Unproven AIDS Therapies: The Food and Drug Administration and ddI
The scientific decision-making process in the case of the acquired immune deficiency syndrome (AIDS) has been characterized by unprecedented involvement of the persons affected by the disease. Initially, the AIDS crisis involved one group of people who were already politically organized—in this instance, gay men. Most other diseases affect a far more heterogeneous population from the start. By the time the first AIDS case was diagnosed in 1981, gay men, together with lesbians, had become a growing political presence in U.S. society and were already psychologically and strategically primed to challenge the system.
Indeed, gay men came to the AIDS crisis with a predisposition to question and distrust the health and scientific establishments. In many of the larger cities across the nation, gay men had formed gay health clinics some 10 to 15 years before the advent of AIDS because of their distrust of the medical establishment in treating gays for sexually transmitted diseases. Past actions by health agencies fueled the distrust. For example, the Public Health Service (PHS), of which all the federal agencies responding to AIDS are a part, including the Food and Drug Administration (FDA), was, until late 1990, responsible for enforcement of a ban on immigration by homosexuals based on the determination that gays and lesbians are mentally ill—even though the
Jeffrey Levi is a Washington-based health policy consultant working with several groups on AIDS issues, including the National Gay and Lesbian Task Force, the Gay Men's Health Crisis, and the Institute of Medicine.
American Psychiatric Association removed homosexuality from its list of illnesses in 1976. It is from this past experience that AIDS activists of all stripes—from mainstream lobbying organizations to street groups using direct action (demonstrations and civil disobedience)—came to challenge some of the fundamental assumptions of the public health community's response to AIDS: from traditional public health control measures to how research ought to be conducted to determining who should make decisions about access to experimental therapies.1
Decision making regarding early release of dideoxyinosine (ddI), a promising antiretroviral drug that has been seen as a potential replacement for the often highly toxic drug zidovudine (AZT) in the treatment of human immunodeficiency virus (HIV) infection, occurred in a context of more general discussions within the AIDS scientific and activist communities about early access to experimental treatments for persons with AIDS that were to be offered on a "parallel track" with ongoing clinical trials. In effect, ddI became a prototype for early access before a model for implementing the broader parallel track program was developed or approved by the relevant government agencies. The success of the parallel track concept will be closely linked to the improvised approach developed for ddI.
The support for early release of ddI and the general notion of parallel track represented a remarkable shift in attitudes among government researchers and regulators, a change inspired by discussions with and pressure from the AIDS activist community. The endorsement of parallel track by top PHS officials occurred in a context of agency jockeying for support from the AIDS activist community. Parallel track and early release of ddI have also resulted in an unusual confluence of interests among AIDS activists, regulators, drug companies, and some key scientific researchers. Government officials saw early release as a means of showing compassion and responsiveness at a time when existing research and regulatory structures seemed rigid and uncaring. Drug companies and top government regulators saw early access as another step toward the reduced regulation of the pharmaceutical industry that was a hallmark of Frank Young's tenure as FDA commissioner. Finally, activists saw parallel track as providing greater autonomy in decision making for persons with HIV infection.
This paper reviews and describes the decision-making process that led to the early release of ddI. It is not an attempt to judge the merits of early release of ddI or the entire parallel track concept. Rather, it is meant to paint a picture of how and why decisions were made, with the hope that it might provide a basis for better, more rational decision making in the future.
To lay the groundwork for the discussion of early release of ddI, the
paper first reviews the history of increased drug regulation in the United States. It then discusses the trend toward deregulation during the Reagan administration that coincided with the AIDS crisis, the initial discussions and pressures for creation of a parallel track, and, finally, the process leading to early release of ddI. The paper also assesses some of the motivations of the actors involved.
The sources for the paper are primarily interviews of participants in the process conducted by the author. The outcome is necessarily the author's synthesis—but one that, it is hoped, reflects what actually happened when the analysis of the individual participants and the public record are pieced together.2
THE DRUG REGULATION PROCESS
Federal regulation of the drug industry in the United States 3 is a product of twentieth-century legislation that, until the 1980s, produced increasingly tight restrictions on pharmaceutical companies. Prior to this century, regulation of the food and drug industries was left to state and local governments, but as the federal role increased, the state role in this area began to disappear.4 Each significant federal initiative for closer regulation of the drug industry coincided with a major catastrophe with a drug, which brought political pressure for greater premarketing protection of consumers.
