To add to this, it's been said that ECT is like pounding on the closed hood with a sledgehammer until something happens.

This formulation is such dreadful English. Please think about what you are trying to say.

1 in 20 would be expected if errors are random and if no biases are in place like publication bias. We know that's not typically the case, so we can indeed expect it's worse than 1 in 20. The more I read about anti-depressant trials, the more I think we will come to view them as particularly egregious examples of deliberately badly designed trials. I've just read Bad Pharma by Ben Goldacre and it's mind-boggling what happens in real world trials. One little example: to take advantage of the 1 in 20 chances, simple hold 20 little trials and publish the one that gets good results! Or ignore your stated trial goals and measure and just cherrypick any good result afterwards. If you have many variables you're bound to get some correlations post hoc, just by chance alone.

It's important to realise what screening is and is not. Testing someone who is sick is not screening. Screening is testing the whole population (or most of it) even though they have no signs at all of the disease. The usefulness of screeing depends dramatically on the accuracy of the test used as well as the actual incidence of the disease in the target population. Testing someone who has signs and symptoms of the disease is symptomatic testing which is very different, not least in the statistics involved, as well as the particular test used. For instance, (asymptomatic) bowel cancer screening would use a fecal occult blood test to check for invisible blood in the stool. This is cheap, non-invasive, self-administered and moderately accurate. If on the other hand, you had symptoms like overt blood in the stool for a long time plus unexplained weight loss (say) then you would likely be sent straight for a colonoscopy which is expensive, invasive and pretty accurate.

There was never any serious proposal to raise the corporation tax rate in Ireland, it's the one non-negotiable position of the Irish government and always has been. Having a low rate is also popular generally, though many would feel the rate could go up a few percent (though still well below other many other European countries) so that corporations do contribute more, but no-one seriously proposes raising it dramatically. What should perhaps be changed are the rules about moving profits around so that multinationals actually pay the 12.5%. I read this week Google only paid a few million in tax last year. It's not the rate, it's the application of the rate. Germany may have a nominal rate much higher than Ireland but the effective rate is often much lower due to special tax breaks in certain sectors.