What kate3 said.
The sample size is small. They did no tests of significance, which I find concerning (at best, they didn't find a statistician, at worst, they ran the tests and nothing was significant so they didn't report their p values - my guess based on the sample size is the latter).
There's no information about the groups. Are there underlying differences? It makes sense that vaccinated children see their GP more. Are they getting diagnosed more because they see the GP more? Are those parents more likely to go to the ER? Is there a difference in their attitude toward medical professionals? Are there socioeconomic factors influencing their ability to see a doctor? Is what one person sees as 'aggressive' seen by another parent as normal 3 year old behavior? There's no attempt made to model the data to tease some of these issues out.
My disdain here isn't for the underlying thoughts, but it irks me to see data misused and misrepresented. As a study, this is meaningless.