How is this even remotely good science? So far as I can tell, they didn't admit the possibility that this can be either complete noise (if you have, say 100 diseases, and you're looking for time-series correlations, or performing tests with too low an alpha or power) or caused by a confounder.
The task of making causal claims is a difficult one, and not one to be made lightly. The literature is replete with examples of this going terribly wrong, even with perceived physical justifications. See
The task of making causal claims is a difficult one, and not one to be made lightly. The literature is replete with examples of this going terribly wrong, even with perceived physical justifications. See
[0]: tylervigen.com
[1]: http://nerdsonwallstreet.typepad.com/my_weblog/files/datamin...
[2]: http://www.huffingtonpost.com/2013/04/16/reinhart-rogoff-aus...
[3]: http://www.economist.com/node/5246700