When pharma spin docters go cherry picking

Pharma saves no trouble producing all kinds of "evidence" to prove to the GP that their products are "efficient" (so sales could go up followed by share holders value). To do this they often invent very good new products but also often try to repack evidence in a statistical plausibility cloth.

Neurosceptic unmasked a hidden "regression to the mean" falacy  example in a recent study sponsered by GSK about their antidepressive paroxetine flagship.

Nice example to keep for the record.