Thread by @tylerblack32

1/ California Kids: Suicides in COVID

The CA contrarians like the fearmongering & almost-always-wrong-in-two-months Monica Gandhi ghoulishly used child suicide #'s to advocate for reopening.

As shown, there was *no* unexpected increase of suicides during the yearlong lockdown.

2/ [/1]

When taken as a whole (red - March 2020-Feb 2021 vs blue - March 2019-Feb 2020), the overall increase was not significantly higher (95% CI -7.4% to +46%), though as the pandemic wore on things appear to increase than earlier.

Grey: 2013-2017

[/2]

3/ Rates in the 2nd half of the lockdown response in Sept-Feb 2020 are higher than 2019's rates (+2% to +169%, 63%), but it should be noted that this is not significantly higher than 2018, 2017, or 2016's rates.
[/3]

4/ As i've frequently said, comparing one year to another is problematic, especially in (even large) states, where variance in numbers affects things greatly. 2019 makes a very low comparison year, 2018 makes a moderate comparison year, and 2017 is actually quite similar.
[/4]

5/ When trying to interpret data, it's always important to go back in time and look for two things: trend and variance. If you don't know those two things, comparing two yearly numbers is a big problem.

6/ Trend: if there is a trending increase or decrease, the new rate needs to account for it. (IE: if the US population grows by 1.5% every year, a 0.5% increase could be a significant *decrease* in the trend)

7/ Variance: rare events can cluster.

Consider child suicides in CA in November:

'16: 16
'17: [redacted, less than 10]
'18: 12
'19: [redacted, less than 10]
'20: 16

So sure, 2020 was up from 2019, but that number is up and down by >45% and 16 happened 5 years ago too.

8/ While schools reopening can have benefits, "preventing child suicide" is likely not one of them. NB: i can say that embargoed data from CA after march 2021 looks reassuring.