Wednesday, February 24, 2021

Comments by Sparkytus

Showing 1 of 1 comments.

  • Dear Mr Whitaker. I enjoyed this article and wholeheartedly support the views expressed. I just wondered of you could clarify one of the figures. On page 5 you state that 83% of subjects in the Leucht study fall into the “harm” category. This consists of the 59% who don’t respond to drugs and the 24% who would have responded to placebo. I’ve been puzzling over this. If I give 100 patients antipsychotics and 41 respond then, as you say, 59 are now at risk of adverse effects for no discernible gain. This means that you are classifying some of your responders as now at risk of harm. However, I don’t think it can be 24 patients as surely the only ones at risk of harm are 24% of the 41 who responded (roughly 10 patients) – who would have responded to placebo if they had been given this. Your harm category would then be about 69% (still a substantial figure, obviously).

    Is it possible that another figure might be 41.5%. My imaginary cohort is actually 200 patients. I give 100 the antipsychotic and 100 nothing. Overall, 59 from the first group are now at risk of harm and 24 from the second group are also at risk (as they would have responded to a placebo) this means that the calculation is 83/200 = 41.5%

    None of this addresses the assumption that you are never at risk of harm from a placebo and we know that some patients in trials do suffer side effects from placebos. It’s difficult to explain this in terms of physiology but it does happen. The assumption that those who don’t respond to placebo are not at risk of any harm is therefore perhaps a theoretical one rather than one that is borne out by study results.

    I’d be interested in any thoughts you have, It’s entirely possible that I’ve just misunderstood the figures completely (not an unusual occurrence for me!)

    I wondered if the true figure for those in the harm category should be