Avandia - The Predictable Billion Dollar Mistake

Getting past scapegoats to system solutions

Another blockbuster drug with hundreds of millions of annual sales, taken by hundreds of thousands of patients, likely doing more harm than good - does this sound familiar?

Mounting evidence suggests that increased death rates, strokes, and heart disease are occurring in patients taking the diabetic medicine Avandia. The FDA is reconsidering the case of Avandia again this month. This is obviously a huge issue. Millions were spent to make this drug available, billions have been spent buying Avandia, and yet it appears, patients are worse off. How can this be?

In 2004, David Graham, then Associate Director for Science for the FDA, stated that the FDA was - "virtually defenseless" against another "terrible tragedy and a profound regulatory failure" like Vioxx. The proven concerns against Vioxx were the same as those now with Avandia - heart attack, stroke, and cardiac death.

In passing out blame, pundits often charge that the FDA is incompetent, or even worse, in bed with the greedy pharmaceutical firms. I believe that those answers are results, not underlying causes. Well-designed systems work despite a profit motive and entrenched or even shady bureaucrats. However, we now have a system that delivers predictable results from the system's design - biased research that favors expensive drugs whether they are beneficial or not. What happened?

Before the 1980's, the main sources of pharmaceutical research dollars were the U.S. government and private foundations. Improvement in medical knowledge was viewed as a public good and a driver of overall productivity and therefore supported. However, over the last few decades, a major shift in drug research funding occurred. Now most of the funding to develop new drugs comes from the pharmaceutical firms themselves. The stated purpose for this switch in policy was to make drug research more cost effective.

With the funding comes the power to determine the focus of research design. A cost effective bias with a profit maximizing mode seeks to answer the question, "How do we design the least costly study that will assist us in selling the most drugs at the highest price?"

The problem with this "cost effective" approach is that industry looks much harder for a reason to use a medicine than for reasons not to use it. It is incredibly inefficient for a company (although very beneficial for society) to drop a drug that does not work, after spending millions of dollars trying to prove that it does. There lies the rub.

There is good data to show that this industry bias influences results. A British Medical Journal article showed that industry-funded research was 4 times as likely to report data that was favorable to the drug under study. This efficiency bias is so strong that one begins to distrust the results of all industry-sponsored research. Thus the research becomes incredibly ineffective, no matter how cheaply it is done, by eroding the trust in the whole system.

A second important question for industry in creating a cost effective study would be, "How do we create enough data in the least costly manner so the FDA will let us sell the drug"? In the case of Vioxx, this included strategically ignoring some key data altogether. Billions of dollars of drug were sold before the importance of the "doctored" data was discovered.

The capitalistic fact is that pharmaceutical firms must make a profit or go out of business. What would happen if a drug company made the decision to be extremely careful so their research designs did not have a bias towards getting positive results? Unfortunately, over a short period of time, competition would assure that they would go broke. Let me see -- millions of dollars in profits versus going broke. It is easy to see how the studies are designed with a bias to get positive results.

To conclude, we have a system design that has delivered the expected results - biased research that favors expensive drugs whether they are beneficial or not.

In this scenario, the FDA is not the problem. Expecting the FDA to police the validity of research results against such hidden built-in biases is bound to fail. Putting up too many safeguards to protect against this bias creates expensive bureaucracy -- also not the answer.

To avoid future serial blockbuster-drug-gone-wrong disasters, a major overhaul in pharmaceutical research funding is necessary. It is time to push the pendulum back from industry-sponsored toward more publicly-sponsored pharmaceutical research. The methods to accomplish this are beyond the scope of this blog, but it is abundantly clear that the current system is not performing the goal of cost-effective discovery and promotion of the best treatments for our patients.

Expressly yours,

(If one wishes a more in-depth analysis of Vioxx and other drug research failures, I would suggest the very readable and well-documented book, Overdosed America, by Dr. John Abramson.)