Tuesday, June 07, 2005

Scientific Integrity.

In relation to the earlier post, I'd like to elaborate a bit more on why it's important for scientists to write into journals such as Nature, so their work can be critiqued by other scientists. Science ultimately isn't magic, it relies on other groups being able to faithfully and accurately replicate the results of an experiment. If I discover nuclear fusion in my basement for example, it's important that I am able to tell other people how to replicate and verify that my claims are true. If it turns out that I'm full of rubbish, peer review is one of the most efficient ways of discovering this. This is because the real importance of peer review isn't so much as announcing what you have done; it's more important that your work can be analysed and criticised by other scientists.

Unfortunately, sometimes this system is abused and companies deliberately attempt to not put all of their results in for such examination. An example of this occurred when GlaxoSmithKline (GSK), a large pharmaceutical company, was sued for fraud based on accusations that it had not released all the results of it's clinical trials to the public. The drug in question was called paroxetine, which was originally made to treat depression but was found to be ineffective and to make matters worse, increased the chances of treated individuals committing suicide. The trial in question was not released however until they sought approval for using the drug to treat another disorder. This unleased a proverbial storm of criticism and undermined public confidence in pharmaceutical research, and very possibly scientific research in general.

In many respects, this has resulted in a general review of the peer review process, particularly in how 'transparent' drug companies are about publishing all of the results of their trials. This highlights the issue as to what scientists should be presenting for critique in the literature. Negative results may not always be useful, indeed it might not even be relevant to demonstrate every failed trial or no result that comes along. At the same time, such openness would at least provide a method for analysing the experimental techniques and increase public confidence in these companies.

Unfortunately, it doesn't appear as if the lessons that were learnt with paroxetine have been fully realised. A recent report in Nature noted:

Drug giants fail to name compounds in trial database.

But the editors say that drug firms are inserting a "meaningless phrase" instead of the names of drugs, so patients aren't getting the full picture, including any negative data.

The New England journal's editor-in-chief, Jeffrey Drazen, says that Merck, GlaxoSmithKline (GSK) and Pfizer in particular "didn't meet the sniff test" in a review conducted early this month by Deborah Zarin, the database's director. Zarin found that specific drug names were missing in scores of trials, which used the phrase "investigational drug" to describe their products. Drugs weren't named in 36% of 75 Pfizer studies reviewed, in 53% of 55 GSK trials, and in 90% of 132 Merck trials.

Once again casting doubt upon the integrity of the scientific process and if these companies are being readily transparent. It is vitally important, particularly in keeping the publics confidence, that results are made available and people know exactly what they are getting into with such trials. In the end however, it is ultimately on the scientific community to regulate and ensure that proper ethics and scientific standards are applied in drug trials, especially when individuals lives are at stake.