The FDA: What Happens When You Starve the Beast

In October, I talked to a source inside the FDA who suggested that the agency was having a hard time keeping up with work-flow.  I quoted him on this blog as explaining that since the FDA has committed to reviewing applications for approval of a new drug within 10 months, drug-makers have been submitting “shabbier” applications that contain less evidence about risks and benefits.

“For the drug-maker it’s a gamble. The company is betting that, because we want to make the 10-month deadline, we won’t send the application back,” said the source. And often, he acknowledged, the drug-maker is right. “If you find a problem or there is something missing and it doesn’t seem terribly material, there is a tendency to overlook it. Because if you don’t it will just delay the whole process.”

In the past, he added, a company submitting an application knew that if the application wasn’t up to snuff, the FDA would send it back. But those standards have fallen: “Now we send it back [only] if it’s really crappy.”

Yesterday the FDA Science Board dropped a bombshell in the form of a report which suggests that standards at the FDA haven’t just fallen—they’ve fallen off a cliff. The title of the report says it all: FDA: Science and Mission At Risk. The problem, according to the report: a lack of funding. The Coalition for a Stronger FDA,  co-chaired by the last three secretaries of Health and Human Services (the department that oversees the FDA), says the FDA needs a 15 percent boost in funding per year for the next five years. 

Here are just a few highlights from the report:

• “The Information Technology situation is problematic at best—and at worst it is dangerous.”
• “The FDA has substantial recruitment and retention issues”.
• “Critical data…including valuable clinical trial data...are sequestered in piles and piles of paper documents in large warehouses."
• “The FDA has an inadequate and ineffective program for scientist performance."
• "The FDA has inadequate funding for professional development to ensure that staff maintain scientific competence."

William Hubbard, a former FDA associate commissioner who supports the Coalition for a Stronger FDA, told ABC News that the report stands out because of the "intensity of the feelings" expressed by the subcommittee.

"These people were horrified by what they found," he added.

The report found the agency's computer system, which tracks hazardous foods and drugs, is so antiquated, it's constantly breaking down. "Imagine having an e-mail system so old, they have to bring technicians out of retirement because current technicians have never seen equipment that old,” said Hubbard.

Many front-line employees don't even have a computer. Inspectors must write urgent reports by hand, which often end up lost in huge warehouses. The report does not blame the FDA but rather Congress, for continually cutting the agency's budget while asking it to assume more and more responsibility.

A leading FDA critic in Congress, Rep. Rosa DeLauro, said, "The essence is, send me a plan, ask for money, and we'll consider it. But they never ask for more money or authority."

The report blames Congress for not funding the FDA. But DeLauro seems to be suggesting that the FDA itself has been less than proactive in demanding the funds it needs.

Why would the agency be shy?

In recent years, not everyone in the Bush administration has been enthusiastic about having the FDA do its job. In 2001 Dan Troy became the agency’s chief counsel. A Bush appointee, Troy had been a longtime foe of FDA regulation. Just months before joining the agency, he had represented Pfizer in a battle with regulators. A U.S. News & World Report headline summed up his career change: “Mr. Outside Moves Inside: Daniel Troy Fought the FDA for Years; Now He’s Helping to Run It.”

During the first two years of the Bush administration, Troy operated in a power vacuum. The FDA had no permanent commissioner and “while the White House and Congress publicly argued over who should lead the agency it was being quietly transformed by appointees such as Troy, who needed no congressional confirmation,” The Boston Globe observed in December 2002. U.S. News corroborated the report, saying that Troy “operated as the de facto head of the FDA” between September of 2001 and November of 2002. During that time, the magazine added, Troy held “at least 50” closed-door meetings with representatives of drug companies and others regulated by the FDA. Troy resigned his FDA post without explanation on November 16, 2004.

In 2002, while Troy was running the agency, it looked as if Alistair Wood, a respected clinical pharmacologist at Vanderbilt University, was close to being nominated as FDA chairman. But some industry executives—along with some members of the Wall Street Journal’s editorial board—objected. When Senator Bill Frist was asked why Wood was dropped from consideration for the top post, he summed up the opposition: “There was a great deal of concern that he put too much emphasis on safety.”

Since then, drug recalls, failed defibrillators, and questions about many other products that the FDA has approved suggest that it needs the resources to put a new “emphasis on safety.”

Comments

[Gregory D. Pawelski]

I sent Maggie some material that would probably give a few people headaches from trying to understand. To others, it would have brought back memories of 10+ stat. and measurement courses in grad. school.

Basically, the Bayesian method has to do with "conditional probability." The probability that event E (an effect) and C (a cause) will both occur is the product of the event C occurring, times the conditional probability of an event E occuring. An example: The chances of being hit by a truck and bleeding to death is the product of the probability of being hit by a truck and the probability of bleeding to death if you get hit by a truck. Well, so what?

