Exactly - I literally just had this discussion here yesterday (https://news.ycombinator.com/item?id=5475037). ADHD is incredibly tough (read: expensive) to diagnose properly, and almost impossible to self-diagnose with any appreciable degree of confidence.
My frustration comes from the fact that people first misdiagnose themselves or their kids, or diagnose themselves through inappropriate tests, which then leads people to conclude that ADHD doesn't really exist (see randallsquared's reply).
In other words, "our [cheap] classifier model is incapable of making accurate predictions reliably; therefore the latent variable we're predicting must be imaginary", instead of "our [cheap] classifiers are incapable of making accurate predictions reliably; therefore those diagnostic models are inadequate".
To be fair though, it can be difficult to diagnose even in the medical community, particularly given that it seems to be the default fallback for any parent wanting to medicate a child that may be going through a natural hyperactive phase.
On the other hand, I wasn't officially diagnosed until I was 20 so I spent 18+ years in the dark. Others I know went through similar (longer) periods of quiet self-diagnosis before they finally decided to go through what would be considered an appropriate test.