The Food and Drug Administration
The FDA Meets the 21st Century
by Thomas Maeder

MOST PEOPLE don't know what the FDA does because the FDA usually does it so well. We don't get up from dinner saying, "Thank god I survived another meal!" We take for granted that our food is safe. Drugs, cosmetics, and medical devices generally do just what they're supposed to do, safely and consistently. When something goes wrong, it is quickly and unobtrusively caught, contained, and corrected.

What lies behind this uneventful experience that we in America unwittingly enjoy is, however, extraordinary. The U.S. Food and Drug Administration's responsibilities are dizzying in their scope, and touch countless aspects of our daily lives.

You wake up in the morning and brush your teeth with a toothbrush (a medical device) and fluoridated or unfluoridated toothpaste (an over-the-counter (OTC) drug and a cosmetic respectively). Bath soap, deodorant, shampoos and hair conditioners, hair dyes, shaving cream, eye liner, foundation, rouge, lipstick, nail polish, shaving cream, cologne, perfumes, breath fresheners, foot powders, depilatories, mud packs, and suntan lotion are all regulated by the FDA as cosmetics, though suntan lotion with sunblock is an OTC drug.

Almost everything at the breakfast table, from cereal and the labeling on its box to milk, sugar, and out-of-season imported South American fruit is regulated and inspected by the FDA at many stages during production and distribution. Even pots and pans, plates, cups, forks, spoons, knives, and food packages themselves fall under FDA jurisdiction, because if substances leach out of them and contact the food, they become in that instant food additives, regulated under the Federal Food, Drug, and Cosmetic Act - though for practical purposes the agency exercises this authority only in rare cases where, for example, lead or cadmium from ceramics pose significant health hazards. The water from your faucet is regulated by the Environmental Protection Agency, but bottled water is FDA.

If you have an infant, his or her bottle nipples, formula, and teething rings are all under the FDA, though not pacifiers, which fall under the Consumer Product Safety Commission, which began its existence more than thirty years ago as part of the FDA, then was spun off as a separate agency.

If you feed your animals before leaving for work, pet food and medicated bird seed come under FDA authority, as, in fact, do psittacine birds themselves - parrots, cockatoos, macaws, and parakeets - because of the risk of psittacosis. The FDA even regulates turtles, and bans the sale of those smaller than four inches long, because the Centers for Disease Control and Prevention determined that it was impossible to guard against salmonella.

Warm your coffee in the microwave or turn on the television to check the weather and traffic reports, and you have two more items that fall under FDA's jurisdiction over radiation emitting devices, a category that also includes laser pointers, surgical lasers, the price-scanning devices at supermarket check-outs, and airport security X-ray machines.

Throughout the day, virtually everything you eat or drink is regulated by the FDA. Technically the agency's authority covers only items moved through interstate commerce, though, as one field investigator mentioned to me, one is hard pressed to find any food at all - even homemade cookies sold at a bake sale - that does not contain at least one ingredient that had crossed a state line, whether sugar, flour, vanilla, or food colorings, a whole other area of FDA authority.

Sometimes another agency's more specific authority takes precedence over that of the FDA, so that, for example, the U.S. Department of Agriculture inspects most, though not all, meat and poultry. Cattle, sheep, swine, goats, and horses are USDA, though the FDA retains responsibility for bison, rabbits, deer, moose, game and zoo animals. The USDA inspects domesticated fowl such as chickens, turkeys, ducks, geese, and guineas, though wild ducks, wild turkeys, wild geese, and emus are FDA, and while the two agencies share jurisdiction over shell eggs, the FDA is charged with insuring egg safety. Sometimes the boundaries are unfathomable and comical. Open-faced sandwiches shipped across state lines to lunch carts or cafeterias are regulated by the USDA, while closed face sandwiches are FDA. Cheese ravioli is FDA; meat ravioli is USDA. A pepperoni pizza migrates from FDA to USDA jurisdiction if the wet weight of meat exceeds 3% of the pizza's total. Hot dogs and chicken noodle soup are USDA. Beer, wine (unless under 7% alcohol), and liquor fall under the specific authority of the Bureau of Alcohol, Tobacco, and Firearms, though if you drink to excess and are pulled over by a cop, the breath alcohol test system that determines your fate is regulated by the FDA.

Visit the dentist, and all of the instruments as well as the operating light, amalgam, false teeth, abrasives, cements, and dental floss fall under the FDA.

A doctor's office or hospital is a hotbed of FDA regulation, from every drug and solution to all tongue depressors, ice bags, hot water bottles, examination gowns, patient mattresses and covers, thermometers, canes, crutches, wheelchairs, surgical instruments, sutures, eye charts, bone cements, casts, artificial joints, diagnostic devices, monitors, drug tests, centrifuges, blood collection tubes, electrocardiographs, pacemakers, surgical gloves, face masks, operating tables, hearing aids, and even data processing equipment used for clinical purposes, and refrigerators and freezers used to store blood.

At your local pharmacy, not only every prescription drug and non-prescription remedy from headache pills to toe fungus treatments comes under FDA, but also the boxes of bandages, vitamin pills, cosmetics, candy bars, gum, and almost everything else in the store except toys (Consumer Product Safety Commission) and greeting cards (nobody at all).

Despite the agency's responsibility for so many of the most important products in our lives, the FDA suffers from a chronic, acute shortage of money and a strange, unacceptable lack of authority in many key areas, for which it compensates with hard work, ingenuity, and unofficial triage of its duties. Oddly, if the FDA suddenly ceased doing its job and we were faced with the awful consequences, with deaths from tainted and improperly handled food, allergic reactions to undeclared ingredients, medical casualties from useless or dangerous drugs, defective devices, and haphazardly collected and transfused blood, a horrified government would doubtless be moved by the catastrophe to appropriate funds far in excess of the agency's current budget. Congress might even hastily grant the agency powers beyond those given by the existing Food, Drug, and Cosmetic Act. Fortunately for the American people, this has not happened, though our relatively uneventful daily lives mean that we are protected by an agency stretched to the limit and beyond.

The FDA regulates a trillion dollars worth of products each year - twenty-five cents of every consumer dollar spent in the United States. Yet the agency's total budget exceeded a meager one billion dollars for the first time in 1999, an amount less than half the budget increase for the National Institutes of Health that year, or of the annual foreign aid allowance to Egypt, and less than the cost of a single B-2 bomber. Though the 2002 budget has inched up to $1.35 billion, including over $180 million in industry user fees to hire additional reviewers to shorten new drug approval times, several drug companies regulated by the FDA sell individual products with revenues far greater than the FDA's entire budget: domestic sales of AstraZeneca's drug Prilosec, for the treatment of heartburn, generate over $4 billion per year. Attempts to secure additional funding for the agency are subject to strange and tortuous battles because, through an artifact of the agency's governmental lineage, it is part of the Department of Health and Human Services, but receives appropriations through the Agriculture committee. Though all government agencies complain of insufficient funding, in light of the critical nature of many FDA functions and the growing complexity of its tasks, it seems inexplicable that the agency's cost-adjusted budget and actual manpower have gone down.

Most of the FDA's money goes to salaries and rents for about 9,000 employees working at agency headquarters and centers in and around Washington, and in 167 field offices of various sizes throughout the United States and in Puerto Rico. Fewer than 1,100 FDA investigators must inspect about 120,000 domestic establishments, from mom and pop medical device companies in a garage to multinational food conglomerates and pharmaceutical giants. A lack of funding and personnel forces the agency to exercise inspectional triage, visiting companies at only a fraction of the legally mandated frequency. The problem has progressively worsened with flatlined hiring and explosively growing demands, particularly the tremendous increase in imports and costly overseas inspections. Investigators sample more than three million import entries ranging in size from a single item in an express mail envelope to entire freighters full of produce, and collect random samples of food and drug products to be analyzed by understaffed FDA laboratories, many of whose equipment and supply budgets are smaller than those of a single modest university research lab. Meanwhile, internal reviewers at the Center for Drugs evaluate New Drug Applications so massive that, before the days of CD-ROM and electronic submissions, they were delivered to loading docks by tractor trailer and transported internally by forklift, as well as smaller but more numerous applications for new therapeutic indications of previously approved drugs or approval of generic equivalents. When product tampering cases arise, such as the Tylenol poisonings or suspected cyanide-laced Chilean grapes, or when an outbreak of food poisoning occurs, whether botulism infected vichyssoise soup or tainted potato salad at a church social, the FDA field staff drops its routine duties and, working in coordination with local health authorities, rushes in to identify the problem and trace it back to its source.

