courtesy: howorthgroup
There is a growing concern, in the medical field, that clinical trials published to aid doctors in making decisions to help their patients might be rigged.
If that sounds like a loaded word, it is. But this evidence, according to a new book by a specialist in the field, Ben Goldacre, is raising some very important questions, and just as many ominous doubts.
But as Dr. Goldacre relates in his book, this is not a new fact. As a matter of fact, it has been known for decades.
But let's start from the beginning.
There are many different types of trials. In the field of pharmaceutical science there are industry funded trials and independently funded trials.
That alone, should send people running for the hills. Because, yes, the industry trials, although pertinent and necessary, are made by the same people who make the medicines they 'try', and we, the patients rely on their findings to decide whether or not to take a medical compound.
But there is a problem with industry funded trials: and that is that they are much more likely to produce a positive result than a trial conducted by anyone else.
And that should worry a whole bunch of people.
In the book, Goodacre cites the recent case of three researchers from Harvard and Toronto who found all the trials for five major classes of drugs, then measured a key feature: were the trials positive for the drug, and were they funded by industry?
The results were jaw-dropping: of 500 trials they found, 85% of industry funded studies were showing positive results for the drug, in other words, it concluded that the drug was effective and safe, while only 50% of independent studies found their trials to show positive results for the same drugs.
Then in 2007, researchers focused on a drug that has become nearly ubiquitous in modern medicine: statin compounds. These drugs, who are given routinely in the management and prevention of cholesterol, are one of the most commonly used drugs in the world.
The researchers comparing trials in the case of statins found that of 192 trials, the industry funded trials were 20 (yes that's twenty) times more likely to show the drug positively compared to non industry trials.
Just for good measure mr. Goodacre cites another study done in 2006, which focused on pshychotropic drugs over a 10 year period. in this case the researchers found 542 trials in total. Industry trials were 78% positive, while independent trials only showed the drug as being effective in 48% of the trials.
This is a dismal report, and an even more dismal reality. Because one must realize that the trials that have been examined were the ones that were easily found.
And that's another ugly reality of clinical trials. Many trials are never published, and invariably those are usually the ones that show negative results for one drug or another.
The only way to truly examine a drug's effectiveness is to compile data from all trials, published and unpublished, and create a 'systematic review', that is a review that looks at all possible trials ongoing, finished and scrapped. Only when we compile all the data together the real picture comes into focus.
But there is a real obstacle to doing this. First of all trials that are not published are not easily found, and that is by design. And most of what is found easily as published is positive, also by design.
So researchers have to either be very lucky or dig themselves into files from agencies, such as the FDA or other sites where the trials might have been deposited. One thing is for sure. Industry will not publish a list of unpublished trials.
But in 2003, when some of the systematic reviews were finally completed, they showed that industry funded trials are 4 times more likely to produce positive results.
Now you would think that such a finding would have shaken the sector or the regulatory agencies to the core. But it didn't. Regulatory agencies are powerless to reign in such a problem, and without stricter laws, the pharmaceutical sector does not have to change a thing.
More importantly, one of the researcher's finding was that in some cases, specifically with some medications whose results were forwarded to the American College of Rheumatology the case studies did not contain all the raw data, ranges or all the details usually found to buttress the conclusions. There was only the conclusion, because nothing else was needed, since all the conclusions were favorable. It read:
The results from every RCT (trial) favored the drug of the sponsor.
What that meant was that since every industry sponsored trial has a positive result, that all you would need to know was whether it was industry funded to know the outcome.
When looking deeper into this matter, and looking for the reason why the industry trials were all positive, they discovered that trials are sometimes flawed by design.
Yes, you read right. By design. But this is nothing new in science. A scientist can skew the results by choosing parameters that he knows favor a positive result.
For example, in a clinical trial you can compare your new drug with something you know to be worthless, such as an inadequate dose, or an existing drug that is proven to be ineffective, or a placebo. You can also choose your patients carefully, so that they are likely to have a better response to your treatment. And you can also stop your trials when the results look positive, since the longer the trials run, the more probability for the figures to change exists.
These expedients are known as 'statistical poison'.
Another method employed by the industry is to fail to publish the result of a trial if it is unflattering to their product. But this habit of withholding negative information can have serious consequences, which we will deal with later.
Mr. Goodacre, a doctor himself, makes a case for himself by citing a commonly used drug, Reboxetine. The doctor did his due diligence, and read the trial data carefully, all of which pointed to good efficacy for the drug. So he prescribed it to his patient.
However, in 2010, a systematic review of the drug proved that it was no more effective than the previous less expensive drug, in treating the ailment. The doctor in fact found that of the trials conducted, seven had been done against a placebo. Obviously all patients showed positive results.
But when six more unpublished trials were were found that had been conducted on ten times more patients, all of these trials showed that the drug was no more effective than a sugar pill.
What's a doctor to do then? This doctor dug some more, and found that not only the medication was ineffective, but that it had more side effects. The side effects were marked as not unusual in the seven published trials, but in the unpublished ones, they showed significant side effects.
If a doctor can be fooled, how can the patient be aware, or even able to be informed?
What is worse, again, is that no one broke the law. Negative data goes missing all the time, in all areas of science, and regulators and professional can do zip about it.
And as we cited before, evidence that is withheld can be dangerous.
One example is that of the infamous trial of TGN 1412. In March 2006, in the UK, a trial for a new drug took place in London. Twelve people were given the drug, all at once, without staggering for possible side effects.
Withing a week all patients were dying and were saved by a veritable miracle, but most of them had amputated limbs, because the drug caused a severe adverse reaction that led their immune system to shut down organs, and their limbs went into gangrene.
This episode could have been avoided if the previous trial for a similar trial conducted ten years earlier had been published. The trial would have been avoided or at least conducted in such a fashion that would have minimized the harm it could do.
These events, in which a drug is approved for which there are negative trials that remained unpublished have meted many victims.
A famous case cited was that of a large scale trial for Lorcainide, a drug that was given to people who had heart attacks. Because of the lack of information on the effectiveness and adverse reaction of precursor drugs, 100,000 people died after taking the medication.
And what is more disturbing, researchers can bury any result that are not to their liking.
One of the solutions to the problem offered by several scientists is to create a registry of all clinical trials, and demand that people register their study before they start, so that one can easily make reference to it, or find it, and order that the results be published, good or bad.
But such legislation implies a willing government body who is not influenced by the sector target of the legislation. And that alone, could prove impossible.
Our regulators cannot regulate if they cannot identify or access data that has been withheld by the pharmaceutical companies on the drug's effectiveness or adverse effects.
In some cases, the same regulators who are supposed to police these problems, refuse to share the information they have. In the US, the FDA has a wide cache of unpublished trials. But without knowledge of their existence, the data they contain never sees the light of day.
Source: Scientific American; 2.17.13
Source: Ben Goodacre: BAD PHARMA (Faber and Faber.2013)
No comments:
Post a Comment