Conflicts of interest: Finally a formula that solves it!

It is now difficult to know what to do with the information provided in the disclosure section of journal articles.  I am puzzled as to how are we supposed to train residents and fellows into how to incorporate this information into the decision making process.  After all, we want the readers to formulate impressions about disclosures so that they attain objectivity about content.  And objectivity is about measurement.  So how is one supposed to do so?  Until now no one knew for sure.

As background I would suggest we needed a simplified statistical system for an audience who is known on occasions to struggle with stats.  For example many very smart colleagues still struggle with the significance of the p value.  Less than 0.05 Biblical!  Anything else discard immediately...  Let alone understanding a Bonferroni correction.

So to try to solve this problem I tackled a recent paper in the NEJM where C1 inhibitor deficiency is successfully treated by a new intervention (one of 3 papers in the same issue).  (Zuraw et al N Engl J Med 2010; 363:513-522August 5, 2010).  By the way congratulations to the authors on such important study.  (This comment is about the irony of disclosures not the value of their paper, which is truly outstanding).  The methods did not tell me how to account for the disclosures (Top figure).  So if you look at the results section there were 1006 words.  The disclosure text was almost as long (798 words).  There must be a way!  After many hours I took it upon myself to calculate the significance of disclosures and suggest the following formula for correction.

For the sake of providing due credit I thought about the "CANDAB correction" which is an acronym with the first letter of some of the most intense pharmascolds. (HINT:  Solving this is much simpler than my formula).  So here it is finally! Do this for every author.  X stands for the age of the author, y is the net compensation received in the last 12 months and n is the number of times they have participated in CME or consulting.  You add it all and move the p value as many decimal points as the result shows to know the truth. 

 Late breaking news: I heard JAMA will include automatic computation in their website.