Leaving Us to Our Own Devices

Alexander Volokh
RT: The Journal for Respiratory Care Practitioners, December/January 1995

Each year, thousands of US residents die because the Food and Drug Administration (FDA) fails to evaluate potentially life-saving treatments quickly enough.

The medical device approval process is slow and uncertain, evidenced by the more than 1,100 device applications currently being held for processing. In fact, medical device regulation is so burdensome that 45% of manufacturers are considering leaving the country.

FDA Commissioner David Kessler's proposed solution of levying user fees -- some as high as $52,000 -- on device manufacturers will not solve the problem. The proposed fees would be used to increase FDA staff and hopefully reduce the backlog. But even if this tax on medical innovation produced the intended results, thousands would still die needlessly.

The backlog problem

User fees would work primarily by reducing the turnaround time for device applications. Under current law, premarket approvals (PMAs) for new devices should take fewer than 180 days, but in fact they take an average of 437 days. Premarket notifications for devices similar to pre-existing ones, or 510(k)s, should take 90 days. Instead, they lag to 173 days.

Proposed fees range from $3,200 for 510(k)s to $52,000 for PMAs ($26,000 for small businesses), with intermediate fees for PMA supplements. They would fund, among other things: device review and approval activities; policy document and test standard development; application counseling for manufacturers; and postmarket activities such as reviews of adverse medical effects.

The dark side of FDA

Kessler paints a rosy picture of the FDA and would like us to believe that user fees are needed to protect us from snake-oil peddlers and mad scientists. But the FDA has a dark side -- just as real but not so brightly illuminated.

In February 1993, for example, the FDA banned US testing of the Ambu CardioPump, a toilet-plunger-like CPR device. Mandatory in French and Austrian physician-staffed ambulances, use of the CardioPump is encouraged just about everywhere and may increase CPR survival rates by 10% to 50% (Wall Street Journal, August 2, 1994). Yet the FDA, calling it a "significant risk device," requires informed consent for its use. Since cardiac arrest victims cannot offer informed consent, the United States has no CardioPump. This policy's toll is estimated at 7,000 deaths per year (Wall Street Journal, August 2, 1994).

Similarly, in April 1989, the FDA denied approval for the Omnicarbon heart valve, made by Medical Inc., Minneapolis, Minn. People with mechanical valves can suffer from thromboembolism and drugs that reduce clotting increase the risk of hemorrhage. The Omnicarbon valve reduces thromboembolism and hemorrhage rates by half; it has been available in some countries for the last 10 years and is now marketed in most of Western Europe, Canada, and Japan. Estimated toll: 900 deaths per year and untold unnecessary hospitalization costs.

For 9 years, the FDA has kept the Sensor Pad, which helps women detect suspicious breast lumps, off the market. This device, manufactured by Inventive Products, Decatur, Ill., is made of two plastic sheets with lubricant in between; like soap and water, the currently recommended self-examination method, all it does it reduce friction. While the Sensor Pad has been approved in most countries, the FDA requires that the company conduct exhaustive clinical tests to show that the product is not only a self-examination aid but can also diagnose breast cancer. The manufacturer has spent hundreds of thousands of dollars in legal fees and may have to close down the plant. This potential toll of this policy is unknown.

Close encounters of the first type

These case studies point to the fundamental problem at the FDA. When the agency evaluates a device, it can screw up in two ways. Either it can reject a life-saving device or it can approve one with unexpected adverse side effects. In the statistical business, we call those Type I and Type II errors. Now, 20,000 people who die because they did not have a good device are just as dead as 20,000 people who die because of a bad one. But politically, the second error will lead to public outcry and angry congressional hearings, while the first will produce no response at all. That is the reason why regulators consistently make Type I errors -- a problem public-interest attorney Sam Kazman dubs "deadly overcaution."1

With watchdogs like these, who needs evil scientists? The FDA is trying to give us safe and effective devices, but there is no such thing. Degrees of risk and types of benefits differ, and patients view risks and benefits in different ways. People with a high risk of dying in the near future, for example, are willing to take bigger risks. Given the broad range of willingness of accept risk, a one-size-fits-all strategy fits no one.

