Great points the placebo dropouts I’d greatly outnumbering dosed patients.. Would have skewed these numbers .. And wouldn’t those patient not seeing any change be the first to drop out? They would be placebo as dosed patients saw almost immediate benefit..
The zero responders would have brought down the placebo response and create clinical significance.
Well the main issue was the strong placebo response, so I find it hard to make sense of your question.
Per Kaufman the issue is believed to be a disproportionate (poor) baseline scores in the placebo versus dose arm patients. This could happen simply be chance especially in a small trial. In any event dropouts must be accounted for including their baseline scores and any missing trial duration data, which the statistical methods used should ensure.