Publication bias

In the ‘Publication bias’ section of Sismondo’s article ‘How pharmaceutical industry funding affects trial outcomes: Causal structures and responses’, Sismondo describes how sponsorship bias may influence the reporting and publication of industry-funded trials. Sismondo notes that positive results are over-reported relative to negative results for industry-funded trials, and provides evidence for this from a case of trials of five SSRIs. In this case, 42 trials separated into 21 positive trials and 21 negative trials, although the 21 positive trials produced 19 articles, whereas the 21 negative trials produced merely six articles. This evidence certainly supports the conclusion that positive results of industry-funded trials are at least numerically over-reported compared with negative results of such trials. But without further evidence or information, this does not necessarily entail that positive results of such trials receive more biased emphasis of a qualitative rather than quantitative kind than do the corresponding negative results. For example, while the positive trials mentioned above spawned 19 articles relative to the mere six articles of the equally numbered negative trials, perhaps the six negative articles were written in a much stronger, lengthier, and more emphatic fashion than the 19 positive articles. If this were the case, then we could not conclude that industry-funded trials over-reported positive results rather than negative results in a manner transcending mere quantity, at least in this context. So, in order to properly determine whether or not publication bias truly exists in a substantial manner for the data reporting of industry-funded trials, we must also examine evidence concerning the quality of the relevant reporting, rather than just the quantity of such reporting.

This entry was posted in Uncategorized. Bookmark the permalink.

One Response to Publication bias

  1. Pingback: 14 to 40 percent of medical research are false positives (Yet Another Calculation) | Alea Deum

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s