The first major federal legislation designed to protect consumers was the Food and Drugs Act of 1906. No premarketing approval was proposed; the law merely required that drugs meet official standards of strength and purity. The next major legislation was the Federal Food, Drug, and Cosmetic Act of 1938, prompted by the elixir sulfanilamide scandal in 1937. In this case the drug was marketed without any safety testing and caused the death of 107 persons because it contained a chemical commonly used in antifreeze. The 1938 law required that a manufacturer prove the safety of a drug before it was marketed. Thus, from 1938 until 1962, regulation was focused on safety and involved minimal review by the FDA. A manufacturer simply sent a new drug application to the agency, and if there was no response within 60 days, the drug was deemed to be approved. The FDA had authority to veto an application, and it also had the option of using informal authority to try to convince a company not to market a product.
In 1962, in reaction to the thalidomide scandal overseas, the Kefauver amendments to the Food, Drug, and Cosmetics Act resulted in a requirement that manufacturers demonstrate not only the safety of a drug but also its efficacy.5 This requirement led to far more complex clinical trials and the regulatory system that is the basis for today's drug approval process.
The Drug Approval Process Today
A pharmaceutical company wishing to begin human clinical trials of a new drug must first submit an investigational new drug (IND) application to the FDA. The IND must contain laboratory data on the drug, the results of any animal studies or foreign trials, and the proposed research protocols for trials in humans. Trials may begin within 30 days of submission of an IND if the FDA does not order a hold on clinical trials.
Although not required by regulation, there are generally three stages to the clinical trials process. During Phase I, initial safety studies are conducted, usually among a small number of healthy patients, with gradual increases in dosage to determine safe levels. These trials average between six months and one year. Because they normally take place among healthy research subjects, usually no data are collected regarding efficacy. If a drug is considered toxic, however, as is the case for many cancer and AIDS drugs, Phase I trials are conducted among those who are already sick. As a result, some information regarding efficacy may be obtained, which is the basis for some of the pressure for early release of experimental drugs. These trials were not designed to test for efficacy, however, and as a result it is risky to draw any firm conclusions about the potential usefulness of a drug from Phase I trials. About 29 percent of drugs do not continue in trials past the Phase I stage.
Phase II trials test the effectiveness of the drug and provide further evidence on safety. These studies, which usually take up to two years and involve several hundred patients with the disease for which the therapy is intended, are designed to assess the value of the drug as a treatment. Another 39 percent of drugs under development fail at this stage.
Phase III trials are long-term safety and efficacy studies, involving thousands of patients at numerous research centers, that are designed to assess the risk-benefit value of the drug. They take between one and three years. Very few drugs (3 to 5 percent) fail at this stage. Indeed, there are some in the scientific community who believe that the distinction between Phase II and Phase III trials is often artificial and that, in fact, the two phases run into one another at a certain point in the research process.
At the end of Phase III, the FDA begins its formal review of the data. A new drug application (NDA) is the basis for the agency's final approval of a drug for marketing. The NDA summarizes the findings of all the research, and the FDA has 180 days to approve the application. After the drug is placed on the market, the FDA and the pharmaceuti-
cal company monitor it for unexpected toxicities, which could cause the FDA to revoke the marketing approval. In some instances, the FDA may even require what are called Phase IV, or postmarketing, studies to learn more about the drug.
Speeding Up the Process: The “Bush Initiative”
The length of the drug approval process has been a source of frustration to the drug industry for many years and a source of anger to many advocates for patients with life-threatening diseases. Partially in response to this pressure, and with the urging of the President 's Task Force on Regulatory Relief, chaired by then Vice President George Bush, the FDA announced in the fall of 1988 new regulations to expedite availability of promising therapies for patients with life-threatening and serious diseases. In a sense, the purpose of the regulations was to short-circuit the three clinical trials phases and, in the words of then FDA Commissioner Frank Young, “be able to reach a scientifically defensible decision to approve or disapprove marketing of drugs intended to improve the outcome in such diseases, based on the results of well-designed Phase II controlled trials ” (Young, 1989).
