Read on Twitter
Someone just shared this w me. It seems to be going viral globally because we are all desperate for hopeful predictions but it's based on completely unsound epidemiology and is totally totally wrong. If you see people sharing it please let them know.
1/ http://ddi.sutd.edu.sg
1/ http://ddi.sutd.edu.sg
Basically it assumes that the reason infections are now going down in many countries is that covid has spread through the whole pop so they are approaching herd immunity - this is how the bell-shaped curve produced by SIR models works.
2/
2/
This is not what's happening w covid - cases are ARTIFICIALLY going down in many countries because distancing policies have suppressed the disease BEFORE a large portion of the population gets infected.
3/
3/
This means their proposed epidemic end dates are totally meaningless because at those points huge numbers of any population that suppressed covid successfully are still 'yet to be infected'.
4/
4/
The way they've used the SIR model here is completely inappropriate and wrong and it's clear that no epidemiologist had any input into this analysis.
5/
5/
But bad science happens all the time - the reason this worries me so much is that they've had 1 million unique visitors to the site in its first 7 days of running.
6/
6/
I worry that many of these million people will make bad decisions (personal or policy) as a result of trusting this misinformation because it's misleadingly framed as well-informed and sciency.
This. Will. Cost. Lives.
7/
This. Will. Cost. Lives.
7/
Information is our most powerful tool against coronavirus and we all have a responsibility to use it wisely. Dangerous things happen when we decide our science degree in one area qualifies us to speak as experts in a field we know nothing about.
8/
8/
I can't express strongly enough how epidemiologically illiterate these methods are. Please debunk them if you see them. If you want proper covid predictions, look at @JohnsHopkinsEPI where they actually know what they are doing.
9/9
9/9
I realise this thread comes across as v scathing. I don't mean for one second to fault the intentions of the team who worked on this and I understand why they are taking the tools they know and applying them to this unprecedented global situation to see if they will help.
I'm just really worried about misinformation like this influencing a huge audience.
There's a lot of Maslow's Hammer research going on atm by people who are not epidemiologists (data scientists are particularly susceptible) and we need to be careful
https://en.wikipedia.org/wiki/Law_of_the_instrument
There's a lot of Maslow's Hammer research going on atm by people who are not epidemiologists (data scientists are particularly susceptible) and we need to be careful
https://en.wikipedia.org/wiki/Law_of_the_instrument
