Barry Blackwell: The Lithium controversy: A historical autopsy


Barry Blackwell’s reply to Ken Gillman’s comment on the interaction between Barry Blackwell vs Paul Grof and Jules Angst


       I appreciate Ken Gillman’s comment on the relevance of the previous discussion about the placebo response and lithium controversy. He amplifies this with concern about the “strangely ill-conceived concept of the placebo response” and raises a few specific issues relating to studies of depression.

       There is, in fact, a vast literature on this topic from which I cite two publications.

       One, with relevance to antidepressant studies (Kasper and Dold 2015), in which the authors analyze factors contributing to increasing placebo responses in antidepressant drug trials based on the findings in 11 studies completed between 1993 and 2015.  They distinguish between two major possibilities.

       First, a decline in the newly marketed drugs that are less effective than older compounds. This is a likely outcome of the well-documented deterioration in the quality and reliability of industry-sponsored research contributed to by corruption, complicit academic ghost writers, KOL’s and lax journal editors in thrall with advertising revenue.

       Secondly, possible elevation of the placebo response influenced by the following:

1.    An increase in monitoring requirements and attention to subjects.

2.    A decline in severely ill or suicidal subjects due to increased ethical oversight.

3.    U.S studies using subjects enrolled by advertisements compared to non-U.S sites.

       A second publication is even more informative (Enck, Klosterhalfen, Weiner et al. 2010). They review the world literature on the placebo response begnning in the 1940’s. The number of annual publications on the topic has increased from virtually zero in 1953 to more than 140 in 2005 and beyond, amounting to a total of more than 2,000 studies investigating the “neurobiological and psychological factors underlying the placebo response.”

       It examines the following issues:

1.    The way addition or subtraction of the placebo response from the drug response influences outcome.

2.    Is the placebo response a function of the likelihood of receiving drug or placebo?

3.    Can the treatment response in drug trials be taken as evidence in clinical routines?

4.    Novel testing designs for drug testing

       The conclusions are:

1.    The assumptions we make about placebos in research routinely are at least questionable or false.

2.    Among them is the central proposition that true drug efficacy can be assessed by subtracting the placebo from the drug response.

3.    Increasing randomization of patients to drug rather than placebo increases the placebo rate and the number of patients to be included.

4.    Placebos are frequently used clinically and efficacy is often determined by physician behavior but is seldom evaluated.

       The last conclusion provided the motivation for a study using students as subjects in a pharmacology course more than 50 years ago. (Blackwell, Bloomfield and Buncher 1972).

       The study was conducted at the University of Cincinnati. Students attending a regular pharmacology class were invited to volunteer for a study comparing a single dose of one or two blue or red capsules identified as either a mild sedative or stimulant (all were placebos). Working in pairs they assessed the effects on pulse, blood pressure, mood and any adverse symptoms. Faculty predicted the outcomes based on a literature review of the placebo literature and sealed them in an envelope to be opened at the next class after the results had been tabulated and statistically analyzed.

      Results confirmed all the predictions and were published in the Lancet. The Head of the Department subsequently ruled the study “unethical,” but the students gave me their annual golden apple award!

       The doubts about the placebo response, expressed in the literature, support the prevailing suspicion that double-blind, placebo-controlled randomized clinical trials of investigational drugs have outlived their usefulness and should become a thing of the past. Such studies on highly selected and small sample sizes lack generalizability for clinical practice on the population at large and fail to detect rare, serious and sometimes fatal side effects. The rational alternative is to require that new compounds be tested against existing drugs known to be effective, with well identified side effects and which are almost always far less expensive. The basic goal is to determine effectiveness, not just efficacy. It is high time the FDA revise guidelines promulgated well over half a century ago.



Blackwell B, Bloomfield SS, Buncher CR. Demonstration to Medical Students of Placebo Responses and Non-Drug Factors. Lancet, 1972; 1:1279-82.

Enck P, Klosterhalfen S, Weiner K, Hong B, Zipfel S. The Placebo Response in Clinical Trials: More Questions than Answers. Philos Trans R.Soc. Lond. Biol. Sci. 2011; 366 (1572):1889-95.

Kasper S, Dold M. Factors Contributing to increasing placebo responses in antidepressant trials. World Psychiatry, 2015; 14(3):304-6.


July 30, 2020