The Bayesian method turns this calculation around. That is, it tries to calculate the probability of C, given that E has occurred. Bayes' Theorem is useful and reasonably well accepted for applications such as testing whether the assumptions of probability are valid. For instance, if you flip 100 coins in the air at once, and only get tails 5 times, you have to assume that they aren't "fair" coins. The whole idea of it all is to get more accuracy out of analysis.

[Maggie Mahar]

Gregory --
I agree with everything you are saying here--
I'd like to read more about the Bayesian method; can you direct me to something that would be accessible to a layperson?

[Gregory D. Pawelski]

You're going to have to change the culture and the incentives by which the FDA operates. You're going to have to establish different standards for how a drug gets approved. Not only should a pharmaceutical company show that their drug is effective, but they need to show that it works as well or better than other drugs that treat the same indication. You need to set up stringent standards of evidence that leads to the approval of safe drugs that actually have benefits to the population.

The pharmaceutical companies pay a substantial amount of money to the FDA at the time they bring a drug application for review and approval. But the real expectation coming from the company is that they've paid their money, now approve our drug. That's basically how the FDA reacts as well. Funding for the FDA should be independent of the industry that it's regulating.

Pharmaceutical money is influencing the decisions that get made, and it creates this incentive structure. You have this culture, you have these expectations, and then you have pressure from Congress. All of them come to a head at the FDA and all of those incentives are in the direction of "approving the drug." That's what happens, so I believe that the FDA is unduly influnced by industry and that undue influence is in part the result of industry money funding the FDA operations.

The FDA should be more vigilant in trying to eliminate conflicts of interest on panels. They argue that these individuals have the best knowledge of drugs, and that often is true, but it seems those who have the knowledge and connections ought to be in a purely information providing and not a voting capacity. With some effort, you could assemble panels with fewer conflicts of interest.

It was suggested last year to take a fresh look at how they evaluate new medicines and treatments for cancer. Rethink how we design and conduct clinical trials in the United States. To adopt a system called the Bayesian method, a statistical approach that is more in line with how science works. Bayesian approaches are being used routinely in physics, geology and other sciences. They've already put this approach to the test at M.D. Anderson in numerous phase I and II clinical trials.

Doctors want to be able to design trials to look at multiple potential treatment combinations and use biomarkers to determine who is responding to what medication. They want to be able to treat a patient optimally depending on the patient's disease characteristics. Using the Bayesian approach, trial designs exploit the results of ongoing trials and adapts itself based on interim results.

The most important factor in whether the Bayesian approach will gain acceptance in clinical trials reporting is whether the FDA will accept Bayesian approaches in making determination of safety and efficacy of new treatments. It is possible to reduce the exposure of patients in trials to ineffective therapy and reduce the number of patients required for a trial by using the Bayesian approach, thereby reducing the risk to patients and the cost and time required to develop therapeutic strategies.

Bayesian methodology is a rigorous science. It is no stranger to the technology of cell culture assays, a functional biomarker test. In fact, it is what gives credence to the accuracy of assay tests. The absolute predictive accuracy of a cell culture assay varies according to the overall response rate in the patient population, in accordance with Bayesian principles. The actual performance of assays in each type of tumor precisely match predictions made from Bayes' Theorem.

Maybe the standard for the drugs that come later should show that they're actually better than the drugs on the market because we've already got these other drugs that work. That would create incentives for industry to develop drugs that are better than the ones already there. Currently, the way the incentives are for industry, it's safer to do a "me too" drug, another drug in the same class. It seems like a lot of hype is wrapped up in the problem of so-called "me-too drugs."

We have too many "me-too" drugs. This problem relates to rising prices for new drugs. Scores of new cancer drug applications are for "me-too" drugs which might show only miniscule clinical improvement in trials, yet they somehow gain approval. Since they are marketed as if they were important new breakthroughs, they have very high prices.

In academic centers, the patients are entered into clinical trials of square peg in a round hole therapy. In the private sector, patients are treated with drugs which generate the most revenue for the treating oncologists, overtreat with infusion chemotherapy, and encourage the patient to receive 2nd, 3rd, and 4th line chemotherapy, regardless of the likelihood of meaningful benefit.

It would be helpful to require the FDA to test new drugs against other drugs on the market (head-to-head comparisons) rather than against placebos. And tighter regulation of advertising. Advertising should be reviewed "before" it is aired, not weeks "after."

Pressure should be put upon Congress to implement these changes to the culture and the incentives by which the FDA operates.

[maggie mahar]

P.S. -- I didnt' mean to suggest that the FDA insider who I spoke to yesterday is wrong in defending the agency . .

First of all, if you're inside the FDA, this report is not a huge shock. The media presents it as extraorindary news because they haven't been following the FDA.

