Statistics class 101: get the largest sample size you can across as broad a range of factors as possible. 14 cases out of 30-40,000 is statistically so insignificant that you can dismiss it as any sort of guideline.

It would take hundreds or more to even make me look at the study. Step aside from the cancer forums and read some statistics and probability wikis and you'll feel a lot better. Anything could be factored into the OP's cited study, there could be an environmental factor at work, quality of care received, heck anything can heavily skew the 'results'. Medical stats are some of the hardest to get accurately and the worst to remain relevant because of the nature of medicine's constant improvement.