Surveys regularly place the FDA among the most highly respected government agencies. Paradoxically, it is also one of the most regularly and vehemently attacked, sometimes simultaneously for diametrically opposite reasons. "One of our problems, and one of every regulatory agency," says Dr. Janet Woodcock, director of FDA's Center for Drug Evaluation and Research, "Is that our stakeholders violently disagree with each other, and their method of disagreeing is to blame us." When I first began studying prescription drug regulation and new drug approval in the late Eighties, a survey of articles in the popular press revealed that, at least in the area of pharmaceuticals, the FDA was almost universally criticized - about half the time for withholding beneficial drugs from sick patients through a needlessly protracted approval process, half the time for allowing greedy drug companies to rush inadequately tested, dangerous drugs to market.

Debates over agency policies towards dietary supplements and alternative medicine have generated more mail to Congress than any other issue, including Watergate, and the FDA routinely finds itself embroiled in hard fought battles between consumers, industry, and special interest groups over issues such as gene therapy, contraception and abortion technologies, genetically modified foods, vitamins, alternative medicines, irradiated food, tobacco, the compassionate use of unapproved drugs, artificial sweeteners, nutritional labels, the right of consumers to make their own choices, and the role of any government regulation at all in a democratic, free market society.

None of us likes to have others presume to look out for our best interests. Long ago, however, society and the production of certain types of goods grew so complex and the chain from supplier to end user so long that government oversight of some essential products was imperative. Consumers, even the most cautious, best educated and informed, lack the information, time, skills, or resources to validate the purity and quality of food or the identity, safety, and efficacy of medicines and medical devices. A "let the buyer beware" philosophy would, and, in the past, often did result in many dead or disabled buyers - wary or not. Absent an organized system for collecting, analyzing, and disseminating information about product safety, even tragic consequences do not necessarily result in greater consumer caution. Tobacco was used for four hundred years before researchers generally accepted its dangers. Anything that does not produce harm with a frequency that can be noticed in one's own community may well pass undetected without a special means of surveillance. Even trained physicians have failed to discern within their limited practices serious drug adverse reactions that caused hundreds of deaths nationwide.

At its inception in 1906, the FDA, then part of the Bureau of Chemistry of the U.S. Department of Agriculture, was less of a scientific organization than a police agency focused on accuracy in labeling. Its intransigent tone was set by its first director, the irascible Dr. Harvey Wiley, a man so pure in his thinking that one critic claimed he would object to the confection called "lady fingers" unless they actually contained amputated female digits. When an exasperated ketchup manufacturer disputing Wiley's inflexible stance on preservatives like benzoic acid remarked that the Almighty, in His infinite wisdom, put more benzoic acid in cranberries than his company wished to add to ketchup, Wiley, aspiring to standards higher than those of the deity, retorted that cranberries would be more wholesome had God left it out.

As far as drugs were concerned, decades passed before the agency could demand proof of safety, and many years more before any meaningful requirement was established that drugs must actually do what they claimed. In the Thirties, the basic set of active drugs - opiates, quinine, digitalis, insulin, arsenicals, mercurials, aspirin, and a few vaccines and antitoxins - could be counted on the fingers of two hands, while most remedies relied in large measure on the powerful placebo effect. After the Second World War, however, the advent of highly potent new medicines like antibiotics, steroids, and tranquilizers inaugurated a true pharmaceutical revolution and transformed a profitable but sleepy industry into one of the fastest growing and most lucrative business sectors of all. The tragedies wrought in the Fifties and early Sixties by drugs like chloramphenicol and thalidomide, in which thousands of people were killed and maimed through the irresponsible marketing and use of potentially dangerous drugs, compelled the FDA to adopt a more rigorous medical and scientific approach to the evaluation and regulation of drugs, at the same time that the social climate of the time demanded higher standards of safety for all consumer products, and higher levels of accountability on the part of the corporations that made them.

Beginning in the mid-1960's with the appointment of James Goddard, the first commissioner brought in from outside rather than promoted through the ranks, and a physician with a public health instead of a chemistry background, the FDA began a protracted period of change and modernization that continues to this day. What had been a regulatory agency, grown moribund, accused by some of knowing how to do little more than board shrimp boats and search warehouses for insect parts and rat droppings, suddenly needed to transform itself into a scientific agency and grapple with subtle issues of risks and benefits, cause and effect. Yet science alone was not enough. "I went to FDA believing that the rigorous application of good science would make less controversial and better the decisions that FDA had to make in many areas," said Alexander Schmidt, FDA commissioner from 1973 to 1976. "Much to my dismay, what I learned was that, by and large, that was not true." Though the FDA prides itself on being a scientific agency, science is only one part of decisions that also involve ethics, emotions, politics, and the unavoidable influence of money.

The domain that the FDA regulates has drastically changed and continues to change at an ever increasing pace. Biotechnology, gene therapy, space age medical devices and biomaterials, and transgenic foods and animals pose challenges in simply keeping up with the field of knowledge. Progress in the field of pharmacogenomics suggests that in the not distant future we will have carefully targeted drugs for smaller and smaller therapeutic subcategories, even customized drugs for individual patients, which raises regulatory issues never dreamt of in the past. The globalization of the food, drug, and device marketplaces impose an increasingly unbearable burden on an already undersized and overworked inspection staff, which must now deal with escalating imports and costly inspections of foreign facilities. Meanwhile the FDA, historically a lofty isolationist agency with the power to demand that the rest of the world meet its standards, is now struggling with "harmonization," trying to establish acceptable mutual agreements with countries that often differ in their attitudes towards medical practice, acceptable risk, product liability, and human ethics. The Food and Drug Administration Modernization Act of 1997 (FDAMA) forced the agency to relinquish some of its hard-nosed attitudes and become more cooperative and friendly towards industry, a move that has combined distinct benefits with the fear that more of these benefits will accrue to industry than to the public the FDA was formed to protect. At home, the FDA's range of concern has ineluctably grown from mere oversight of labeling and policing of facilities to a role of educating consumers who want greater control over the treatments they receive, collaborating with industry to help devise better products and procedures, and seeking ways to guarantee that new drugs, once approved as safe and effective, remain safe and effective when placed in the hands of medical practitioners.

Using Safe Drugs Unsafely

In the United States we now take for granted the anomalous status of most important pharmaceutical products, which are given only under a doctor's supervision. Prior to 1938, however, American consumers could purchase any non-narcotic medicine without a prescription. Until that time, of course, aside from narcotics, relatively few medicines existed that provided significant risks or benefits. The inadequacy of the existing laws was highlighted in 1937 when an inept company chemist formulated Elixir Sulfanilamide-Massengill, an early sulfa drug preparation, in a diethylene glycol solution - antifreeze - causing the deaths of 107 people. Congress hastily passed the Food, Drug and Cosmetic Act of 1938, which demanded premarket proof of safety and required manufacturers to include appropriate cautions and instructions for use with every package of drugs that they sold. Acknowledging that some powerful newer drugs might pose risks when used by lay people without supervision, the Act carved out a category of "legend" drugs for which manufacturers were expressly prohibited from including instructions for use, and which could be dispensed only on a doctor's orders. The 1951 Humphrey-Durham Amendment formalized this class of prescription drugs, which grew to include virtually all potent therapeutics. The alternative to prescription medicines, over-the-counter (OTC) drugs, were those compounds for which average consumers were deemed able to reliably diagnose their own conditions and use the remedies safely. This does not mean that OTC drugs are completely safe - witness deaths from aspirin overdoses - but that they are deemed safe and simple enough for self-medication without the help of a learned intermediary.

Along with the benefits of prescription drug restrictions, however, came certain drawbacks. Consumers sacrificed some of their autonomy, lost ready access to information that might have benefited them, and were subject to sometimes unpleasantly paternalistic treatment at the hands of the medical profession. The prescription status of drugs also indirectly contributed to the rising cost of pharmaceutical products, which until relatively recently were completely uncoupled from normal free market forces: as Senator Estes Kefauver remarked during 1959 hearings on administered prices in the pharmaceutical industry, "the drug industry is unusual in that he who buys does not order and he who orders does not buy."

Drugs are unusual among consumer goods in that the FDA can demand proof of safety and efficacy before they are marketed rather than merely recalling unsafe products after the fact. The catch is that no substance is ever completely safe and perfectly effective - if water and air were prescription drugs, their package inserts would list cautions, contraindications, and side effects. And society's notion of acceptable risk-benefit ratios is driven along by our own successes: in the 1950s and 1960s doctors were aghast that the broad-spectrum antibiotic chloramphenicol killed one in 10,000 - 20,000 people who took it, though a mere decade earlier they would happily have accepted a hundredfold higher death rate for a drug that so effectively treated bacterial meningitis, typhoid, Rocky Mountain spotted fever, and other routinely fatal infectious diseases. Nowadays safer alternatives exist for almost all of chloramphenicol's indications.

