />It intuitively seems like the answer would vary quite a bit depending on how much the first group mixes with the second group. And I think that's the more realistic model fo this sort of heterogeneity. There's more than zero contact between 25-year-olds and 75-year-olds, but the two groups don't mix equally with each other as with peers of similar age (at least not in the U.S.).realisthttps://www.blogger.com/profile/15859878286157071663noreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-31303894141618699742020-05-22T05:07:07.267-05:002020-05-22T05:07:07.267-05:00How would your model change if we introduced the f...How would your model change if we introduced the feature that for a large group of the population they faced no risk of death themselves but they could pass the disease on to others that would have a high risk of death - eg if we broke the population into two groups: say 90% with no risk of death and 10% with a 10% risk of death - and you knew to which group you belonged? The no-risk-of-death people might still care about their chance of infecting and killing the high-risk-of-death people, but they might weight that less than the high-risk-group themselves (especially with respect to those high-risk people they didn't know personally). What does the model look like in that case?Andrew Lilicohttps://www.blogger.com/profile/10609184492700616517noreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-68953635119169383732020-05-17T11:50:38.143-05:002020-05-17T11:50:38.143-05:00Thank you. I learned a lotThank you. I learned a lotAnonymoushttps://www.blogger.com/profile/16837835969219773168noreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-60645687497221883892020-05-16T10:34:24.351-05:002020-05-16T10:34:24.351-05:00The following article may be of interest: "Es...The following article may be of interest: "Estimating the burden of SARS-CoV-2 in France", Henrik Salje, et al. Science, 13 May 2020: eabc3517; <br />DOI: 10.1126/science.abc3517 .<br />https://science.sciencemag.org/content/early/2020/05/12/science.abc3517<br />Supplementary Materials: https://science.sciencemag.org/content/sci/suppl/2020/05/12/science.abc3517.DC1/abc3517_Salje_SM.pdf<br />Old Eagle Eyehttps://www.blogger.com/profile/05270080708077871311noreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-4125315300871137552020-05-14T11:02:58.712-05:002020-05-14T11:02:58.712-05:00From Taiwan's vice president, who is also a le...From Taiwan's vice president, who is also a leading epidemiologist: <br />"It’s not necessary to stop all activities. As long as more than 50% of the population reduces 50% of their social contacts then the outbreak can be controlled. They can go to school, to work but must reduce non-essential recreation and social contact."<br />https://www.telegraph.co.uk/global-health/science-and-disease/taiwans-vice-president-chen-chien-jen-countrys-fight-covid-19/<br />SDHhttps://www.blogger.com/profile/01681474314097558883noreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-86144487934962164802020-05-14T10:46:42.724-05:002020-05-14T10:46:42.724-05:00Interesting
https://www1.cbn.com/cbnnews/israel/2...Interesting<br /><br />https://www1.cbn.com/cbnnews/israel/2020/april/top-israeli-professor-claims-simple-math-shows-covid-19-runs-its-full-course-after-70-daysAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-582368152716771238.post-61080065160992598822020-05-12T00:02:17.869-05:002020-05-12T00:02:17.869-05:00Instead of classical epidemiology, this behavioral...Instead of classical epidemiology, this behavioral economic model is better. What about the discrete modeling of Covid within the network science community that gets insights from statistical physics? https://www.networkscienceinstitute.org/covid-19rudyhttps://www.blogger.com/profile/01590261145499718019noreply@blogger.com