Letting Big Pharma Review Its Own Drugs &#8212 What Could Go Wrong?

We can't trust drug companies to disclose product safety hazards when they stand to gain so much from fudging the facts.

A lab technician handles H1N1 specimens at a GlaxoSmithKline facility in Dresden. (Ho New/Reuters)

When British drugmaker GlaxoSmithKline (GSK) agreed to pay $3 billion in criminal and civil fines for illegal marketing of several drugs and hiding safety from the U.S. Food and Drug Administration (FDA), it probably seemed to many people like fitting punishment. It's the largest fine ever levied against a pharmaceutical company, and the civil and criminal charges against GSK included marketing its antidepressants Paxil and Wellbutrin for unapproved (read: unproven) uses, giving Medicaid false information about prices it was charging for the drugs, and failing to report patient safety data about diabetes medication Avandia to the FDA. But even $3 billion is probably not going to get GSK to change its ways -- and fines do virtually nothing to protect patients.

When it submitted false price information to Medicaid, GSK was harming taxpayers, a financial crime that is amply repaid by the $300 million the company is paying for that offence. But discouraging future bad behavior is a different story, because GSK is one of the world's biggest drug makers, with annual revenue of more than $40 billion. A $300 million fine starts to look like just the cost of doing business when spread over the period of time when the company was bilking Medicaid.

Fines are even less effective when it comes to preventing patients from being physically harmed. GSK is accused of illegally marketing two antidepressants -- Paxil for depression in adolescents, and Wellbutrin for weight loss, substance addiction, ADHD and other problems. Paxil has been shown to have significant side-effects, including triggering suicidal thoughts and behavior in some patients, and neither antidepressant is much more effective than a sugar pill in the vast majority of patients with depression. Fining the company after the fact doesn't do much for the patients who have already been hurt by their marketing practices.

That's even more the case for GSK's egregious behavior regarding its diabetes drug Avandia. When a new drug comes on the market, sometimes the FDA will ask or require that the manufacturer to do a study, called a post-marketing trial, to look for rare side effects and establish whether the drug is safe over the long term. That's important because often the clinical trials that are done before a drug is approved aren't big enough, and don't last long enough, to say whether a drug is safe to use over a long period.

With Avandia, GSK fudged the results of those post-marketing trials. The company claimed that its RECORD (Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes) trial showed no evidence that Avandia caused heart attacks or other cardiac diseases. But when outside researchers looked at the data they found massive irregularities and plenty of cases of harm, including patients who died and then mysteriously disappeared from the clinical trial records.

GSK hasn't admitted anything more than "inadvertently" failing to inform the FDA of problems, and the fine isn't going to protect patients from future Avandias. What we really need is a better system for tracking drugs once they are on the market.

Right now, we leave this crucial task to the very companies that have every reason to fudge data, ignore problems -- even deaths -- and keep doctors and patients in the dark, because they are making too much money to do otherwise. At its peak, Avandia earned GSK $2.2 billion a year. It was a phenomenally lucrative drug. The $250 million levied against the company for that part of the fine amounts to less than a slap on the wrist.

The solution isn't stiffer punishments for fraud -- we should take away the power to do harm in the first place. That means we need to stop giving drug companies the opportunity to sell out patient safety in exchange for another few months or years on the market. The simplest way to do that is to have automatic, external post-market surveillance for all new drugs.

Any new drug should be considered by both patients and doctors to be experimental for the first few months, or even years it is on the market, and the FDA should treat the approval of the drug as provisional. (The same goes for new medical devices.) For at least two years, any patient who might receive a new drug or medical device has to be fully informed of its known side effects, and of its experimental status. Data on those patients should be tracked (anonymously), and compared to patients receiving other treatments. Most importantly, automatic algorithms can search those data for patterns revealing any negative effects of the new product, so safety problems can be caught quickly.

An automatic review system would be a huge improvement on our existing review process. It would take the burden of reporting off manufacturers, and let them focus on developing new treatments instead of monitoring old ones. It would remove the financial conflict of interest that encourages companies to hurt patients in order to squeeze a few more months or years of sales out of a new product before problems finally surface. (The only reason Avandia's tendency to kill patients came to light as quickly as it did is because a sharp-eyed cardiologist named Steve Nissen at the Cleveland Clinic happened to be looking at European data that clearly showed the drug was dangerous.)

Perhaps most importantly, it would force us to admit that we don't truly know that our treatments are safe when we start using them -- and that would give patients a chance to be more informed about their choices. That seems like a better system that letting companies kill patients, and then pretending to fix the problem by slapping them with fines every few years.