It doesn't matter what the author's intention was. Until you can at least see an abstract, you might as well act as if the study doesn't exist. Without knowing what study we're talking about, there's no way to discuss it. Hell, the author says "medication", but doesn't even say which medication was studied. As far as we're concerned, this is about as good as "I heard a guy on the train say..."

Edit: As a side note, damn that's a stupid editorial policy. The opposite should be true; any scientific claims in an op-ed should be required to have a citation, even if it's a poor one. I guess it's just one of those places where journalistic and scientific epistemology clash. Journalists say "this is what an expert thinks!", while scientists say "here is some data!"

Here is link to the study showing no long term benefits to adhd meds: http://www.ncbi.nlm.nih.gov/pubmed/19318991

OK, so the elephant in the room here is this: you cannot confirm a null hypothesis (in this case: "Ritalin does not work in the long term") by failing to falsify it. This study is simply not designed such that it can provide strong evidence that ritalin is ineffective in the long term. It can only fail to provide strong evidence that it is effective in the long term.

Furthermore, the author refers to it as a "well-controlled study". It's not. The original 14-month study was controlled, and it showed short-term improvement. This study is a prospective follow-up. It's not a controlled trial. It's not trying to be a controlled trial.

Finally, looking at the full text, they appear to have used linear modeling. That's also a red flag. Obviously, sometimes a linear model is the right choice. But most often it is used because it is the convenient choice.