Advertisement
Research Article

Misrepresentation of Randomized Controlled Trials in Press Releases and News Coverage: A Cohort Study

  • Amélie Yavchitz,

    Affiliations: INSERM, U738, Paris, France, Centre d'Épidémiologie Clinique, AP-HP (Assistance Publique des Hôpitaux de Paris), Hôpital Hôtel Dieu, Paris, France, Université Paris Descartes, Sorbonne Paris Cité, Faculté de Médecine, Paris, France

    X
  • Isabelle Boutron mail,

    isabelle.boutron@htd.aphp.fr

    Affiliations: INSERM, U738, Paris, France, Centre d'Épidémiologie Clinique, AP-HP (Assistance Publique des Hôpitaux de Paris), Hôpital Hôtel Dieu, Paris, France, Université Paris Descartes, Sorbonne Paris Cité, Faculté de Médecine, Paris, France

    X
  • Aida Bafeta,

    Affiliations: INSERM, U738, Paris, France, Centre d'Épidémiologie Clinique, AP-HP (Assistance Publique des Hôpitaux de Paris), Hôpital Hôtel Dieu, Paris, France, Université Paris Descartes, Sorbonne Paris Cité, Faculté de Médecine, Paris, France

    X
  • Ibrahim Marroun,

    Affiliation: Department of Internal Medicine, Hôpital Foch, Suresnes, France

    X
  • Pierre Charles,

    Affiliation: Department of Internal Medicine, Hôpital Foch, Suresnes, France

    X
  • Jean Mantz,

    Affiliation: Department of Anesthesiology and Critical Care, Beaujon University Hospital, Clichy, France

    X
  • Philippe Ravaud

    Affiliations: INSERM, U738, Paris, France, Centre d'Épidémiologie Clinique, AP-HP (Assistance Publique des Hôpitaux de Paris), Hôpital Hôtel Dieu, Paris, France, Université Paris Descartes, Sorbonne Paris Cité, Faculté de Médecine, Paris, France

    X
  • Published: September 11, 2012
  • DOI: 10.1371/journal.pmed.1001308
  • Featured in PLOS Collections

Reader Comments (4)

Post a new comment on this article

Open Access to Raw Data is Needed

Posted by marywoodward on 29 Sep 2012 at 19:49 GMT


Randomized controlled trials are supposedly the holy grail of research developed to avoid the pitfalls of bias and confounding. However sadly, the results of such trials are often distorted when presented as press releases for the public information. A team of researchers led my Isabelle Boutron found that in a study of 70 press releases relating to the results of clinical trials, 47% contained spin(1). Yavchitz A et al identified that where the press release was misleading, the authors were responsible in half of the instances because of the spin they had added to their abstracts. The culprit in the remaining half was not specified(2).
The problem of misleading abstracts can be tackled easily by alert editors of journals and peer reviewers. However the fact that the spin was deleted from the abstract may not prevent authors and their sponsors from making such claims in press releases. In an ideal world medical research would be done for purely altruistic reasons. However when vested interests prevail this will be impossible to achieve.
The Cochrane Collaboration was developed to give unbiased assessment of all available research looking at primary data where possible to avoid spin. Yet the spin in the abstract of the meta-analysis on surfactant therapy was published in the BMJ some time back but the abstract was not retracted (3).
It is well known that sponsoring agencies exert considerable influence over the published data. However the instance of the sponsoring agency giving a press release at variance from the published paper has also been published in the BMJ. The press release was issued by WHO, jointly with GAVI, USAID and John Hopkins among others, after the Bangladesh case control study on the effectiveness of Hib vaccination .This press release suggested that the vaccine is useful, whereas the study itself showed no benefit. No statistical difference was seen in the vaccination status of those with pneumonia or meningitis compared to
controls. A post-hoc analysis presented without proper multiple testing
was used to bolster the erroneous claim(4).
Journal editors alone cannot be the sole gambit in this vicious whirlpool. Stringent measures that hold the authors and their sponsors responsible for misleading press releases are needed. Open access to raw data of published RCTS may be a means of making the authors more circumspect.
REFERENCES
1. Boutron I, Dutton S, Ravaud P, Altman DG.Reporting and interpretation of randomized controlled trials with statistically nonsignificant results for primary outcomes.JAMA 2010 ;303(20):2058-64.
2.Yavchitz A, Boutron I, Bafeta A, Marroun I, Charles P, Mantz J, Ravaud P.Misrepresentation of randomized controlled trials in press releases and news coverage: a cohort study. PLoS Med 2012 ;9(9):e1001308. Epub 2012 Sep 11.
3.Tiwari L, Puliyel JM, Upadhyay P. Truth and evidence based medicine: spin is everything. BMJ 2004 ;329:1043.
4 Puliyel JM. GAVI and WHO: Demanding accountability.BMJ 2010;341 :266





No competing interests declared.