Good article, but I'm concerned the premises of the article have been skewed by misleading graphs. The initial graph of GDP to medical expenditure is reasonably convincing, the graphs later in the article much less so. I could be wrong about this, but it feels a little like the data selection and curve fitting have been selected to supported the hypothesis rather than the other way around (something that could easily happen accidentally).
On a tangential note - that pie chart he uses in each section is an unintentionally good example of what not to do when displaying data. I have to confess, I don't really understand it.