Drug clinical trials, introduced by the FDA in the early 1960s and now duplicated throughout much of the developed world, cannot be a complete guarantee of safety. Egregious risks surface during preclinical animal studies, though people are people and not dogs, rats, or monkeys, so that certain biochemical adverse effects, and subtler psychological reactions, will emerge only once drugs go into humans. Even then, as a rule of thumb a study with X number of subjects will reliably discern an adverse event that occurs in 1/3X patients, therefore surprises must be expected. The largest clinical studies enroll fewer than 10,000 subjects, giving confidence of detecting events that occur in more than one out of 30,000 people. Added confidence is provided by the fact that companies and the FDA do not enter clinical trials blindly, but already know many types of drug side effects or drug-drug interactions that are likely to accompany certain types of compounds. Yet there are still occasional surprises when a drug that has been tested in a carefully selected test patient population and administered by highly trained physicians and staff under controlled conditions with careful monitoring is then released for use by hundreds of thousands or millions of patients with far less supervision of doctor and staff training, concomitant disease conditions, other drugs or dietary supplements being used, and so forth. Moreover, because clinical trials are experiments, and our society is ethically uneasy about human experimentation, we have traditionally shied away from testing new drugs on particularly vulnerable populations - children, the elderly, and pregnant women - who nonetheless appear among the patients who eventually take it. In 1992 the fluoroquinolone antibiotic Omniflox, which looked good during clinical trials using an unusually large and diverse test population, was pulled from the market three months after approval after reports of fifty severe adverse reactions, including three deaths.

Twenty years ago the FDA was accused of "drug lag" - enmiring itself in obsessive-compulsive bureaucratic nonsense that kept beneficial drugs off the market long after they were approved in Europe. Having made tremendous progress, both on its own initiative and under pressure from FDAMA, the 1997 Modernization Act, the FDA now usually beats the European Union approval process and, ironically, has recently been accused of precipitous approval of risky drugs that later need to be withdrawn (while in July 2001 the EU announced efforts to slash its approval time in half). What is the right amount of time? In some sense, it's irrelevant. Adding six months or a year of deliberation, or doubling the number of clinical subjects, might marginally increase our level of comfort, but at suffocating cost to pharmaceutical manufacturers while causing inordinate delays in bringing therapies to patients. Cutting the time and study population might stimulate drug development and give patients rapid access to drugs, but with greater risks involved. Perhaps the really important consideration is not the arbitrary moment that a drug is "approved," but a more careful and dynamic monitoring of what happens after approval, and guaranteeing effective mechanisms for responding to real world events quickly, whether this means changing the label or the formulation, or removing a drug from the market.

The most obvious reason for a drug to be withdrawn is the emergence of unforeseen risks. But in the past few years, a number of approved drugs have been pulled from the market because, according to CDER director Dr. Janet Woodcock, though they were safe and effective when used as directed, the physicians of America, despite repeated warnings, failed to use them that way. Questions were not asked, necessary tests were not made. At the same time, the FDA was not always quick enough, and lacked the sweeping authority to catch abuses promptly and take forceful steps to correct them.

Duract (bromfenac) was a Wyeth-Ayerst non-steroidal anti-inflammatory drug approved in July 1997 for the management of acute pain for ten days or less, a time limit imposed because of liver enzyme elevations noted in some patients who took the drug for longer periods during the clinical trials. Appropriate information was included in the product labeling.

Shortly after Duract went on the market, the FDA and Wyeth-Ayerst received several reports of severe hepatitis and acute liver failure - some necessitating liver transplantation - among patients who had taken the drug for more than ten days. A strong warning was consequently added to Duract labeling, and hundreds of thousands of Dear Doctor letters were mailed, both specifically emphasizing the importance of not using the drug for longer than the approved time.

Doctors, however, continued to prescribe Duract for the extended treatment of chronic conditions like osteoarthritis and rheumatoid arthritis. Eight patients had liver transplants and four people died, all but one of them after inappropriately long use of the drug. Duract offered no irreplaceable advantage over other drugs in its class, and the FDA and the company decided that enforcing the time limit would be impractical or impossible. Duract was withdrawn in June 1998, after less than a year on the market.

Rezulin (troglitazone) was a Parke-Davis drug for the treatment of type 2, or adult onset, diabetes introduced to the market in March 1997 after a six-month fast-track approval. It was prescribed for about 650,000 patients by late October 1997, when the FDA and the manufacturer issued warnings of rare, idiosyncratic, severe liver damage caused even by appropriate use of the drug - a total of thirty-five cases, ranging from mild elevations in liver test results to one liver transplant and one death. Though potentially life-threatening, the liver damage was usually reversible, thus warning letters and new labeling recommended regular monitoring during the first one to two months, regularly but less frequently thereafter, and liver function tests at the first hint of liver malfunction.

More cases of liver damage surfaced, mostly among improperly monitored patients. In late July 1998, Parke-Davis, at FDA's recommendation, issued more stringent cautions. Patients with even moderately elevated liver enzyme levels should not take the drug at all. Levels should be tested monthly for eight months, every two months for a year, and periodically thereafter. If liver enzyme levels rose appreciably outside of the normal range during treatment, testing should be done every week, and treatment discontinued if they exceeded a certain limit.

In March 1999, the FDA again changed the labeling, further tightening the testing schedule, limiting the types of patients for whom the drug should be prescribed, and requiring FDA-approved patient information sheets to alert users of the risks and of warning signs when something might be wrong. Nothing helped. On March 21, 2000, after reports of at least 63 deaths, seven liver transplants, and 20 patients who had recovered or were recovering from liver failure, the FDA asked Park-Davis to withdraw Rezulin from the market. Two other recently introduced diabetes drugs seemed to offer the same benefits as Rezulin with fewer risks. In light of the new therapeutic options, according to Dr. Woodcock, "continued use of Rezulin now poses an unacceptable risk to patients."

The withdrawals of Duract and Rezulin, among others, might be seen as FDA triumphs: drugs that were once worth the risk are withdrawn when something safer comes along, and those that pose previously unforeseen dangers are eliminated when the risk-benefit balance shifts. Many critics, on the other hand, said that the FDA's actions on Rezulin were far too tentative and slow, the result, perhaps, of the post-FDAMA kinder, friendlier attitude towards industry. And anyone should find disturbing the fact that the agency was powerless to convince doctors to use the drugs wisely - whether this would have helped enough or not. But going beyond individual decisions, viewed from a broader perspective, the philosophical underpinnings of the drug approval and drug regulatory processes seem woefully incomplete and mired in a Sixties perspective that subsequent decades of experience should have helped us outgrow.

What consumers want is safe and effective medicines. The safety and efficacy of a drug depends on two things - the pharmacological profile of the chemical compound itself, which is competently addressed by the new drug approval process, and the way in which those compounds are used, a critically important factor that, remarkably, lies wholly outside of the FDA's jurisdiction. This applies not only to new drugs, but to all drugs on the market.

At the time of a new drug's approval, the FDA makes a binary decision - yes, or no - that, barring extreme circumstances tends not to change, unlike the fluid evolution of scientific and medical knowledge. Gross safety issues are determined with a fair degree of accuracy. "We discover obvious reactions within months," says Robert Temple, MD, Director of the Office of Medical Policy at CDER, "Rhabdomyolysis, aplastic anemia, acute liver failure are very rare in the absence of drug reactions. So we're very good at finding those." With increasing knowledge and experience it even becomes easier to anticipate which classes of compounds might be suspected of causing liver damage or cardiac arrhythmias and to pay closer attention to these functions during clinical trials. But other significant risks are more subtle. "All the adverse effects I described are obvious without a control group," Dr. Temple continues. "But what if a drug increases the rate of heart attacks a little bit? The only possible way to detect that is through a controlled trial. Suppose it increases the rate by forty or fifty percent? The recent study of Vioxx is interesting and thought-provoking. [A study at the Cleveland Clinic found that COX-2 inhibitors such as Vioxx and Celebrex, used to treat arthritis, doubled the rate of heart attacks compared to that in patients taking aspirin-like drugs.] It's the first really large, long study of anti-inflammatory drugs there's been, and it picked up a non-trivial difference between the two treatments. So the question you have to ask is: What else might there be out there? And the answer is: We don't know. The other part of the answer is that it's really hard to find out."

"Most of the publicity and concern, and FDA's and the manufacturers' attention are, by law, focused on newly marketed drugs," concurs Brian Strom, MD of the University of Pennsylvania, an expert on pharmacoepidemiology and drug safety issues. "They have seen a number of recently marketed drugs lost because good drugs are being used badly. The public health problem is not rare reactions, but the common reactions to drugs that are already marketed. And that's nobody's responsibility."