Secondly, if you're working at the FDA and trying to do your job, you can't tell yourself "this is just a disaster." If you did, you would walk away.

So I didn't mean to criticize an insider for hanging in; he or she is one of those knowledgable civil servants who keep things from falling apart.

[maggie mahar]

Gregory--
Very informative comments-- thanks very, very much.

I agree with virtually everything you say.

When David Kessler took over as head of the FDA in the early 1990s I wrote a string of stories about the agency, interviewing him and others who were there.

And since then I have kept up with some sources inside.

I totally agree that these are some of the brightest, most dedicated people in govt'. And let me add, many career civil servants are in fact very intelligent people who are working hard to protect people and do the people's business.

It's the political appointees who were appointed purely for political reasons who tend to have their own agenda.

Unfortunately, sometimes the poltiical appointees are in charge, and morale among the hard-working, knowledgable career civil servants falls.

This is what has happened in recent years at the FDA. Yesterday, I talked to my newest source inside the FDA who defended the agency,saying that it sounded like the report made things sound worse than they were . . .

At the same time, he said that they are understaffed; he gave the IT system a "B" and said "Morale is not great. But it's not as bad as it is at Homeland Security."

Wonderful.

This, of course, is not what we want to hear about the agency responsible for protecting the nation's health. . .

Regarding the balance between access to new drugs and risk, I'd say this: The manufactuers of new drugs do a very good job of pushing for access.
That's why I think the FDA should be putting more emphasis on safety.

You need people on each side of the table--the industry pushing to get its most innovative products to market, the FDA on the other side of the table, worrying about safety and protecting the public..

From that dialectic, you are likely to get a balance.

The problem now, as you say is that "The Prescription Drug User Fee Act has worsened the structure of the FDA by allowing drug companies to pay money to the FDA so they will review and approve its drug. How much more of a conflict of interest can that be"

AS a result, it seems that now everyone is on one side of the table.

You also point out that "People who approve the drugs also oversee the post marketing regulation of the drug. When they see there is a safety problem with it, they are very reluctant to do anything about it because it will reflect badly on them."

I hadn't thought of that, but it makes perfect sense. This is not about corruption, but it's just that people are reluctant to realize that they made a mistake, particularly when that mistake could threaten the health of many people.

You are right: the people responsible for keeping an eye on the drug after it gets to market should be a different group of people.

I also agree that different drugs should be held to different standards.

I was all in favor of "fast-tracking" legitimate AIDS drugs in the late 1980s. We were facing a plague. The patients had little to lose; most understood that they were guinea pigs and were happy to try to help by becoming part of the experiment, even if the new drug wouldn't save them.

IF you have any more information on these structural problems within the FDA t--or sources within the FDA who could talk about this--please get in touch with me.

Thanks again -

[Gregory D. Pawelski]

The FDA is really one of the better agencies in government. Dedicated, smart, hard working group of people. They want to protect the American people from
harm and, at the same time, get the drugs out to the American people that have the potential to really help them.

The FDA has more often been criticized for being too restrictive than for being too permissive. The trade off is between access to an effective drug versus exposure to unforseen toxic effects.

The truth is that drugs can't be said to have passed the test of time until they have passed the test of time. I do think that a drug for arthritis (e.g. Vioxx) has to be held to a different standard than a drug for erectile dysfunction (e.g Viagra) and both need to be held to different standards than a drug for pancreatic cancer.

But it will always be a trade off between access versus risk. And the FDA will periodically be criticized for coming down too hard on one side versus the other.

There are structural problems that exist within the FDA. People who approve the drugs also oversee the post marketing regulation of the drug. When they see there is a safety problem with it, they are very reluctant to do anything about it because it will reflect badly on them.

The Prescription Drug User Fee Act worsened the structure of the FDA by allowing drug companies to pay money to the FDA so they will review and approve its drug. How much more of a conflict of interest can that be?

Add the burden of over-extension. A very small proportion of drugs represent a new drug that hasn't been marketed before. Most of those drugs are no better than the ones that exist. The really "breakthrough" drugs, the ones that really make a difference in patients' lives and represent a revolution in pharmacology, are maybe one or two a year. Most of the drugs aren't breakthroughs, but they get treated as if they are.

Scores of "new" cancer drug applications are for "me-too" drugs which might show only miniscule clinical improvement in trials, yet they somehow gain approval. Since they are marketed as if they were important new breakthroughs, they have very high prices.

Case in point with cancer drugs. The FDA's approval of Bristol-Myers Squibb's Ixempra harkens back to the old-school for breast cancer patients who aren't responding to other treatments. "Try this!" Ixempra would be used alone or in combination with Xeloda, in patients who have failed two or three other chemo drugs, instead of identifying the most effective chemotherapy for the individual cancer patient, the first time around.