Restoring Study 309: Figures don’t lie, but . . .

This week the BMJ published an article written under the “Restoring invisible and abandoned trials initiative.” This 2013 initiative was undertaken to bring unpublished trials to light and to correct misreported trials, “. . .to see whether access to and reanalysis of a full dataset from a randomized controlled trial would have clinically relevant implications for evidence based medicine.”  The BMJ article details the reanalysis of  Study 329, a 2001 study funded by GlaxoSmithKline (GSK) (then SmithKline Beecham), published in The Journal of the Academy of Child and Adolescent Psychiatry (no hack journal), and written, not by any of the 22 named authors, but by a GSK-hired ghostwriter. Study 309 purported to show that GSK’s antidepressant drug paroxetine was both effective and safe for the treatment of adolescents with depression. GSK’s marketing campaign based on the study touted its “REMARKABLE Efficacy and Safety,” and in the next year alone over two million prescriptions were written for children and adolescents in the United States.

The researchers for the BMJ article went back and laboriously dredged up original, previously-unavailable patient-level data from the trial and re-analyzed it using the original study protocol; essentially, they re-performed the original study, with two crucial differences: they weren’t being paid by the drug’s manufacturer, and they did not repeat the subterfuge employed in reporting the original study to mask the outcomes. Based on their independent analysis of the more-complete data set, their conclusions are that paroxetine is no more effective than placebo, and that it results in significant harms, including suicidal ideation.

The views, opinions and positions expressed by these authors and blogs are theirs and do not necessarily represent that of the Bioethics Research Library and Kennedy Institute of Ethics or Georgetown University.