Essentially, the new approach assumes that patients and physicians will use a different risk-benefit analysis for drugs to treat life-threatening illnesses and will be willing to use those drugs with less complete data than are usually available for approved drugs. The FDA's new regulations called for early consultation between the FDA and drug sponsors in the preclinical stage to ensure a clear understanding between regulators and industry as to the data required for approval. The regulations also permitted submission of an NDA after Phase II. If the drug were approved, the FDA might still require postmarketing studies (referred to as Phase IV above).
As mentioned earlier, the distinctions among the clinical trials phases are considered artificial by some, and so it is difficult to assess how dramatic this change will be. On the other hand, the more active involvement of the FDA in the design of trials is also dependent on the willingness of the sponsoring company to work actively with the FDA early in the process and on the availability of FDA personnel, who are already stretched thin owing to agency understaffing. It is too early to know what effect the new procedures are having on drug approval times.
Under ordinary circumstances, access to a drug that is still in the investigational (IND) stage is limited to those patients participating in
the clinical trials. The FDA, however, has a number of regulatory options to make the drug more widely available under certain circumstances. The primary vehicle is the treatment IND, which allows distribution of a promising therapy to limited populations before completion of an NDA.
Prior to 1987, the uses and criteria for a treatment IND were relatively vague, which left tremendous discretion to the FDA regulators. The FDA issued new regulations in 1987 that attempted to clarify procedures and standards. According to these regulations, a treatment IND can be issued at any time between the end of Phase I trials and the submission of an NDA for promising therapies to treat the desperately ill—those with immediately life-threatening illnesses—provided there is no comparable or satisfactory alternative therapy that has already been licensed.
The FDA also has a special mechanism for early access to promising cancer treatments. Known as Group C drugs, these agents are investigational drugs under development at the National Cancer Institute that are shown to have a certain level of efficacy but are some time away from full NDA approval.
The primary criticism of the treatment IND process offered by patient advocates, particularly in the AIDS community, was that the standard of proof for treatment INDs was almost as high as that for an NDA and that in practice the treatment IND was functioning simply as a bridge between Phase III and the NDA while the FDA reviewed the data. Some in the FDA and in the research community were concerned that if treatment INDs were granted any earlier, the ability to collect good research data might be compromised and would delay even further the final NDA for a drug. This concern was based on the assumption that if a drug were available outside of clinical trials, few people would enroll in randomized trials because there was no guarantee of actually receiving the drug in question.6
It is in this context—with frustration over the length of time it took new drugs to move through the entire drug approval process and a feeling that existing mechanisms for early release of promising drugs were not being sufficiently employed—that AIDS activists began to call for a new mechanism for earlier release of AIDS drugs in the drug development process. This new mechanism is what became known as parallel track.
The concept of parallel track—a system for expanded access to experimental therapies—originated in the AIDS activist community and
posed some fundamental challenges to assumptions within the scientific community about how clinical research should be conducted and how access to drugs should be regulated. The most formal iteration of the parallel track concept came from the Treatment and Data Committee of the New York-based AIDS activist organization ACT UP (AIDS Coalition to Unleash Power), a direct action group more often seen protesting government policies than developing them. But the Treatment and Data Committee had also acquired a reputation for substantive understanding of AIDS drug development issues, and, more often than not, this arm of ACT UP could be found at the conference table at the National Institutes of Health (NIH), the FDA, or the PHS, particularly during the course of the discussions around parallel track and early access to ddI.
Jim Eigo, a leader of the Treatment and Data Committee, defined parallel track as follows: “At the beginning of phase two (efficacy) trials, this parallel track would make investigational AIDS drugs available to people with HIV disease who are ineligible for the drugs' clinical trials and have no reasonable treatment alternatives” (Eigo, 1989). As Eigo explains it, the parallel track concept grew out of the failure of existing mechanisms to “deliver drugs to people with serious or life-threatening conditions who have no treatment alternative before full FDA marketing approval” (Eigo, 1989). It was first presented in April 1988 to principal investigators and officials of the National Institute of Allergy and Infectious Diseases (NIAID), the lead NIH agency working on AIDS. This presentation was the start of a long series of discussions and an education process directed at the NIH leadership, in particular NIAID's director Anthony Fauci, primarily through conversations with ACT UP and Project Inform's Martin Delaney. (Delaney, who heads the San Francisco-based group, is one of the leading voices for greater patient access and autonomy in decision making regarding the use of experimental therapies.)
