Advertisement
Research Article

Initial Severity and Antidepressant Benefits: A Meta-Analysis of Data Submitted to the Food and Drug Administration

  • Irving Kirsch mail,

    To whom correspondence should be addressed. E-mail: i.kirsch@hull.ac.uk

    Affiliation: Department of Psychology, University of Hull, Hull, United Kingdom

    X
  • Brett J Deacon,

    Affiliation: University of Wyoming, Laramie, Wyoming, United States of America

    X
  • Tania B Huedo-Medina,

    Affiliation: Center for Health, Intervention, and Prevention, University of Connecticut, Storrs, Connecticut, United States of America

    X
  • Alan Scoboria,

    Affiliation: Department of Psychology, University of Windsor, Windsor, Ontario, Canada

    X
  • Thomas J Moore,

    Affiliation: Institute for Safe Medication Practices, Huntingdon Valley, Pennsylvania, United States of America

    X
  • Blair T Johnson

    Affiliation: Center for Health, Intervention, and Prevention, University of Connecticut, Storrs, Connecticut, United States of America

    X
  • Published: February 26, 2008
  • DOI: 10.1371/journal.pmed.0050045

Reader Comments (48)

Post a new comment on this article

Results of the meta-analysis may also depend on outcome measures

Posted by plosmedicine on 31 Mar 2009 at 00:21 GMT

Author: Cyril Hoschl
Position: Professor of psychiatry
Institution: Prague Psychiatric Centre & Charles University, Prague, Czech Republic
E-mail: mail@hoschl.cz
Submitted Date: February 27, 2008
Published Date: February 28, 2008
This comment was originally posted as a “Reader Response” on the publication date indicated above. All Reader Responses are now available as comments.

Dear Sir,
The “law of initial value“ says that the worse is the patient at the beginning of the study, the higher is the improvement (= difference between initial and final measure). At the same time, however, we often see that the worse is the patient at the beginning the worse is also at the end of the study. In other words, the correlation between the initial value (of a scale) and the improvement is NEGATIVE, while the correlation between the initial and final value is POSITIVE. This is avoided when the effect is expressed as a RATIO of final and initial value, which is unfortunately only exceptionally used. I wonder how the results of your meta-analysis would change if you took into account not the difference, but only a final value of HAMD. Could it significantly change the conclusion? It would be clinically relevant because we do not care as much how the patient improved in a study but rather how he/she is finally doing! Moreover, our experience shows that effect of placebo decreases with time. It means that the biggest differences between placebo and an active drug do not occur during the 5-8 week period of clinical trials, but later on. Could it be possible from your data to calculate such a trend and hypothesize that the difference between placebo and an active drug could be observed later on during the treatment? The interpretation that with the severity of depression the effect of placebo decreases and the effect of active drug remains unchanged seems logically unacceptable in spite of your convincing figures: patients in double-blind RCT namely do not know if they take placebo or active drug, so the decrease observed for placebo applies for active drug as well. If it does not occur, it means efficacy (not inefficiency) of the drug.
Cyril Hoschl, MUDr., DrSc., FRCPsych
Prague Psychiatric Centre

Competing interests declared: Lectures for Lilly and BMS
faculty member, Lunbeck Institute
research consultant, Servier
no shares of pharma industry