School and Suicide

There’s a well-known seasonal effect to suicide in that people are more likely to end their lives during the spring and summer months and less likely to do so during the winter. However, when looking at youth suicides (ages 18 and under) the pattern reverses during the summer vacation months, the opposite of the adult pattern. Which could be interpreted as evidence that school is really, really bad for one’s mental health.

The trend of increasing suicides is evident from the trend-line in the following charts:

But what we are really interested in is the seasonal pattern.

When plotting the monthly means of the number of suicides from 1999 to 2017 the different patterns are readily apparent: Youths are less likely to kill themselves during the summer months compared to adults whose suicide rates peak just when youths are the least likely to kill themselves.

We can also examine the association between summer vacation and suicides by fitting a poisson regression with the month of year included as a categorical variable. The model includes an offset to account for the uneven number of days in each month.

For completeness sake I’m including the suicide deaths means of college age adults

I also fitted a regression model to the 2014-2018 suicide data to see how the marginal effects of the interaction age * month affect suicide rates. It’s pretty much what you expect, June and July have the lowest suicide rates when people are younger, but once people leave school their rates start rising.

I added a spline to the model cause the effect of age isn’t linear as should be obvious from the plot.

Copy 
1fit2 <- glm(rate ~ splines::bs(age) * month_fac, data = df)