The intrinsic safety of the drugs themselves when properly used is one issue. How doctors use them in reality is quite another. FDA's Janet Woodcock agrees that there are serious deficiencies in physicians' knowledge of drugs. "Doctors being trained today are not being trained to use medicines. We have a theory that the learned intermediary is supposed to know everything; that's why everybody is supposed to be hands off, and they're controlled only by medical liability or the medical boards in their state, and we know how effective that is. In the last twenty years, drugs have become the mainstay of medical treatment, but even the hypothetical godlike physician will not know how to treat people all the time."

The question of whether the FDA should have authority over the practice of medicine has been timidly raised and firmly trampled into the dust many times over past decades, largely from fear of the powerful doctors' lobby. The pharmaceutical industry, moreover, is doubtless delighted to have physicians extend the use of approved drugs at their own risk. Aside from rare cases, such as Accutane, an acne medicine that could cause severe birth defects, or thalidomide, the most infamous drug of all, recently reintroduced to treat side effects of leprosy, the agency has neither sought nor attained enforceable guidelines on how a drug should be used. The FDA has, in fact, repeatedly and explicitly stated that it does not want to regulate medical practice, though it is constantly frustrated by the misuse and overuse of the products that it does regulate. A recently completed, ten year national survey, for example, documented the grossly promiscuous, medically inappropriate use of the most expensive antibiotics to treat sore throats, which are usually caused by viruses that do not respond to antibiotics at all. Though every physician should know better, armies of doctors were prescribing lousy treatments that raised costs, incurred risks of side effects in return for no benefits, and collectively contributed to the gradual increase in the prevalence of antibiotic resistant organisms. It was, in short, a pathetic display of collective medical incompetence.

The FDA's strongest weapon in the control of appropriate drug use is the labeling, an increasingly amorphous domain that originally included only what was written on the package insert, but that now covers almost all representations in any medium made by a manufacturer to its customers. A company files a New Drug Application for a specific formulation and application: marketing an antibiotic to treat anthrax, as we have recently seen, requires different clinical trials and regulatory review than selling it for the treatment of chronic bronchitis or urinary tract infections. Clearly minoxidil needs different clinical trials for use as a blood pressure lowering agent and as a baldness remedy. Even some cancer drugs may be approved for the treatment of one solid tumor but not another. This makes sense, because one disease is unlike another, and the same drug may be used for different treatments at different doses and durations, in patients of varying age and health status who are taking a vast range of concomitant medicines. But because clinical trials for drug approvals are time-consuming and costly, drug companies typically select only one or a few specific indications and rely on doctors to use it for others.

There are perfectly legitimate reasons why many widely used drugs are commonly prescribed for unapproved indications based on information not included on the label. Science and medicine continue to evolve after the label is printed. Ongoing research and clinical experience yield important information that physicians communicate among themselves through professional meetings and peer-reviewed journals. Far from discouraging "off-label" uses of drugs, the FDA recognizes their necessity and, under new provisions of FDAMA, even permits manufacturers to disseminate valid scientific literature on such uses, subject to review and to certain conditions.

This patchwork situation, unfortunately, is riddled with flaws. Because drug labeling was originally conceived as a regulatory rather than an educational tool, even at its best it does a poor job of informing. "The label is very much a legal document," says Dr. Woodcock. "Anything in the label can be used as a basis for advertising, so we have to be extremely vigilant. A little line in the label becomes a very prominent ad. But that attitude is in direct conflict with our desire to inform clinicians through the label of everything that's known about the drug. The label is a complete disaster in some ways. We've talked about how physicians don't have the time or knowledge - you look at the label, and it starts with the chemical structure of the drug and goes downhill from there."

The FDA is currently planning an overhaul of drug labeling that will make it more like food labels, which provide user-friendly nutritional information as well as a basic list of ingredients. Yet however much the label improves, the disparity between compulsively controlled labeling of approved uses and the lack of control over off-label applications and real world use in general is troubling. In a world where databases and on-line reference works can be updated at any time, it seems peculiar to have the official level of scientific and medical knowledge on a drug frozen in time at the submission of the New Drug Application. If the conclusion is true - and it seems inescapable - that most physicians lack the time or motivation or resources to keep up to date on the medicines they use in their practices, one is reluctant to trust their judgment about off-label uses of drugs, where they need to work much harder to find useful, reliable information. Years ago, people hoped that managed care would help improve the practice of medicine rather than simply bludgeoning down its costs. Many hospitals exercise limited control over drug use through formularies, guidelines, and prescription reviews, and recently federal agencies such as the Agency for Healthcare Research and Policy (AHRQ) have worked on developing guidelines to improve the quality of care and patient safety, but it seems preposterous that the agency with by far the most information and expertise on drugs is allowed to carry its participation only so far and then inexplicably vanish.

"It's clear to me that for the FDA to continue to have a positive impact we can't just be a gatekeeper for drugs getting onto the market," remarked Dr. Woodcock. "We have to be players in drug information - informing clinicians and consumers about drugs."

If premarket approval of new drugs is supposed to afford greater protection of the public than post-disaster recalls, this advantage is partly negated by the utter lack of control over what is done with them, as cases like Duract and Rezulin, among others, so clearly show. Why shouldn't the FDA have regulatory authority over the use of drugs? Or if not the FDA, shouldn't someone?

Herbal Tea for Socrates

The FDA's origins as an agency that fought quackery and fraud left it with habits that tend to die hard. For decades the agency's position on vitamins and mineral supplements was influenced more by suspicion of marginal and alternative therapies, which it lumped in with patent remedies, than by a scientific assessment of available evidence. Even during the Sixties and Seventies, when alternative treatments enjoyed burgeoning popularity and segments of the medical profession began to view them with open-minded interest, the FDA's arch-conservative attitude was that the amorphous class of substances known as "dietary supplements" were useful only to treat nutritional deficiencies, but that nutritional deficiencies in America could be avoided by eating a proper diet, a closed logical system that proved them to have no value at all.

The FDA tried on many occasions to limit the claims that were made about dietary supplements, but frequently lost in court, and its position was gradually eroded by the changing tenor of the times. By the 1980s the agency even began to allow health claims on cereals and other foods, as long as they were backed by scientific evidence and did not use specific types of statements that would hoist them into the realm of drugs. But even as the FDA attitude loosened, it grew clear that regulation was needed in an area that was populated by crooks and incompetents along with honest researchers and manufacturers, much like any other industry. Some health food store products contained ingredients as potent as those in OTC or prescription drugs, yet were subject to none of the clinical trials, quality checks, or other regulations. The fact that they were natural did not make them mere foods - "Drink up Socrates, it's natural!" reads the caption on a cartoon that appeared on FDA bulletin boards several years back - and there was good evidence that some products were hardly as natural as they seemed, being laced with actual prescription drugs or inadvertently tainted with heavy metals or other toxins.

Matters came to a head in late 1989, when the FDA learned of several dozen cases of a debilitating new disease of the muscles and blood, eosinophilia myalgia syndrome (EMS), that seemed exclusively associated with the consumption of L-tryptophan, a dietary supplement recommended as a sleep aid. Within a month, the FDA recalled all L-tryptophan, but eventually over 1,500 cases of EMS were reported, including thirty-eight deaths. Subsequent investigations suggested that the culprit was a contaminant in the L-tryptophan produced by one Japanese manufacturer rather than the main ingredient itself, but this was little consolation to the FDA, which had seen people sicken and die from substances it was not adequately empowered to control.

The FDA promptly sought greater authority over dietary supplements, including the right to preapprove health claims and to handle at least some of them with the sort of restrictive supervision applied to prescription or over-the-counter drugs. To bolster its arguments, the agency hastily assembled a report of "Unsubstantiated Claims and Documented Health Hazards in the Dietary Supplement Marketplace." Unfortunately, the report was so riddled with errors and unsubstantiated claims of its own that it proved a public embarrassment to the FDA. At the same time, a crusade led by Senator Orrin Hatch of Utah, home to many dietary supplement companies, and a grassroots movement - urged on by the industry and health food stores, and publicized in alarmist television commercials showing FDA cops kicking down doors to confiscate people's vitamins - roused public sentiment against the FDA's paternalistic attitudes and brought in more letters to Congress than Watergate. In the end, the FDA efforts backfired, resulting in the passage of the Dietary Supplement Health and Education Act (DHSEA) of 1994, which placed vitamins, minerals, amino acids, herbs and botanicals, other dietary substances, and concentrates, metabolites, constituents, extracts, or combinations of the above firmly on a par with foods rather than prescription or OTC drugs, without even the premarketing safety requirements of food additives or colorings.

"There are two things that the people and Congress want," I was told by a senior FDA official in the Office of Compliance, "Access to anything that can possibly help, as soon as possible. And they want assurance that anything out there is safe."

