Contact Matrix for long closure fails to open

[jbinary]

Hi, It seems in https://github.com/rajeshrinet/pyross/blob/master/examples/deterministic/ex4-age-structured-SIR-india-with-interventions.ipynb the last example has a flaw in contact matrix that it does not actually lift the lockdown which leads to an incorrect assumption that 49 days of lockdown is enough to stop the spread? Thanks.

[rajeshrinet]

@jbinary it is now updated to lift the lockdown. Earlier, I had mimicked with dotted lines, what will happen if lockdown continues. But probably, as you have noted, it is better to show what happens once the lockdown is lifted. Please note that in SIR or SEIR model, once social distancing is removed the cases will grow again. So social distancing alone can not completely remove the total number of cases. Either herd immunity or vaccine is needed.

[jbinary]

Yes, surely. The thing is people already made a fake picture claiming 49 days enough to fight it: https://twitter.com/JBinary/status/1248726482467328003 So probably it's a very good idea to update the GIF

[ronojoy]

@jbinary many thanks for bringing this to our attention. Please could you point us to the source of the manipulated image ?

[jbinary]

I don't have the source, someone have sent it into a twitter thread but I'm pretty sure they were not the author.

[jbinary]

And actually it really was not clear from the image, so the only sin of them is they said it's from Cambridge University.

[ronojoy]

They also added the side captions and the erroneous interpretation that "cases will not increase". This is not something we have said in our paper. Everything in this repository is under MIT license but that does not absolve the users from being irresponsible.

[jbinary]

Absolutely. We are lucky though: https://www.indiatoday.in/india/story/coronavirus-lockdown-india-cambridge-mathematical-model-extension-1661321-2020-03-30

[jbinary]

Oh wait sorry, it's not the source. I'll try to ask the guy where did he take it from.

[nbfigueroa]

Hi Rajesh and Ronojoy,

Just a suggestion: in order to avoid any further misinterpretations of your plots and predictions, perhaps you should also lift the lockdown for the three closure case; i.e.

def contactMatrix(t):
    if t<21:
        xx = C
    elif 21<=t<42:
        xx = CH
    elif 42<=t<47:
        xx = C
    elif 47<=t<75:
        xx = CH
    elif 75<=t<80:
        xx = C
    elif 80<=t<98:
        xx = CH
    else:
        xx = C
    return xx

and maybe increase the simulation time to >98. This way, for all the simulated scenarios, the plots will represent what really happens when the lockdowns are lifted, without any assumptions. This will hopefully avoid tweets like the one @jbinary pointed out.

Best and great work! Nadia

[ronojoy]

@nbfigueroa thanks for this suggestion (we'll keep this in mind for future work) and for your kind words. All the best!