They didn't report the period and the "normal" rate of dying.
If the normal chance of dying in one year is 1/10000, a change to 1.2/10000 nothing to worry about. Their result might be statistically significant according to the rules of statistics (which apparently are deeply flawed too), but it certainly looks like noise. Even the doubling for diabetes is probably insignificant.
Slashdot Mirror
They didn't report the period and the "normal" rate of dying. If the normal chance of dying in one year is 1/10000, a change to 1.2/10000 nothing to worry about. Their result might be statistically significant according to the rules of statistics (which apparently are deeply flawed too), but it certainly looks like noise. Even the doubling for diabetes is probably insignificant.