|From the WSJ article.|
The whole point of science is that it must be reproducible. So you do something, you document your data, and you publish your conclusions. Then others can process your data and see if they get the same results. They can collect their own data and see if they get the same results as you did.
But apparently this is mostly no longer true for academic medicine.
As you can see from the image above 64.2% of 67 studies from academic journals examined by Bayer were not reproducible.
I wonder why this is?
The WSJ offers some ideas - the most generous being that the lab equipment and techniques varied slightly between the published study and the equipment used by the lab attempting to reproduce it.
In this modern age of billion dollar research budgets and projects, post doc academic researchers are somehow using faulty or, alternatively, magically-capable lab equipment not available to Bayer? Techniques not written down that cause their academic research to work in their own labs?
Possible but this seems unlikely.
Most likely is cheating, incompetence or "publish or perish" - and none bodes well for us - the recipients of the results of these studies.
Now let's look at submissions to the FDA. A study, related in the same WSJ, indicates that of 33 submitted drug trials where the submitting company must supply all study data only about 75% of the data was ever supplied.
Most surprising of all was that the data not submitted was, gasp, unfavorable with regard to the drug in question.
My own personal knowledge of the FDA approval process is that virtually all companies are submitting data via complex, powerful software systems that track all data, responses, comments, notes, results and other data associated with the studies.
Perhaps these programs are just "losing" the 25% of the data that reflects unfavorably on the drug being tested.
So this says the problem is not just academic and its not just corporate.
So what is common to both?
The human element. The people. The researchers and their training.
The first question you have to ask is how long has this been going on?
The next question is does the processing of training these researchers have anything to do with it?
Certainly there is and has been a history of falsified research. A quick check of Wikipedia reveals a long and detailed document on the history of academic cheating.
Interestingly cheating is well documented into ancient Chinese times (thousands of years ago) for civil service examines - even when the penalty for cheating was death.
More interesting is the statistic that about 70% of all high school students admit to cheating, about 35% of the teachers do as well and about 56% of MBA's.
Adding the 70% of high school students to the 56% MBA cheaters, divide by two, and you get 63% - just about what the level of cheating would appear to be for academic studies that cannot be repeating. Probably about the mix of college business and science graduates and their average rate of expected cheating...
What an interesting coincidence!
More than likely the "cheating" has become a lifestyle choice all through college and into work - at least in the healthcare and drug industries.
(Not that this is new - when I was in college in the mid 1970's one of the benefits of joining a fraternity was access to its library of papers for submission as your own.)
So the Bayer results would seem on par with what is acceptable at today's academic institutions as far as cheating is concerned.
Sadly for me the nuns eliminated any desire to cheat from my life in first grade (thank you belt and yard stick).
At any rate these statistics should leave little wonder about the reliability of today's medical knowledge.
(So, in case you've an academic education and can't figure it out, about 1/2 of everything we read, are told, or think we know if we got it from somewhere else is probably wrong....)