Thread by @tylerblack32

US Pediatric Suicide in 2021
By Region
=-=-=-=-=-=

With historical context, changes can be best interpreted. This is very important as many media/pundits/presenters lazily present year-over-year trends as if it can tell the full story. It can't.

Here's the US since 1999:

/1

Suicide rates during the pandemic are within expected trends, & no increase is significant. The highest increase in the past 21 years was in 2017 for boys and 2004 for girls. The 2021 rate is a record for American girls, but again, well within the trend we would expect.

/2

We'll now look at things regionally, and start with the HHS region 1 (CT, ME, MA, NH, RI, VT). We can see, that trends are within expected for boys and girls, and these noisy rates bounce quite a bit. (missing values for the girls in earlier years is due to CDC rule)

/3

HHS Region 2 (NJ, NY are represented only, the CDC data does not include Puerto Rico or US VI) shows that the pandemic period is a relatively lower rate of suicide, with the records and record increases occurring earlier).
/4

HHS Region 3 (DE, DC, ME, PE, VA, WV) saw records for boys in 2021 (following the previous 15 year trend).

Girls were pretty much on average for the past 8 years.

/5

HHS Region 4 (AL, FL, GA, KY, MS, NC, SC, TN) had quite a jump for boys, and is the strongest worsening change for boys among the regions, though I wouldn't exactly call it outside of the trend since 2006.

The rate for girls essentially unchanged for 7y.
/6

On the other hand, Region 5 (IL, IN, MI, MN, OH, WI) has a huge drop for boys overall since 2018, and 2020/2021 really within longstanding trends.

For girls, the "slow steady creep up" continues but neither pandemic year a record.
/7

HHS region 6 (AR, LA, NM, OK, TX) seems to have a decreasing trend for boys the past few years.

Girls suicide hit an all-time high in 2021 as part of a well established pre-pandemic trend. No "pandemic leaps," but certainly a concerning march.

/8

HHS region 7 (IA, KS, MO, NE) seems to be on trend for continued increases, but lower than previous years both for girls and boys.

/9

HHS region 8 (CO, MT, ND, SD, UT, WY) shows no significant suicide rate change in boys during the pandemic relative to the previous 5 years, and girls. In fact, girls on a general trend decrease since 2017.

/10

HHS Region 9 (only including AZ, CA, HI, NV) has no pandemic effect on boys and an increase for girls, though again it follows trends going back to 2007.

/11

Finally, we have HHS Region 10, which shows no significant pandemic effect, and the record increases and rates are well before the pandemic for both boys and girls.

/12

Overall, the changes during 2020/2021 are pretty much exactly what we would expect with random variation. Remember, rates can only go up/down, & most of the time, changes are small.

Comparing to the previous 21 year trend, the majority of pediatric suicide rates are lower.

/13

Comparing previous 2 years (2018, 2019), 2020&2021 again follow an expected random distribution of rate changes.

Overall, girls up by 3% (95% confidence interval -5% to +13%, non-significant)

Overall, boys down by 3% (95% confidence interval -8% to +3%, non-significant)
/14

It is possible for p-hacking (NOT WHAT WE DO) to create "significant numbers" in either direction.

Girls:

HHS 1,5,6,10 (13.9M girls represented)
- higher by 20% (6%-35%)

HHS 2,3,7,8,9 (12.3M girls represented)
- lower by 15% (2%-27%)

/15

Again, this is unethical stats creating these "significant numbers", and the statistical correction needed would remove all significance.

Boys:
HHS 1, 3, 4 (14.9M boys represented)
- higher by 11% (0%-23%)
HHS 2,5,6,7,8,9,10 (25.3M boys represented)
- fewer by 9% (2%-16%)

/16

Simply put, there is no evidence of pandemic-related changes to American pediatric suicide rates in 2020 and 2021. If anything, it is exactly on trend.