The public, alas, got one but not the other. Some dietary supplement proponents claim that the FDA has all the power it needs to protect the American public. This is true in a technical sense, because the agency can remove from the market products proven to be dangerous. But a huge difference in effort, cost, and time exists between keeping an unsafe product off the market by denying premarket approval, and pursuing a host of individual cases in court after already marketed products reveal their danger, especially when the companies are small operations that can readily disband, start again under a new name, and carry on as before. Moreover, proving danger is unusually difficult with dietary supplements. The Dietary Supplement Health and Education Act does not require reporting of adverse events by physicians to manufacturers or by manufacturers to the FDA, as is the case with drugs. Many manufacturers have admitted that it never occurs to them to tell the FDA about problems, which are often collected through their marketing departments anyway. Consumers may not consider "natural" products to be entities similar to drugs, and thus don't always link adverse effects to them, don't report them to their doctors, and remain blissfully, sometimes dangerously unaware of potential harmful interactions between certain dietary supplements and drugs. By one estimate, only about one percent of adverse reactions to dietary supplements are reported, which according to a health department official testifying at an August 2000 public meeting on dietary supplements containing ephedra, suggests that 300 - 3,000 people each month suffer serious, sometimes disabling or fatal side effects to ephedra compounds alone.

Ephedra is a disturbing example of the challenge the FDA faces in attempting to bring dietary supplements under control. Ephedra comes from the ephedra seneca plant, primarily found in Asia, and is marketed as an herbal product under many names, the most common being "ma huang." The Chinese have used ephedra for thousands of years to treat colds, asthma, and congestion. The most prevalent of the five main active alkaloids in the plant, ephedrine, a chemical cousin of methedrine and other amphetamines, is used in FDA regulated OTC decongestants and bronchodilators. Thus, ironically, the same active ingredient is regulated in a form where it is manufactured under drug industry standards, given in a standardized form, with appropriate instructions for use, after being evaluated for safety and efficacy, but is not regulated when sold in a form that is not standardized, produced under unknown conditions, for a large and ambiguous set of indications, often without adequate instructions or cautions. The legislative justification for this is clear, though the real world logic is unfathomable.

Industry, nutritionals retailers, and devoted consumers aver that ephedra has been used safely for thousands of years. The FDA, many health and consumer organizations, and those who have suffered its ill effects point out that this traditional use was for relatively low doses taken for short periods of time to treat specific complaints. The major current uses in the U.S., on the other hand, are for weight loss and to improve athletic performance (ephedra was recently banned by the NCAA and NFL). Ephedra products are used in combination with other active compounds for long periods of time by dieters and athletes eager for quick results who often start at high doses and push them still higher. Between January 1993 and February 2001, the FDA collected 1,398 reports of serious reactions to ephedra compounds, including 81 deaths, 70 seizures, 69 strokes, and 95 heart attacks or cardiac arrhythmias, many of them in young and otherwise healthy people.

The problem, of course, is a lack of reliable information. The FDA data is compiled from incomplete voluntary reports. Critics distrust it because of the FDA's supposed bias and the erratic nature of the reports. As the president of the American Herbal Products Association pointed out at the 2000 public meeting, in the past year, fourteen surveyed companies had sold over three billion servings (classified as a food, dietary supplements come in servings rather than doses) and received only twenty-five adverse event reports, or a mere eight per billion. In citing this impressive figure he ignored the same profound flaws in adverse event collection methods he had criticized in the hands of the FDA.

The FDA's policy is that no amount of risk is tolerable in the absence of efficacy, which is why drug evaluations look at both factors together and make complex decisions: more risk is acceptable with a drug that treats cancer than in a remedy for headaches or chapped lips. Yet little sound information is available for most dietary supplements. Because most dietary supplements are not patented, the industry lacks the economic incentives and deep pockets that allow pharmaceutical companies to fund large, controlled clinical trials. Some foreign countries are more accepting of herbal and other "alternative" treatments as part of mainstream medicine, and generate scientific studies of them. The FDA reviews this literature, but there are limits to its value. "Diluted as a tea and consumed by an elderly person in an Asian country is different from homogenized, blended with something else, put into a tablet, and consumed by an 18 year old athlete," says Christine Lewis-Taylor, director of the Office of Nutritional Products, Labeling and Dietary Supplements. "So these foreign documents don't necessarily have the evidence that we think is needed." And though DSHEA also created an Office of Dietary Supplements at NIH, it is woefully underfunded and scarcely able to sponsor the type of scientific research needed to evaluate the safety and efficacy of the thousands of compounds and combinations on the market, including an estimated 300-400 different ephedra products sold for weight loss and athletic performance alone. Some scientific evidence from studies at Columbia and Harvard Universities does suggest that ephedra may be helpful in weight loss, though it is hard to translate preliminary results into regulatory policy or guidance for consumers as the FDA is supposed to do in other areas. At the same public hearings where the FDA and some witnesses enumerated deaths and disablement from ephedra use, a stream of industry representatives and customers testified - some of them trim, seemingly happy people telling the audience how they had once been chronically, grossly obese, depressed to the point of suicide, displaying enormous pairs of pants as proof of how massive they had been despite everything they tried, until they used ephedra.

Chemical entities are relatively easy to test for purity, quantity, and identity. This is not true of herbal remedies. At an American Association of Pharmaceutical Scientist meeting on "neutraceuticals" in 2000, a speaker at a session on good manufacturing practices began, "First, you have to make sure that you have the right plant." Toxic or useless plants have been mistakenly or purposely substituted for botanicals, and with many plants not only the plant, but the species, the part of the plant, the way it was grown, the way it has been stored, and any number of other factors affect its final potency. For many botanicals, no one knows which of the many constituents might be the active ingredient. Tests of chondroitin sulfate, a substance derived from animal cartilage used to treat osteoarthritis, have found preparations with anywhere from 0 - 200% of the ostensible content. One can't, quite frankly, have it both ways: either something is potent, in which case the dose matters, or it has no effect, in which case it's quackery.

Certainly the public deserves to be protected against fraud. And to the extent that we find it useful to keep complex, potent drugs under the controlled supervision of medical professionals, and to insist upon demonstration of safety and efficacy and the provision of adequate instructions for use when it comes to over-the-counter medicines, what earthly reason could there be for active compounds sold as dietary supplements?

The issue, of course, returns to the question of how paternalistic the American public wants its government agencies to be. "DSHEA is based on wanting consumers to make their own decisions," says Lewis-Taylor. "But how competent can consumers be? Can a mother make a decision about giving gingko to a five-year-old? Can the agency help her make that decision? What level of protection do we want? And how much do you trust the agency? It's been a long and interesting battle between how much the agency can protect the consumer and how much we should provide information on the label and let consumers decide for themselves." Unfortunately, at present, the FDA is able to do neither.

Mad Cows and Englishmen

Drug approvals and dietary supplements are conspicuous areas where the FDA faces new issues, such as communicating complex information in an accurate and useful way to medical professionals and end users. In addition, the agency retains vast and growing responsibilities as an old-time inspectional organization that protects the public by establishing and enforcing strict safety standards, especially in the area of food. But regulating a complex international market is an unwieldy chore, fraught with many unknowns. And with expanded clinical knowledge and improved diagnostic and detection techniques, we are able to discern and, one hopes, protect against a growing set of new dangers. But greater safety comes at a cost far above what was needed merely to limit the number of insect parts and rodent droppings in our food, and the new challenges threaten to overwhelm the agency's available resources.

Two diseases of cattle have recently emerged or reemerged, both initially in Britain, then spreading to other countries. One affects only the cattle themselves and poses an economic problem of vast proportions. The other affects both cattle and humans, and creates daunting quandaries for international regulatory agencies.

Foot and mouth disease, a highly infectious viral disease spread through physical contact or by airborne or mechanical dissemination of virus, affects cattle, sheep, pigs, goats, and some wild animals. Though rarely fatal, the disease can leave animals permanently debilitated, resulting in a severe loss of meat and milk production. Because of the potentially massive economic impact, though the outbreak was limited to 2,000 confirmed cases, British authorities took draconian measures, slaughtering nearly four million cattle and attempting to contain the spread of the disease by closing national sites and imposing restrictions on travel. These measures eventually halted the spread within the United Kingdom, and strict but straightforward American precautions have thus far fully prevented any contamination of U.S. herds.

Bovine spongiform encephalopathy (BSE), or "mad cow disease," on the other hand, presents a more frighteningly nuanced set of problems, where soft scientific knowledge must be balanced against a real but largely unknown health threat and readily perceptible economic consequences. Animals and humans are infected by diseases in the spongiform encephalopathy family for years before symptoms first appear. No currently available test can diagnose BSE or its human equivalent in its early stages. Preventive measures are therefore based on hypotheses that will only be proved well in the future, when it is too late to correct mistakes. Worse, the infectious agent in BSE, defective proteins called prions, are not inactivated by cooking or even by the heat or chemical treatments commonly used to sterilize medical instruments, raising the dreadful possibility that disease can be spread from one undiagnosed, asymptomatic sufferer to other patients through equipment that ought - by all ordinary standards - to be safe.

