The Attitude of Chicks to Trojans & Horses

October, 21, 2013 | 3 Comments

Comments

  1. As a Doctor of Psychology who has lived long enough to observe the fickleness of research and public gullibility (though these are not the guilty ones), I’m pretty sure the same thing has happened in psychology research, which according to the eternal swing of the pendulum, now points to Behavioural methods, to the detriment of all else. Once a path in research is started, unfortunately, bad conclusions only build on themselves. The most lucrative grants go to those who have started a thread and are already immersed in pursuing it, under the belief that these researched must continue to be supported lest their historical funding investment be lost. Indeed, what can granting committees say when faced with fabricated, or at least doctored, near-perfect research results?

  2. Thank you for this post.
    Another question is: why doesn’t Alltrials campaign ask for IPD?
    Especially when EMA is proposing to discuss IPD public availability.
    Of course, publication bias is obvious, outcome reporting bias too. Getting CSR will provide scientific community with important information. And I signed the Alltrials campaign.
    But there are examples of incomplete or wrong analysis in CSR used for approvals. Information showing bias was only available in FDA reports and obtained from IPD analysis. (Examples here in French : http://www.etudes-et-biais.com/individual-patient-data-ipd-une-necessaire-mise-a-disposition-de-la-communite-scientifique/).
    Should we leave FDA as the sole authority able or allowed to check if CSR are correctly or completely describing the results of trials? EMA is not performing IPD analysis.
    Shouldn’t investigators have access to the IPD of their centre to check if data included in the database do correspond to their patient file and sign for it?

    Without IPD, the only thing we can say about a doubtful trial is that there is a risk of bias but we cannot describe how the potential bias distorted the results and if it is a bias or a fraud.
    Search for a “risk of bias” is the only research we can do. But a risk of bias doesn’t induce exclusion of the trial from Cochrane Meta-analysis which is at the top of EBM pyramid.
    IPD availability could add a new top over the EBM pyramid. I hope it is not the reason why IPD availability is excluded from Alltrials campaign.

Leave a Reply