A bit of digging and one can get to the ultimate source of the 35% increase claim with the originator of the article:
http://www.counterpunch.org/sherman06102011.html
From a purely statistical standpoint, 4 weeks is waaaay too short of a period to draw any meaningful conclusion.
The authors didn't rule out that the 4 weeks they measured wasn't itself a statistical anomoly outside the normal bounds of expected results. The 4 week period they measured could have been a 4 week period for that area with an unusually low indicidence, so that any period following would have seemed like a e.
I can tell you from first hand experience with data regarding infant mortality (something very important to some HMO's), that it can be "lumpy", if one simply looks at one month of data, or even one year sometimes.
Seems like bad science to me. Important question to be sure, but it doesn't pass scientific muster, and definitely requires more fleshing out, before anyone should rely on it.
Unfortunately, bull can spread at the speed of light on the internet.
I will, of course, defer to your take on it, the above is just my understanding.