The first case of BSE was reported in 1986. Afflicted cattle exhibit changes in behavior and coordination, weight loss, and decreased milk production. Symptoms typically first appear from two to eight years after infection, but once the affliction becomes evident animals' conditions rapidly worsen and they die or are destroyed, usually within two weeks to six months. No preventive vaccine or curative treatment exists. To date, BSE has affected about 180,000 cattle in Britain, 1,800 in other European Union nations, and was recently detected in Japan, leading to fears that it will spread to Indonesia, Thailand, and other parts of Asia. BSE is far less contagious than foot and mouth disease, being communicated only through eating body parts or tainted by-products from infected animals, so though the number of affected animals far exceeds those with foot and mouth disease, the same vast preemptive slaughter did not take place. Once the most probable vector was identified - infected meat and bone meal fed to cattle - and appropriate preventive measures were taken, the number of reported new cases began to level off and drop.

Cows in the modern world are not just meat or dairy cattle. A living creature that enters a slaughterhouse is disassembled into marketable parts, like a car going into a chop shop. Meat, bones, hides, blood, eyes, hooves, internal organs, all have uses somewhere, whether as human or animal food, medical research material, or as chemical ingredients for other products. The biggest outlets for non-edible cow parts are rendering plants where fat is processed for use in soap, cosmetics, glue, and other items, while other remains are turned into animal feed.

In March 1996, based on ten cases, British doctors identified a new human disease, called new variant Creutzfeld-Jakob Disease (nvCJD), which was similar to but distinct from the rare progressive, untreatable, invariably fatal Creutzfeld-Jakob Disease. nvCJD had a younger onset, longer duration, and slightly different pattern of symptom emergence, but led to the same eventual result: depression, loss of coordination, dementia, and death. By 1997 scientists concluded that the probable cause of nvCJD was BSE. As of late 2001 there were 112 confirmed or probable cases of nvCJD - 107 in the UK, 3 in France, and 1 each in Ireland and Hong Kong - though because diagnosis is impossible before symptoms appear, and because the incubation period may last from five to twenty years or more, a much larger, invisible epidemic may already exist, waiting to reveal itself.

The FDA's challenge was to prevent mad cow disease and thus nvCJD from entering the United States. The agency had the advantage, points out Dr. Murray Lumpkin, the FDA's Acting Deputy Commissioner, who was given the task of coordinating agency efforts to contain BSE, of having the British experience as an advance indication of what might happen. What it suggested was quick preemptive action. "You can't wait until the first cow shows up ill, because of the incredibly long incubation period. The cow is already out of the barn. So even without evidence of disease in this country, and without evidence that the feed was infected, we had to take prudent preventive measures. That did have an economic impact on the rendering and feed industries, but we got very strong support from industry on this because they were aware of the far greater impact if BSE amplified through our herds."

The FDA has always, and does increasingly, coordinate its efforts with other agencies and organizations, but the BSE threat entailed a truly remarkable degree of collaborative effort between three divisions of the Department of Agriculture, the State Department, Environmental Protection Agency, Department of Defense, Office of Management and Budget, National Institutes of Health, Centers for Disease Control and Prevention, the Customs Department, the White House, and various state and industry organizations.

Even within FDA, keeping the BSE non-problem from turning into one entailed unprecedented cooperation between all five Centers - Food, Drugs, Biologics, Medical Devices, and Veterinary Medicine. In September 1992, the FDA issued an import alert about animal by-products from BSE-afflicted countries. In late 1994 this was superseded by an alert calling for the automatic detention of high risk tissues from the steadily growing list of suspect nations. In August 1997 a rule went into effect banning the use of most mammalian protein in animal feeds for ruminant animals - a categorical preventive measure against the eventuality that BSE might have entered the country undetected. In August 1999 the agency issued guidelines to blood centers, cautioning against the use of blood from donors who might have been exposed to BSE tainted food, which included those who had spent six or more consecutive months in the U.K. between January 1, 1980 and December 31, 1996 - a rule subsequently expanded to include other countries.

To ensure that industry was in compliance with the feed regulations, the FDA and cooperating state agencies carried out nearly 10,000 inspections from 1998 to 2001 of feed mills, farms, rendering plants, haulers, and distributors. The great majority of sites were in compliance, and most of those in violation at the first inspection willingly corrected their mistakes soon afterwards. Because any violation of the rules could trigger a disaster of monumental proportions, and because the FDA could not possibly police every farm, plant, and bag of transfused blood, the agency could not depend on inspections as its primary defense. Knowing that it was in the industry's best interests to comply, the FDA expended a tremendous amount of time and resources on educational programs, sponsoring workshops for veterinarians and feed control officials from all fifty states, Puerto Rico, the Virgin Islands, and Canada, and producing a satellite teleconference and CD-ROM.

"At the end of the day, everybody in a very multi-faceted chain has to do their part," says Dr. Lumpkin. "If anybody screws it up, or says 'This doesn't apply to me,' the whole house of cards comes down."

One challenge was simply figuring out the full range of products potentially affected by BSE, which included not only the obvious meat, animal feed, and pet foods, but transfused blood (possibly from people with nvCJD), bone marrow, cornea, and other tissue transplants (ditto), vaccines (often grown in fetal calf serum), drugs and dietary supplements (animal-derived gelatin in capsules), candy (gelatin again), soaps and cosmetics (tallow), and medical instruments (cross-contamination between patients). In addition, the FDA needed to step up surveillance for smuggled shipments of all classes of suspect goods, and improve methods for tracking the multiple countries of origin of finished goods that came from one place but whose various ingredients might originate elsewhere.

Throughout it all runs a difficult, often unsettling weighing of the intangible risks and uncertain benefits accompanying every decision. Many of the vaccines we use to protect people from serious diseases, for example, are years old, and some come from European sources that could possibly have been contaminated. Though no vaccine is known to have spread BSE from animals to people, there is a small but real risk that this could happen. "You could just throw away all of these vaccines and start fresh with uncontaminated cow material," says Dr. Lumpkin. "That's one very conservative, prudent approach. The problem is that we would be a community without vaccines for a long time, which would pose a very real public health problem." In the end, after long debate and public discussions at an advisory committee meeting, the FDA left existing vaccines on the market, but recommended close review of the source of all bovine materials used in any future biologics.

All of this massive effort and expense was to contain a threat that had never even materialized on American shores, a novel animal disease that has the potential, thanks to modern industry and a global marketplace, to wreak extraordinary havoc in astonishingly diverse ways. And in handling the unwieldy BSE challenge, the FDA enjoyed the unusual advantage of having American industry on its side. Companies stood to lose heavily if the FDA did not win. If BSE slipped into the country and caused even one confirmed case of nvCJD, countless cattle would be slaughtered and the beef industry and many related concerns would suffer tremendous losses. Industry support, however, is far from the norm, and had the multitude of potentially affected companies chosen instead to pit their vast economic and lobbying power against the FDA, the process would have been very different.

We are the World

Even in areas where old style gumshoe type inspections are still necessary, the proliferation of facilities and complexity of manufacturing processes has made it necessary to adopt new strategies to ensure product safety. The practical necessity of budget and manpower constraints and pressure from industry and the government have forced the FDA to take a more collaborative approach with those it regulates, while experience shows that education is often a more effective means of encouraging industry to comply. Instead of focusing on the impossibly vast deluge of products, the agency increasingly looks at the control systems, processes, and executive oversight that goes into making them. One of the earliest systems analysis type of inspectional techniques applied by the FDA was inherited, curiously enough, from the space program.

In October 1958, Congress created the National Aeronautics and Space Administration to launch the American space program. Less than a year into its existence, NASA was immersed in pondering the challenges of manned space flight, including that of providing safe, nutritious, palatable food for astronauts, even though the first American manned mission, a suborbital flight by Alan B. Shepard, Jr., still lay two years in the future and would last only 14.8 minutes, scarcely time to open a package and chew.

One of NASA's concerns about space food was that in zero gravity, food crumbs would not simply drop on an astronaut's lap or onto the floor, but that small particles would contaminate the spacecraft's atmosphere, and might float into and foul delicate instruments. A more serious problem was food safety. Though centuries of explorers had crossed oceans and deserts and frozen wastes for months or years with more vulnerable stocks of food, the notion of leaving Earth altogether, far from medical care, seemed to demand much higher food safety standards, and NASA sought a way to guarantee that space food was completely free from any toxins, bacteria, or viruses. Food safety monitoring up to that time consisted of randomly selecting and testing a certain number of samples from a day's production run. If the samples were all right, one wagered that everything else was, too. Taking a larger number of samples yielded incrementally greater statistical evidence of safety, but without testing every single item in the entire production run and reducing output to zero, the degree of certainty would never attain one hundred percent. Problems that weren't picked up by this sampling, or that developed later in the product's life during shipping or on the shelf, were dealt with when they arose. And no matter how careful a manufacturer was, problems always did arise sooner or later. This was not good enough for NASA, which had developed a "zero defects" policy of quality control.

Paul Lachance, a scientist in charge of flight food and nutrition at NASA, wanted to apply the same sort of systems approach to food that the agency's engineering contractors were using with hardware. NASA approached Pillsbury, which assigned the task to Howard Bauman, a young microbiologist who had previous experience with the challenges of preparing food for submarines.

The crumb problem was easily overcome by producing bite-size foods coated with an edible veneer to hold the product together. Making one hundred percent safe food was a different matter, but eventually Bauman and his colleagues developed the Hazard Analysis and Critical Control Point (HACCP) systems approach. HACCP was adopted by NASA, and also became the standard at Pillsbury. In 1971 the company made a presentation on the program at the National Conference of Food Protection, and soon afterwards won a contract with the FDA to train agency personnel in the method. The FDA applied the principles in high risk food product areas such as acidified and low-acid canned foods, which are susceptible to botulism contamination if not properly processed.

HACCP consists of seven basic principles, which provide a conceptual template that can be applied to any food preparation process. It entails identifying hazards and the critical points between production and consumption where those hazards may occur, setting limits on important parameters (e.g. cooking or storage temperatures), developing techniques for monitoring actual performance, directions for handling errors, and methods for maintaining records and verifying that the HACCP program works and for revising it if it does not. A unique HACCP plan is devised for every process.

In December 1997 a HACCP program went into effect for seafood, another food area that posed significant risks. The USDA implemented HACCP programs for meat and poultry processing plants as of January 1999. Beginning in 2002, large producers of fruit and vegetable juices, in response to chronic problems with contaminated apple cider, must operate under HACCP guidelines, and the FDA is considering extending HACCP to other and possibly all foods that fall under its jurisdiction.

HACCP seems, at first glance, to be nothing special, a gimmick akin to the lists a conscientious parent might make and keep to help pack while preparing to go to the same vacation spot year after year. But then, of course, there's great value in such gimmicks, which is why we tend to use them. A more important feature of the HACCP approach, however, is that it helps food processors think logically and systematically about what they do and why they do it and how things might go wrong. Writing a HACCP plan becomes a self-education process to help people dissect the real hazards and solutions in what they do, as well as a blueprint that the FDA can use to check a company's performance. HACCP is an example of the new way the FDA operates, focusing on processes more than products, and giving industry tools to improve its performance rather than policing adherence to standards, while nonetheless letting them know that enforcement is available when and if needed.

The change did not sit easily with some old time investigators. "When we started the HACCP program," says Susan Setterberg, Director of the FDA's Central Region, by way of example, "Toni Ravelli [a young New Jersey investigator and now the state's Seafood HACCP Coordinator] and I discussed the way she did it. She worried, because she felt that if she could spend X more days she could do a better job, check another system, feel just a little more certain that everything was covered. I felt that she got better results if she spent three days in the plant rather than five, because either you hit every plant for three days, or you hit three-fifths of them for five days. Just being there has a major effect on industry."

Evidence supports the idea that the mere presence of a warm body may be more important than compulsively thorough inspections. Russell Munves, a New York attorney specializing in food and drug law, remarks that he can advise medical manufacturing clients on proper procedures until he's blue in the face, but they're usually too busy to care. Then an FDA inspector visits their plant, and suddenly they care very much. "You get the most enforcement for the least amount of money when the investigators can go out and inspect facilities. It's like the cop in the rear view mirror - that's what makes you obey the speed limit." In the long run, quantity over quality is not an ideal solution, but in the real world, with a financially strapped and understaffed agency, reminding companies that the FDA is watching may be the most effective strategy.

Yet the new techniques can also actually provide useful tools for managing complex issues that would soon grow unwieldy even with a much larger budget. Ravelli changed her opinion of HACCP, which she thinks gives a more comprehensive view of what is happening in a plant than the one-day snapshot she got through a traditional inspection, and helps to redefine the role that the FDA plays towards industry. "The firm has to keep records of what they're doing, what's going on, and anything that changes. Any new process, piece of equipment, is noted. The onus is on the firm to tell you what's going on. You don't just have to find it."

The FDA, nonetheless, does not have to trust a company's records. Inspectors can compare procedures to the firm's written HACCP program, either on a random basis or when there is reason for suspicion, checking that critical control points are correctly identified and that appropriate monitoring is in place, interviewing employees to make sure that they understand and follow the prescribed plan, and taking product samples in a more limited and focused way as a last check of the final result. In extreme cases, nothing stops the FDA from reverting to the old methods and going over every inch of the facility with a fine toothed comb.

Perhaps the most useful thing about HACCP is that it fosters a more collaborative approach to consumer safety between the FDA and the industry, and enables the agency not only to use its knowledge and experience to police industry but to help industry perform better. Even the most cynical old-time FDA inspectors agree that the vast majority of companies want to do the right thing and just occasionally need help in figuring out how. The agency conducts regular seminars for members of industry, and finds that education works wonders at improving compliance. When people understand the reason for rules and regulations they are far more likely to follow them than if they are perceived as mere troublesome and arbitrary fiats.

"Firms now are more likely to call you and ask questions," says Ravelli. "I get calls every day from people asking whether they should use this Critical Control Point or not. Or they'll have training on various subjects at different locations. We're making ourselves more readily available to industry than we were before. Before, if people called, they wouldn't tell us who they were, because they were afraid of being inspected."

Yet in the end, the success of programs like HACCP, with their more analytical and collaborative approach, can succeed only if the agency heeds both halves of Theodore Roosevelt's famous adage, "Speak softly and carry a big stick." A decade ago, no amount of discussion or warning letters seemed to deter the food industry from gradually lapsing into increasingly unjustified claims on food labels. But companies snapped back into line almost overnight after the FDA ordered federal marshals into a warehouse to seize 2,000 cases of Procter & Gamble's Citrus Hill Fresh Choice orange juice on grounds that reconstituted juice was not "fresh." There is much to be said for cooperative approaches, but one cannot forget who is the regulated and who is the regulator.

A world apart from seafood inspection - yet sharing some of its challenges and demanding analogous solutions - the most wildly diverse product category regulated by the FDA is medical devices, which includes everything from tongue depressors through breast implants, artificial hip joints, and cardiac pacemakers to MRI machines. Until passage of the 1976 Medical Device Amendments - urged through by concerns over the Dalkon Shield and defective heart valves - devices were erratically regulated and, unlike drugs, subject only to after-the-fact enforcement actions against products that had already, tragically, demonstrated their danger to patients. After 1976, the FDA gained approval authority, which was exercised differently according to a product's complexity and perceived risks. The Center for Devices and Radiological Health acquired some powers that the Centers for Drugs and for Biologics lack. It can demand mandatory product recalls, whereas those in the other centers are technically voluntary, and can require that a company continue clinical studies even after a product goes on the market, a power that the Center for Drugs has only in exceptional circumstances. Because of past problems with defective pacemakers, heart valves, and other implanted devices, CDRH also has the authority to require tracking of products to individual physicians and patients, which is neither available nor even practical in drugs and biologics.

There are currently about 80-100,000 types of medical devices on the U.S. market, ranging from pieces of plastic used in IV infusion sets to pacemakers, respirators, and life support systems. "You don't want to treat a pacemaker and a bedpan with the same standard," observes Dr. David Feigal, Director of the Center for Devices, which is why devices are ranked in three different risk categories, from Class 1, fairly innocuous products that don't even need to be tested as long as they are well manufactured, to complex Class 3 devices that must be intensively studied in clinical trials before they are allowed on the market.

One extraordinary challenge for the Center for Devices - or the Center for Things, as it is jokingly called within the agency - stems from the vast number and diversity of device manufacturers. Whereas the great majority of drugs are made and sold by a relatively small number of large pharmaceutical companies, there are currently about 8,000 device companies with 13,000 registered manufacturing sites. Two percent of the firms generate fifty percent of the market's annual $50-$70 billion sales, and are either as large as or actually are big pharmaceutical companies, like Johnson & Johnson. Eight percent of firms have eighty-five percent of sales. But the largest number of device companies are small operations, with a median size of about fifty employees and some very small mom and pop operations that make specialty items such as low demand implants, prepacked saline syringes, or that are local repackagers of goods produced elsewhere. To ensure good manufacturing practices and quality products, the FDA must inspect all of these sites on a regular basis.

"We are supposed to inspect companies every two years," Dr. Feigal says. "The problem is a dropping inspectional capacity. That's very recent. I think we kept up until about six or eight years ago. We did 2,600 inspections a year back then. Now the capacity has dropped to under 1000. We get to U.S. firms every four years, and foreign firms every seven years. We're starting to squawk about that now."

The Center for Devices, like the rest of the FDA, would benefit enormously from more money and a larger staff. But unique characteristics of the device world would render the challenges here more daunting than in the other Centers even if the budget were multiplied several times. A good pharmacologist understands the principles behind any therapeutic drug. Drugs fall into a limited number of therapeutic categories and modes of action, and the types of things one looks at - absorption, toxicity, drug interactions, etc. - arise again and again in only slightly different form. The equipment and processes used to make drugs and the instruments to test them are relatively standard. Even when a whole new class of treatments or testing strategies is introduced, the Center for Drugs would need only to bring in or train someone in a new set of skills.

A medical device, on the other hand, can be almost anything. A new resorbable polymer implant that serves as a scaffold for a regenerating body part poses unique issues of chemical engineering and materials biocompatibility. A novel type of ophthalmic laser involves sophisticated knowledge of laser physics and of light-tissue interactions. The rapidly growing number of instruments and devices that use microprocessors requires familiarity with electronics and software. With a truly innovative new medical device, the only experts in the world are often the inventor or the employees of the company that makes it. Whereas a food inspector could, with a little review of materials and perhaps refresher training, go into a cannery, a bakery, or a seafood processing plant, it is impossible for a device investigator to walk confidently into every manufacturing facility armed with the full knowledge needed to evaluate the product and every step of the process by which it is made. Even given an infinite budget and limitless staff, the FDA could not hire enough people with appropriate skills to cover every technology used by the medical device industry.

For many years, device investigators conducted good manufacturing practice (GMP) inspections much as was done in food or drug plants. "We focused on complaints, Medical Device Reports (adverse experience), changes to the device, things of that nature," says Rob Ruff, an investigator with the New Jersey District Office. "If we saw any trends or patterns of problems in that data, we'd try to link it to a GMP deficiency - a problem in manufacturing, or a vendor, or a component. We'd look at complaints first, then work our way back into the system."

Starting in the late 1990s, the Center for Devices shifted to something more like a systems approach - not the same as the food HACCP program, which focused on processes, but one that concentrates on design and management. In 1999 it was formulated into the Quality System Inspection Technique (QSIT) top-down approach that does not exhaustively scrutinize records for every aspect of a company, but instead looks at the way the firm handles seven subsystems and thirty-eight critical areas, then drills down as needed into specific records. The emphasis is no longer on the "gotcha!" form of police inspection, but on making sure that the company "is in a state of control." Old investigations were conducted as though the investigator was constantly thinking about compiling evidence for an injunction. The QSIT approach aims more at issuing warning letters to bring a firm into compliance and then later, maybe, if it proves necessary, reverting to more exhaustive legalistic gumshoe techniques if legal action is the last resort.

"Some people think it's FDA's job to check every process at every firm and tell them what's wrong with their process," says Tim Wells, formerly a branch chief in the Division of Enforcment at CDRH and team leader of the QSIT project. "But there's no way FDA can do that. Firms have to take responsibility for their own processes, and it's the executive management that needs to know that every process on every product in every location is meeting the requirements. We can't be everywhere or check everything. If somebody dies in Pittsburgh because of a product made down the street, is that our problem, because FDA hadn't inspected them within the last six months? There are people in the FDA who feel that way - that we are responsible for making every product safe. I tell them, 'I appreciate that you feel that way, and that you stay up at night worrying about it, but people can't blame the FDA because someone dies because of a poorly made product.'"

What the QSIT approach attempts is to shift the focus from the specific instances where things have gone wrong to the system that allowed them to happen and to the top management at the company who failed to exercise adequate oversight. Companies are not able to claim that the unvalidated process or uncalibrated instrument were fixed or that the untrained worker was educated or fired - stopgap explanations for specific errors that can go on forever without addressing the underlying system flaws. The ultimate objective is not blaming the person who physically did the violative commission or omission, but holding senior executives clearly and directly responsible for the whole system, which is ultimately a more efficient and effective way of making sure that a company knows the proper ways to insure product and consumer safety, and that it is in control of its processes and safeguards.

"Global harmonization" is a jarringly hippyesque term for a federal agency to use, particularly at a time of rampant international terrorism, but it is becoming a simple fact of life. For most of its history, the FDA was a fiercely nationalistic organization, believing that its standards and methods were superior to those in the rest of the world - an opinion that the rest of the world largely shared - showing little faith in any other country's inspectional organization. The fact that a drug or device was approved overseas, even in a major European country, had virtually no standing in the eyes of the FDA, which might accept some foreign scientific and clinical data but insisted that companies begin a new drug or device application all over again before marketing in the United States.

This old approach has been rendered untenable by the increasingly global marketplace. Multinational pharmaceutical companies acquire raw materials and finished products, or conduct clinical trials at sites around the world. Every class of products regulated by the FDA now arrives from abroad as well as being produced domestically, and while the agency increasingly uses systems approaches to monitor the performance of U.S. companies, it relies on sampling and testing food imports at their points of arrival - checking a mere one percent at best - and takes already vastly overburdened drug and device investigators away from their regular work plans to send them abroad for foreign plant inspections that are largely conducted the old-fashioned way.

The FDA'a legal mandates are clear, and old traditions die hard, but the agency lives in a real world, where economics and public wishes must, it seems, be weighed alongside of questions of safety - at least until disaster forces the pendulum the other way. One striking example of the evolving FDA was the 1997 outbreak of cyclosporiasis, a diarrheal disease, caused by imported Guatemalan raspberries contaminated with the cyclospora parasite. At the FDA's request, the Guatemalan Berries Commission halted the export of raspberries, and because of the fruit's short shelf life, existing imported stocks were quickly depleted, but the problem remained of how to manage the problem in the future. Ironically, raspberries were not indigenous to Guatemala, but had been introduced by WHO to help impoverished farmers produce a marketable, high value export crop. A permanent ban on Guatemalan raspberries would have signficant adverse effects on both the Guatemalan economy and American consumers. So instead of regulating, the FDA educated.

"The agency took a lot of flak," says Deborah Ralston, now Director of FDA's Office of Regional Operations, and one of many who feel stuck in the middle while trying to uphold the FDA's traditional high standards in an era that does not tolerate such restrictive policies. "The American public has become used to having exotic fruit and vegetables year round. Some of us had a hard time with the thought that we are here to protect the public health, but we're also supposed to make sure that the American public has access to fruit that isn't normally available. I don't even want to guess how much the FDA spent working with the Guatemalans - a small fortune - but we ultimately came up with a solution." The solution was to send analysts to Guatemala to help test for cyclospora, a difficult parasite to detect in raspberries. The FDA ran training courses, helped the government set up an inspection program, and periodically sent its own inspectors down to audit the program and verify the testing procedures.

In the matter of drugs, matters remain more complicated because of lingering fundamental differences in the U.S. and foreign approaches. The gold standard for American clinical drug trials remains the randomized, double-blinded, placebo-controlled study, yet some countries in Europe consider it unethical to use a placebo when an agent with known therapeutic benefits exists, and may even desire comparative studies of new and old drugs to help national health systems and other end users make rational decisions about the best and most affordable treatments to use. And despite years of mutual recognition agreements, the FDA cannot bring itself to put complete trust in foreign facility inspections. "At one point a few years back," remarks one senior FDA official, "virtually the entire Italian inspectorate was in jail for graft. That doesn't give you much confidence."

Yet the FDA will be obliged - by economics, by consumers, and the realities of a changing world - to become a "kinder, gentler FDA," as some agency people sarcastically phrase it, putting more trust in industry and other countries, and shifting its emphasis a bit more towards education rather than enforcement. Yet in spite of successful new educational and inspectional programs it is not hard to find considerable dissatisfaction within the FDA over the agency's small budget and shrinking staff. FDA employees take their responsibility of protecting the public very seriously, and are genuinely worried by their increasingly precarious ability to do so with the resources at their disposal. While the world grows more complicated and food and drug companies become bigger, richer, and more powerful, the FDA struggles valiantly to keep up. We have, perhaps, grown complacent, because the FDA has so successfully anticipated and protected us from ghastly food and drug disasters that have never been. But one must ask whether we can afford to skimp on an agency that oversees such a huge number of products of such critical importance, or whether it will take another major food or drug catastrophe for us to appreciate an agency that we often take for granted, as the terrorist attacks of 2001 forced us to readjust our opinions on safety, national security, and individual rights.

This article was supported by a grant from Understanding Government.