Slashdot Mirror
Merck's Deleted Data
An anonymous reader wrote to mention a Forbes article describing a drug study tampering proven by software. From the article: "A top editor of The New England Journal of Medicine says that he was stunned to find out that data linking Vioxx to cardiovascular risk was deleted from a major study his journal published five years ago--and that it appears that Merck researchers may have deleted that data ... When you hover the cursor over the editing changes, the identity of the editor pops up, and it just says 'Merck'"
You missed the point entirely. The Journal was given a hard copy of the study by Merck four months before the issue came out. This was in the days before the publication worked from digital submissions.
Merck knowingly gave the Journal incomplete data and the editors have only now discovered the discrepancy by going back and examining the original computer document.
Now why would you allow to publish such inconclusive studies at all ? Is this journal peer-reviewed ? It would be interesting to see if they also publish the comments from the anonymous reviewers ? Did they agree about the paper before it got published ?
Do you expect scholarly journals to reproduce all experiments before publishing the data? It's only inconclusive and misleading because some data was conviniently deleted. There is no way for a journal to know this without reproducing the study.. and that is not the purpose of "peer-review" in the "peer-reviewed journal". The work, itself, is taken on honor, and the "peer-review" is there only to make sure the experiment, as explained, is scientifically interesting and accurate... the data itself is taken at face value until it is either independantly confirmed or denied.
where i work, we enforce use of the Remove Hidden Data Tool to prevent this happening
we once got some documents from DOJ that were supposed to go up on our website that had obvious edditing changes in them
On Slashdot, we often are willing to relate anything anti-MS to funny, insightful, interesting, or underrated.
Excuse my speling.
Making The Bar Project
Yeah... I work in data management for a central lab, and I don't think data can really ever be "deleted". We certainly never delete ANY data (not for at least 15 years anyway), and as a central lab, we don't answer to a single sponsor and DO answer to federal regulations and are subject to random audits at any time. So, we would have the data, and the CRO (contract research organization) would have the data at least. What Merck or any other drug company does with the data is not really our concern but it's effectively impossible for clinical data to really be wiped out totally because it will be in the hands of many independent organizations.
I'm just a database monkey so I'm pretty ignorant about the process with these journals and such, but it sounds like the data was deleted because it was past a cutoff date; maybe at that level thats a no-no, but for us it's pretty much standard procedure that if we have data that has some outstanding issue and we are waiting for some confirmation/reconciliation, we just suppress it until the issue is resolved, which is preferable to sending erroneous data.
Also, to troll it up: maybe if it was possible to recall a drug without necessarily opening up yourself to billions of dollars of liability lawsuits, drug companies would have more incentive to take recall actions sooner rather than waiting until the evidence is overwhelming. By making the price of admitting there MIGHT be a problem with the drug so high, it's inevitable they would try to delay a recall for as long as possible. I'm not defending it - I'm saying it's inevitable and logical. The tort system takes it's toll in lives as well as dollars.
If you RTFA, you will notice that the said article was submitted in 2000.
When I heard this news about Merck, my impression was "deja vu". I recently read a book entitled "Overdosed America: The Broken Promise of American Medicine" by John Abramson. The author is not some quack. He teaches medicine at Harvard Medical School. He makes a strong point that drug tests which used to be funded by unbiased sources and now mostly funded by drug companies. They tailor the tests and massage the results to arrive at the conclusions that they are looking for. Before you take any prescription medicines, I strongly recommend that you read that book!
The former CEO (left early 2005) walked off with some $17 million in compensation. Merck could be penalized hard but the guy who was in charge will have absolutely nothing done to him.
A radio maverick jumps to internet only. The Future of Rock n Roll
While this story makes Merck look bad, idiotically bad, on closer inspection there isn't as much here as you'd think. The data in question are three heart attacks in the final weeks of the VIGOR trial. But the adverse cardiovascular event data in the paper, as published, didn't reach statistical significance, and they don't seem to reach it with these added in, either. On top of that, these data were submitted to the FDA during the drug's approval process, and (according to Point of Law) are on the Vioxx package insert itself.
So, you're saying that if the company can select a cutoff date that's too soon, they can avoid a lot of hassle using your system?
You are right in that optional stopping points are one of the most common methods of medical study fraud.
However, according to the article, internal Merk documents indicate the heart attacks occurred 5 weeks before the study is over. At the end of the article, they quote Merk officals who are not under oath as telling reporters that they were outside the study, and that they reported the deaths to the FDA anyway.
At this point in time, we have 1) a motive, in the money Merk made 2) a means to do it and 3) evidence of Merk officials lying in testimony and getting caught in it from the transcript of the trial in Texas. If you are still posting "in defense of Merk . . . " comments at this point, you must be a Republican.