When they collected the data in 2000, about seven 8 year olds out of 1000 had autism. Those children were born in 1992.
When they collected the data in 2002, about six 8 year olds out of 1000 had autism. Those children were born in 1994. This information was not labelled on the x axis.
When they collected the data in 2004, about eight 8 year olds out of 1000 had autism. Those children were born in 1996.
When they collected the data in 2006, about nine 8 year olds out of 1000 had autism. Those children were born in 1998. This information was not labelled on the x axis.
When they collected the data in 2008, about eleven 8 year olds out of 1000 had autism. Those children were born in 2000.
... I'm too lazy to continue but ...
When they collected the data in 2020, about twenty seven 8 year olds out of 1000 had autism. Those children were born in 2012.
When they collected the data in 2022, about thirty two 8 year olds out of 1000 had autism. Those children were born in 2014. This information was not labelled on the x axis.
I'm not positive i'm reading it right, but that's what I think they were trying to convey with this (terribly labeled) graph.
eta: yeah, rereading the subtitle, those numbers and years make sense
I think there's actually not 2 bars per year, but instead:
The x axis increases by 2 years each entry.
The number after the "|" is only meant to "helpfully" (and confusingly) tell you when those children were born. To take your example: