Civil Justice Memo
No. 6 February 1988
The Liability and Regulatory Threat to Biotech
Huber argues that it is not the stringency of biotechnology regulation that is harmful to development, but rather the chaotic manner in which administrative agencies operate. "One agency's approval may be trumped by a second's disapproval," he writes, while a Federal court may overrule both. The result is a legal system that cannot say "yes" to new technologies, only "maybe." For example, Genentech encountered delays and expenses while the USDA and FDA argued for more than a year over which agency should regulate that company's bovine interferon.
Conflicts also occur among local, state and Federal authorities, and even within single agencies, further disrupting the development process. Moreover, the current liability system "guarantees that the regulatory process is never really concluded. Judgements on a product's acceptability often come only decades after the original research development and commercialization." Today, full compliance with regulations does not afford liability protection in the courts. All this adds up to an environment in which biotechnologists face not only the natural uncertainty of the scientific realm, but of the legal and political as well. The nation's competitiveness in this vital field is in jeopardy.
--Daniel Shuchman, Editor, Technology Review
Disorder in the regulatory process threatens to undermine biotechnology in the United States
Biotechnology and the Regulation HYDRA
By Peter W. Huber
“You almost have to be a lawyer before you can be a scientist.” So says Gary Strobel, professor of plant pathology at Montana State University. Last June, Strobel injected elm trees with bacteria whose natural capacity to secrete a toxin active against Dutch elm disease had been enhanced through genetic engineering. He deliberately chose not to secure the required permission: he characterizes his act as civil disobedience, in protest of the number and complexity of regulations affecting geneticengineering R&D.
While most biotechnologists have condemned what Strobel has done, many share at least some of his frustration. After all, conventional "genetic engineering"such as animal husbandry and systematic crop breedingfaces few regulatory obstacles. But recombinant DNA technology is the target of abundant regulatory effort, even though its end products are often biologically indistinguishable from those developed by more traditional work.
A panel of the National Academy of Sciences recently concluded that genetically altered organisms pose no unique ecological hazards, and that regulation should be tailored to the nature of the organism rather than to the means used to create it. But recombinant DNA products look different from others in the public eye. The regulatory agencies and courts are therefore likely to give short shrift to the academy's study. The system used to govern biotechnology will probably not change much.
At first blush, the regulatory system seems reasonably orderly. Administrative agencies provide oversight before products go on the market, while the courts supervise matters farther down the line. But the structure beneath is much more chaotic. The hierarchy of regulatory powers is so fragmented that the system can never say "yes," only "maybe." One agency's approval may be trumped by a second's disapproval. Approvals by two agencies may be refuted shortly afterward by a federal court. And approvals by all three may be rejected by a liability court following an accident decades later.
Gary Strobel has made his point the wrong way, but he has a point nonetheless. The gravest regulatory threat to the development of biotechnology lies not in the stringency of regulation, but in its ponderous disorder. The economic stakes are high, and the U.S. lead is vulnerable. France has launched a major government program in the area. Japan's Ministry of Trade and Industry is working closely with that country's companies in an effort clearly directed at capturing world leadership. We need more expeditious, predictable, and coherent regulation if we are to preserve our head start in this critical new field.
Struggles among Agencies
The problems begin in the administrative agencies. Conflict among federal regulatory officials over biotechnology is notably common because many agencies may have jurisdiction over a genetically altered product. The federal agencies that can become involved include the Environmental Protection Agency (EPA), U.S. Department of Agriculture (USDA), Food and Drug Administration (FDA), Occupational Safety and Health Administration (OSHA), and National Institutes of Health (NIH). Any of them may conclude that certain risks are unacceptable. Effective regulatory approval can thus require a quiver of licenses.
Agricultural products are the most likely to run into regulatory confusion because of the many occasions for intervention. A genealtered vaccine destined for hogs requires review from NIH while any federally funded experimentation is in progress. When the vaccine is used commercially, USDA has reviewing powers if, for example, the VirusSerum Toxin Act applies. EPA becomes engaged through the Toxic Substances Control Act and other statutes. In addition, OSHA oversees the vaccine if farmworker exposure is an issue, and FDA enters the scene if residues can be detected in bacon.
Genentech reportedly encountered needless delays and expenses while USDA and FDA argued for more than a year over which agency should regulate the company's new bovine interferon. The agencies were unable to decide whether the product was a "veterinary biologic" under USDA's jurisdiction or a "new animal drug" under FDA's control. Genentech finally licensed the product to CibaGeigy Corp. in 1985. (That company is now fieldtesting bovine interferon under FDA's jurisdiction, with the dispute resolved through an interagency memorandum.)
Another recent, muchpublicized incident suggests how varying rules from different agencies may contribute to confusion. Advanced Genetic Sciences, Inc., had deleted genes from a common plant bacterium. The company wanted to test how well this altered organism, when sprayed on certain crops, would reduce the risk of frost damage. Last year, as part of the many tests the government requires for a genetically engineered bacterium, scientists injected the organism into the bark of fruit and nut trees located on the rooftop of a company building in Oakland, Calif. The research apparently complied fully with NIH guidelines for a release into the environmentthe first such rules written. But in this case EPA approval was required instead, and that agency's regulations concerning environmental release were more stringent. EPA decided that the rooftop testing constituted an unauthorized release. It fined the company and required that the tests be repeated in their entirety in a greenhouse.
Confusion and conflict can even occur within single agencies. In January 1986, after two years of review and field tests, USDA's Animal and Plant Health Inspection Service licensed Biologics Corp.'s pseudorabies swine vaccine for commercial use. The inspection service claims it complied with all regulations on the books at the time, but the vaccine was not processed through the department's Recombinant Advisory Committee. According to news reports, critics of the inspection service in USDA's Science and Education Division told the press that the vaccine had not received proper approval. A lawsuit was subsequently filed, and sometime this fall a federal district court in Washington, D.C. is expected to determine whether the inspection service violated any regulations. A ruling against the service will undoubtedly set the stage for lengthy appeals.
Agricultural products are not unique. A drug for human use may require approval from FDA because of the patients, NIH because of researchers, and OSHA because of production workers. An industrial process using genealtered enzymes may invite EPA, OSHA, and FDA oversight.
Some semblance of order in the system started to emerge following the publication of a major study by the White House Office of Science and Technology Policy (OSTP) in June 1986. The report, entitled the "Coordinated Framework for the Regulation of Biotechnology," summarizes relevant federal laws and policies and concludes that no new legislation is needed. It urges that agencies adopt scientifically consistent definitions and standards, and that a single agency exercise responsibility over a given product whenever possible. For cases in which more than one agency has statutory authority, the report establishes a lead agency and provides for coordinated regulatory review. At least one interagency group, the Biotechnology Science Coordinating Committee, has formed as a result of OSTP's work. This group has worked on defining terms that heretofore have had different meanings for various agencies, such as "release into the environment."
The OSTP initiative and the interagency coordination that has started to take shape are commendable. But the record of agency conflicts, and the confusion that they have created in the industry, are not encouraging. As the number of biotechnology products increases, interagency conflict will likely recur.
Relations among federal, state, and local governments raise more delicate questions regarding jurisdictional turf and bureaucratic sensibilities. One early local initiative was a threeweek moratorium by Cambridge, Mass., on recombinant DNA research at M.I.T. and Harvard. Later, in 1976, the city council passed an ordinance making NIH guidelines for governmentsponsored research applicable to any projects conducted in the city. The ordinance also imposed additional safety requirements and banned deliberate releases of living, genealtered products and socalled "biosafety level 4" or "BL4" experiments, which NIH considers to involve especially dangerous or contagious organisms.
Local initiatives of this type undoubtedly have an impact. For example, Advanced Genetic Sciences originally wanted to test its frostsuppressant bacteria in Monterey, Calif., but county officials blocked the experiments. Then, after the company received the goahead from two other California counties, the EPAapproved tests were performed under laughably excessive regulation. The scientists who sprayed the wholly innocuous bacteria were required by the California Department of Food and Agriculture to wear suits seemingly appropriate for a moonwalk: fully sealed and complete with helmet, gloves, and breathing pack. Both state and federal regulators also required a barrier around the test site to prevent inadvertent or deliberate trespass.
Monsanto encountered comparable problems with local regulation after it isolated a gene that holds great promise for controlling the black cutworm, a serious corn pest. The company inserted the gene into a microbe so that a protein lethal to this parasite could be created. The county in which the first tests of the resulting microbial pesticide were to be conducted supported Monsanto, but city officials in nearby St. Charles, Mo., passed a resolution opposing the company's efforts. The resolution was purely advisory since the tests were to be conducted outside town lines, but EPA has delayed approval nonetheless.
It's interesting to note that Monsanto meanwhile inserted the same gene into a tomato plant to make it resistant to certain parasites, then secured USDA approval for experiments on this crop. The tomato field tests were conducted across the state line in Illinois, where rural communities proved accommodating.
State regulatory initiatives are evolving, too. New Jersey, for example, is considering legislation to establish a Commission on the Release of Genetically Engineered Microorganisms. That organization would not only monitor compliance with federal regulations but review the adequacy of existing state laws. California has considered several regulatory bills of different severity, but so far the state legislature and a special task force have concluded that the existing matrix of environmental regulation suffices. Other states have also considered initiatives.
The international reach of U.S. safety and environmental standards is yet another troublesome issue. At least two genealtered vaccines synthesized in the United States have been deliberately taken abroad for testing under more hospitable regulatory oversight. The Wistar Institute, a Philadelphia biomedical institution that has long pioneered vaccine development, performed early tests of a new bovine rabies vaccine in Argentina. And researchers at Oregon State University went to New Zealand to conduct field trials of a vaccine against the insect-borne Sindbis virus, which plagues animals. NIH is investigating whether these offshore tests have violated federal rules. A finding that they have could convince American companies to conduct joint ventures with concerns based elsewhere. Some biotechnology companies are already talking about collaborating with Japanese firms, so that tests can be conducted abroad without U.S. government oversight.
The Courts as Gatekeepers
Relations between the federal agencies and the courts regarding biotechnology are at least equally confused. The National Environmental Policy Act (NEPA), passed in 1969, empowers courts to review agency actions that will have a "significant impact" on the environment. The act also requires agencies to prepare an environmental impact statement in connection with such projects. NEPA's reach extends to agency approval of private ventures as well as federally funded projects. The phrasing of NEPA is sufficiently hazy to permit courts that oversee it to act as fullfledged gatekeepers in the approval process.
The courts have already done so on several occasions. In May 1984, Judge John Sirica of the Washington, D.C. federal district court invoked NEPA to block the NIHapproved test that researchers at the University of California wanted to conduct on a frostinhibiting bacterium similar to the one developed by Advanced Genetic Sciences. NIH had not filed an environmental impact statement for the test. Geneticengineering opponents are also trying to use NEPA to halt University of Pennsylvania experiments in which genes that produce human growth hormone have been transferred into mouse, pig, and sheep cells. Several NEPA suits have been filed by one organization critical of genetic engineering, the Foundation on Economic Trends. We can expect NEPA litigation to burgeon.
The nuclear power industry knows this situation well. That business has had particularly vivid experiences with the game of "NEPAgoseek," in which a reviewing court repeatedly finds something wrong with an agency's environmental impact analysis and sends officials off to investigate another detail. After the Natural Resources Defense Council sued the Nuclear Regulatory Commission (NRC) over rules it developed for nuclear waste disposal, the circuit court of appeals in Washington, D.C., struck down the regulations. NRC took the case to the Supreme Court, which reversed the decision and remanded the case to the circuit court for further consideration of certain issues. The court struck down the rules again, this time on different grounds, and in an appeal the Supreme Court reversed the decision again. The process took about a decade.
Judicial secondguessing of administrative licensing decisions can also come through the tort system, which addresses one category of civil wrongs. This part of the regulatory picture is less clear because of the dearth of significant accidents with genetically altered products. But recent experience with pharmaceuticals and herbicides suggests that even if the companies or individuals releasing these materials are not negligent, they will be held liable for all damages that juries may conclude were caused by exposure.
Tort law expressly accepts that the agencies and the courts may send very different regulatory messages. The 1964 Second Restatement of Torts, assembled by a distinguished panel of lawyers on behalf of the American Law Institute to summarize liability law, flatly declares that "compliance with legislative enactment or an administrative regulation does not prevent a finding of negligence." In 1984, for example, the Karen Silkwood lawsuit established that complete compliance with NRC safety standards does not protect a nuclear operator from an award of punitive damages in a state court. Thus, a jury may decide that conduct deemed fully prudent by federal regulatory standards is in fact "outrageously careless." And a judge or jury can then assess an unlimited fine against the responsible party. Biotechnology entrepreneurs must recognize that under current law, complete compliance with all federal and state regulatory requirements will afford no liability protection in the event of an accident.
More troubling is the implicit idea that any accidenteven if it does not harm people or propertyis probably serious enough to deserve monetary compensation. Following the incident at Three Mile Island, for example, the utility paid millions of dollars to settle personalinjury and property damage lawsuits out of court. All subsequent scientifically reputable studies by the Commonwealth of Pennsylvania and others have concluded that the accident had no measurable impact on the physical health of nearby residents.
Even wholly unsuccessful suits can sink a valuable product if brought in sufficient number. In 1983 a cascade of costly litigation drove the drug Bendectin, prescribed for morning sickness, off the market. On the basis of more than 20 epidemiological studies, FDA believed Bendectin was safe. While a handful of juries voiced their disagreement through verdicts requiring multimillion dollar payments, the vast majority of judges and juries reached the same conclusion as FDA. Merrell Dow Pharmaceuticals, Inc., the drug's manufacturer, found that even if it won suits, it spent much more defending and insuring itself than it earned in profits from sales.
Erratic Regulation: Drag on Development
One obvious problem with our multilayered system of regulation is that all phases of the process are painfully slow. Delays running into months and years are now routine. And the liability system as it currently operates guarantees that the regulatory process is never really concluded. judgments on a product's acceptability often come only decades after the original research, development, and commercialization. A firm that embarks on a venture to bring new biotechnology products to the mass market is betting its future on this unpredictable system.
A second problem concerns the expense and competitive disadvantage of having to hurdle high regulatory barriers very early in a product's lifecycle. A company seeking to develop a genetically altered pesticide, for example, must invest large sums at the very beginning to satisfy agency requirements. But the same manufacturer can test hundreds of conventional chemical alternatives and select only the most promising before having to do battle with local and Washington officials.
Most worrisome is the fact that the uncertainty permeating the regulatory system can be exploited by those seeking to retard biotechnology. Emphasizing the unknown potential for risks, and taking advantage of the system's divisions, these people achieve their greatest successes when decisionmakers in the courts and in several state and federal agencies collide.
Biotechnologists thus face uncertainty in both the scientific and regulatory domains. A company cannot be sure that a product it is trying to make will work. Nor can it be sure that regulators will approve the product. And the firm has no idea of the ultimate cost to defend itself in lawsuits that may materialize later, whether or not they are justified. While uncertainty on the scientific side creates both problems and opportunities, uncertainty in the regulatory system offers only negative returns.
When Genentech, for example, undertook research on tissue plasmogen activator (TPA), which dissolves blood clots following heart attacks, the company knew little about how effective the drug would prove, or how feasible or costly it would be to develop. But there was also the possibility for enormous profit if all worked well. TPA could be one of the first highrevenue products of biotechnology, with several million dollars in sales a year. But last May Genentech suffered a major setback when an FDA advisory committee concluded it had insufficient data to support TPA approval. Although the drug will almost certainly be approved eventually, it is now undergoing more studies at further cost to the company.
Any endeavor can tolerate only so much uncertainty. Compounding scientific doubt with unnecessary layers of regulatory unknowns will sink many undertakings regardless of their scientific and economic merits.
Pressing for More Collaboration
A strong biotechnology industry ultimately depends on a regulatory system capable of making expeditious and definitive judgments. We would do best to switch from local control to national regulatory oversight, because both the benefits and risks of biotechnology transcend state boundaries. Genetically engineered products can hardly be successfully commercialized in the United States if their development can be stopped at the penultimate moment by shifting, nonuniform demands. Administrative power over any particular biotech venture should be firmly lodged in a single, identifiable agencyand in just one division of that agency.
Sweeping new federal legislation to consolidate regulatory authority is possible. But the idea will seem most attractive to those least familiar with the Washington political process. A bill that is well crafted and clean when it is drafted rarely emerges from Congress in the same condition. A new statute designed to streamline could easily end up adding one more layer of decisionmaking authority to the system. For the time being, the best political bet is to stick with the laws on the books while pressing for less local government involvement and more of the interagency collaboration the OSTP report calls for.
It is difficult to specify the means by which the courts that review agency actions through NEPA should hold their authority in check. Clearly they have an important responsibility to describe finite, comprehensible requirements that reasonably conscientious agencies can satisfy. In the end, reviewing judges will have to do what is often hardest for them: resist the temptation to engage in endless finetuning.
The potential liability problems, though the least apparent and immediate today, remain the most intractable. Perhaps the best prescription one can realistically suggest is that the courts should give greater deference to decisions by administrative agencies. An agency's full, unqualified approval of a test or commercial operation should be viewed for what it is considerably more than a routine pleasantry that may be dismissed as soon as a tort plaintiff enters the courthouse. Complete, goodfaith compliance with applicable safety regulations must provide some safe harbor.
No regulatory system will ever be "right" about every issue clouded in scientific doubt. But an industry should be able to ask for a process that will be reasonably quick and definitive. Regulatory judgments should not be subject to revision every time the locus of authority shifts among federal and local agencies and the courts.