So Monica Gandhi and Kyle Hunter made a few mistakes in their piece on kids and the pandemic, specifically in suicidology. With additional context, a lot of their points about MH and the pandemic specifically melt away. I'll enumerate them.
/1
First, she claims that there was an increase of 24% child suicides in california (by rate it's 24-27%) in 2020. She points to data for 2019, 2018, 2017 to show that this is an anomaly.
/2
In fact, this is true. However, some additional context (or, had she consulted a suicidologist) sort of undermine this argument. The suicide rate has increased by 24% or more 4 times prior to 2020 in the past 20 years, and does so with a regular frequency.
excel data, CDC
/3
Next, she makes the point that the CDC just showed a 50% increase among (mainly) girls during the pandemic.
Significant editing is required.
/4
The CDC data has issues (discussed before, but it's data & it DOES show increase in distress in girls for sure. But only girls. not "mainly" girls. For boys, there was no difference. And it was not "during the pandemic."
/5
For the first portion of the pandemic, rates of ER presentations were lower for both girls and boys, and country wide, almost every school was closed.
/6
For the 2nd portion, there was a split in ER presentations b/w girls & boys, with boys returning to baseline & girls up by about 10%. This is during the end of school yr, the summer, and new school year, where school attendance really depended on state rules &conditions.
/7
Certainly in Dr. Gandhi's jurisdiction, mostly closed. But across the US, there was a lot of variance on whether or not kids attended schools during this portion of the year.
In my jurisdiction, BC Canada, schools have been open since September. Our SA/SI pattern was similar.
/8
In the new year (2021+), we see this huge takeoff of girls and ER presentations, but this is during weeks 8-18, where almost half of schools are open across the United states. Remember, this is national data, not California data. And for whatever reason, only girls, not boys.
/9
So for this spike to be related to school closures, it has to explain:
a) why only girls and not boys (unless I missed the memo that boys were totally off the hook and didn't have to go)
b) why the rates are increasing the more that schools are opening
/10
I think it should be concerning to Dr. Gandhi, who has maintained quite the ... sure ideological position... if adding context to the data she provides completely changes the weight of it as evidence of school closures being to blame.
/11
In my jurisdiction, Vancouver, schools have been open full time since September (Trust me, my timeline is full of angry parents kids and teachers about this!!). I see kids professionally who are attending schools. And we have seen almost identical patterns in suicide/MH.
/12
Those that are blaming school closures for mental health distress numbers might do a thing or two and examine their own evidence.
/13
(regarding this data, i made strait population copy from 2020. this being a census year, it's difficult to know what happened with subset populations. the results do not change if the trend continues (27% increase) or reverses (24% increase)
I would also like to say that I believe in the principle of charity. I think the main reason this group of UCSF physicians are so convinced its school closures is because..well.. in their jurisdiction, schools are closed.
I don't think its malfeasance or incompetence. Just bias.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
The CDC made a large mistake in their recent publication. It took data from "nonsuicidal self injury" and "suicide attempts" and combined them to create a metric "suspected suicide attempts."
This is stigmatizing and wrong. I will explain why.
Non-suicidal self injury is a phenomenon that happens in adolescence at a high rate. About 12 to 20% of adolescents and young adults engage in non-suicidal self-injury yearly.
It is **not** suicidal. It differs from suicide in many ways.
/2
Non-suicidal self-injury (NSSI):
* Is recruited for therapeutic purpose
* usually involves superficial injuries that do not threaten life
* usually provides relief (as reported by those who do it)
* When we ask kids, they are clear. "IT MAKES ME FEEL BETTER*
/3
Hey CDC if your article is about suicide attempts why did you present data about nonsuicidal self harm?
Shocked to see the CDC make such an elementary mistake in its titles and headlines, and it was a great opportunity to separate the two **very different** phenomena.
Nonsuicidal self harm is a coping mechanism. It's intent is to feel better, not die. It's a sign of distress and ineffective/failing coping mechanisms and usually people with it are suffering but want to live. It's very common in adolescence. (10 to 22% depending on survey).
Suicide attempts are gestures or actions intentionally taken with the purpose to die. They are signs of hopelessness and overwhelming distress, and represent a much smaller group of kids. By survey, about 6 to 8%, by ER presentation about 2%.
Biiiiig separation between boys and girls in the US for ER presentations for suicide attempts. Very average-to-highish year for boys, big difference for girls. Amplification not surprising: Girls are more likely than boys to gesture/attempt suicide, boys are more likely to die.
The same was not seen (or as dramatically so) in slightly older people (18 to 24)
Please note reports like this do not support your pet theory on what caused it. The increases in pediatric have been continuing since 2015 (30 to 50% year over year) and the pandemic/Lockdowns/whatever may or may not have had an influence on it. Causes yet to be determined.
Because the media is obsessed with Japanese suicide data ("suicidal" Japan is a lazy, one-sided trope in Western news), I'd like to add nuance.
First, here are the graphs for who suicides in Japan for the past 5 years, to February 2021! /1
If we zoom into the men, we can see that there was a decrease in early 2020, an increase at the end, and things are back to normal-ish for 2021. In fact, 2021 is a record low for January and 2nd lowest for February. /2
If we zoom into the women, we can see that there is again an early mild decrease followed by a SUBSTANTIAL increase at the end of 2020. Though things are still elevated in Jan-Feb 2021, much less so and hopefully on the way to normalizing. /3
1/ CW: abuse (please read only if you feel in a good space to, this stuff is tough, but it's important data). Statistics from #TX don't show the predicted "major spike" in child abuse fatalities during the pandemic.
Contd:
2/ Every child death by abuse, neglect, and wrongful supervision is a tragedy. In no way am I underplaying the importance of these deaths. I'm simply pointing out that the increase is not some kind of previously-unseen thing. It's definitely no tsunami.
3/ evidence:
a) as evidenced, the number is quite "noisy" (standard deviation is 35), and the # of deaths falls between the min (151, 2014) and the max (280, 2009)
b) the rate change is lower than most recent years (its lower than 6 and higher than 8 of the past 14 years)