Band-aid reforms

Reform is in the air at the FDA, but will the changes proposed do any good? User fees may result in quicker processing of device applications. Even if they do, it is unclear whether better health care will be the result. The only applications that will be processed faster will be those of paying manufacturers; those who cannot afford the fees will be shut out. Many manufacturers are already shut out. As an entrepreneur who moved to Scotland to avoid the FDA approval process puts it, "If a young man wanted to start a heart value company today in the United States, forget it."2 User fees threaten small companies and entrepreneurial start-up firms. That is no mean threat when 87% of device manufacturers have fewer than 100 employees and 61% have fewer than 20.

But such considerations skirt the issue. User fees are a band-aid solution because they do not address the real problem. Quicker turnaround is of no use for devices that FDA would reject anyway. According to policy analyst Fred Smith, "A faster car will only make things worse if you're going down the wrong road." As long as drug and device approval is controlled by an agency that answers to a congress that answers to media sensationalism, the process will be dominated by sensational disasters. Thalidomide, the sedative that caused birth defects in the early 1960s, is one example. A bigger disaster -- less visible and therefore more likely to be ignored -- is the number of deaths from heart attacks or cancer or any of a host of other diseases that may occur because certain drugs or devices are experimental or only available elsewhere.

Three modest proposals

What, then, should be done? I offer three modest proposals. Backlog is a problem, not because the FDA is underfunded but because it tries to do too much. The first solution, then, is to return the FDA to its original mission: looking out for unsafe drugs and devices. Today, drugs and devices have to be declared both safe and effective. The safety part dates back to the original 1938 Food, Drug, and Cosmetics Act, passed in the aftermath of the sulfanilamide scandal. After the thalidomide experience, the 1962 Kefauver-Harris Amendments mandated that drugs also be tested for efficacy -- although the problem with thalidomide, which never got into the United States, was related to safety and not efficacy. The 1976 Medical Device Amendments, passed after the Dalkon Shield IUD fiasco, extended the safety and efficacy requirements to devices.

According to Herbert Burkholz, author of The FDA Follies, "Testing for safety is expensive and time-consuming enough, but an approval process that mandated testing for efficacy as well promised an exponential increase in both time and money, a promise that we have seen kept."3 Ultimately, the cost of more information outweighs the benefits. The FDA should test for safety, and leave efficacy to doctors and patients.

The second solution is to end the FDA's monopoly in the device-approval market. Regardless of the ultimate purpose of FDA testing, the agency should not have the final say on device approval. It should keep reviewing devices but in an advisory capacity. Unapproved devices could carry warnings -- maybe even a skull-and-crossbones -- but the final choice should be left to hospital staff, physicians, and patients. Physicians, whose reputations and incomes are on the line, would still have an incentive to act responsibly, and so unapproved devices would not be widely used -- but they would be available for those who situation is desperate enough for them to take the risk.

The solution would not eliminate the backlog, but it would reduce it. Moreover, ending the FDA's monopoly is necessary before user fees can even be discussed. Kessler claims that "user fees are based on equity: an identifiable beneficiary of a government service should pay the cost to the federal government of providing that service."4 But we would not pay for the privilege of having our books censored, and we should not pay to be denied access to devices we may want to use. If we cannot summon the will to remove the FDA's veto power, we should at least see to it that the agency allow approved nongovernmental organizations to take over the approval process.

This, the third solution, is similar to the approach used in Europe, where companies compete to inspect medical devices and examine product dossiers. This was proposed in the Bush administration's 1991 FDA reform package, which suggested companies pay voluntary fees to have their data reviewed in such a way. (Companies unwilling to pay could still use the slower, more conservative FDA approval process.)

The FDA is safe, we are sorry

With the FDA, all gall is divided into three parts. First, there is insult -- the loss of consumer choice embodied in the 1976 Medical Device Amendments. Second, there is denial, the FDA's refusal to acknowledge that it is killing anyone. And finally, there is self-righteousness, the FDA's insistence that we pay for the privilege of submitting our medical devices to its scrutiny. The FDA wants to take credit for keeping us safe, but as Burkholz warns, "It is no defense to mouth the obvious 'better safe than sorry.'"3 The best thing the FDA can do is leave us to our own devices.

1 Kazman S. Deadly Overcaution: FDA's Drug Approval Process. Journal of Regulations and Social Costs. 1990; 1:35.

2 George WW. Presented before US Congress, Energy and Commerce Subcommittee on Health and the Environment; July 14, 1994; Washington, DC.

3 Burkholz H. The FDA Follies. New York, NY: Basic Books; 1994:22.

4 Kessler D. Presented before US Congress, Energy and Commerce Subcommittee on Health and the Environment; July 14, 1994; Washington, DC.