Barry Blackwell: The Lithium Controversy: An Historical Autopsy. 

Ken Gillman’s comment on Barry Blackwell vs. Paul Grof and Jules Angst interaction


         I wish to sincerely thank the participants of this discussion (Blackwell, Grof and Angst) for their opinions and reminiscences, hopefully they will enable those in subsequent generations to learn from history and from other people’s “mistakes” (I use that word advisedly, since it is clearly an irregular verb: “I gain experience” whereas, “you make mistakes”!).  As Bismarck said, “Fools say they learn from their mistakes, I prefer to learn from other people’s mistakes.”

         My current writing project is about the strangely ill-defined concept of the “placebo response,” which, as generally used, is an undifferentiated jumble of distinct phenomena and processes.  It is puzzling that such an omnipresent notion should receive so little critical evaluation.

         This made me particularly aware of how such notions were relevant to this issue (e.g., “the problem of using a placebo control based on the high suicide rate in untreated affective disorder”).  Clearly, the relevance of the “placebo-controlled” group is less if the natural history of the condition is one of a steady level of symptoms, or a worsening, and also if the method of assessment is of “objective” categorical outcomes (like suicide), as opposed to subjective symptoms.  A “treatment as usual” comparison group would have been sufficient.  Evidence is mounting that the measure of treatment response (misleadingly and uncritically) labelled as a “placebo response” is greater when considering subjective symptoms and less with objective measurements (viz. dead/not dead).  With this in mind, one cannot help thinking of the comments relating to the lack of necessity for double-blind trials, commented on previously, e.g., by Shorter and Carroll, in relation to imipramine and Thomas Kuhn.  That also accords with the opinions from Sir Austin Bradford Hill about statistics, randomisation and blinding only being necessary when differences are small (Hill 1965; Worrall 2011).

         If one is treating hospitalized melancholic cases and making them well enough in a few weeks to leave hospital and go back to work, one is in a completely different ballpark to modern day trials of mild “illnesses” (more heterogenous and higher rate of mis-categorization) where people are looking for small differences in efficacy or side effects over short periods of time.  In such cases regression to the mean and improvements related to psychological expectation will be proportionally greater contributors to the small improvements in subjective rating scale scores.

         It is high time there was a more sophisticated dissection of the current rather mixed notion of “placebo response”; such an understanding helps in the interpretation of evidence relating to lithium.

         Second, a trivial detail, but I would be interested to know what Jules Angst thought of the silence from Shepherd: what sort of person goes to a meeting like that to say, “no comment?”  It seems odd, to say the least. Was it passively antagonistic?



Hill AB. The Environment and Disease: Association or Causation? Proc R Soc Med. 1965; 58:295-300.

Worrall J. Causality in medicine: getting back to the Hill top. Prev Med. 2011; 53(4-5):235-8.


March 12, 2020