Fauci was probably the most critical player in the transformation of parallel track from an idea advocated by outsiders to one embraced by the federal public health establishment. He initially held the standard scientific community view on these proposals—that scientific controlled trials were the only way to establish the effectiveness of a drug and that expanded access could not be allowed to compete with these trials. Fauci believed the success of those trials was paramount as well as the most compassionate route in the long term, and would ultimately help more people than would be helped through early access.
Through his conversations with the activists, however, Fauci became convinced that expanded access would not have to compromise the integrity of the trials if the parallel track was limited to those who could
not otherwise participate in a clinical trial. There was also, at this time, growing concern within the NIH-funded AIDS clinical trials program that patients were not complying with the trial protocols. If patients did not see the clinical trials as their only means of obtaining a drug, it might be possible to conduct more efficient trials and have better patient compliance.
Although there were hints from NIH officials at the Fifth International AIDS Conference in Montreal in early June 1989 that there might be more flexibility in NIH's position regarding earlier access, the formal declaration of a change of heart occurred in a speech given by Fauci in San Francisco later that month. “My commitment to carefully designed, controlled clinical trials for AIDS has not changed,” said Fauci. “Such trials are absolutely essential if we are going to get the answers needed by physicians who are treating patients.” But, he declared, “[a]t the same time, we have to be creative and flexible so we can provide increased access to promising drugs to patients who cannot participate in clinical trials” (Zonana and Cimons, 1989).
Once Fauci broke the ice, there was almost a bandwagon effect on the rest of the PHS leadership. Fauci had given no warning to his colleagues in the FDA or to Dr. James Mason, the assistant secretary for health. Mason, who was Fauci's boss, reportedly was angry at receiving no advance notice that such a major initiative was going to be announced without prior consultation or approval. FDA Commissioner Young, eager to maintain his leadership position as an early advocate of quick access to experimental drugs, was quoted as saying, “I've been pushing it [parallel track] as much as Tony has” (Kolata, 1989a).
Fauci presented his endorsement of the concept in principle without detailing how the program would actually work. Essentially, he announced that early release would not interfere with NIH's conduct of scientifically sound clinical trials and that personally he felt there was a moral obligation to make such therapies available earlier. He then tossed the ball into the FDA's court, correctly saying that, from a legal standpoint, all decisions regarding early release had to come from the regulatory agency, not the research arm. This maneuver was a source of considerable annoyance on the part of the other branches of the PHS, which were now handed a hot political potato with no prior warning or discussion. Overnight, Fauci became the hero of the activist community and, from the perspective of the FDA at least, made the regulators into the stumbling block to reform.
It was not just the regulators and his political superiors in the assistant secretary's office whom Fauci took by surprise. Little if any groundwork had been laid within the research community or within the NIH AIDS program. This lack of preparation left considerable room for open op-
position (or covert backbiting) to the proposal from the traditional research community. Indeed, some principal investigators in the AIDS Clinical Trials Group (ACTG) of NIAID suggested that Fauci was undercutting good science. Fauci did not have direct discussions on this topic with the principal investigators until the regular meeting of the ACTG the following month.7
There has been some suggestion that Fauci, in announcing the parallel track initiative, was trying to abate the rather harsh personal criticism he had been receiving from the activist community. In a sense the activists were the squeaky wheel that got the attention. Little was done to address the consequences of a parallel track endorsement to the rest of Fauci's political base—or the impact on internal agency morale of such a surprise announcement. It fell to the assistant secretary's office, through its National AIDS Program Office (NAPO), to try to produce a bureaucratic consensus on parallel track and ensure that the PHS was speaking with one voice. Essentially, Fauci was told that, although NIH should participate in discussions on parallel track, FDA and NAPO would be taking the lead roles.
The need for a PHS-wide consensus was heightened by the pressures of a congressional hearing, a common mechanism to force decisions within the bureaucracy. Congressman Henry Waxman called a hearing on parallel track for July 19, 1989. As chair of the House health subcommittee, Waxman was the key AIDS legislator in the House and was concerned about the momentum Fauci's idea had gained. He was also wary of how such a policy was to be implemented and how the broader research community would react.
At the hearing, Mason presented an administration position on parallel track that was astonishingly similar in its broad outlines to that of the AIDS activist community. Said Mason, “As contemplated, the availability of investigational therapeutic agents through this mechanism would be limited to those persons for whom there are no satisfactory alternative drugs or therapies available to treat that stage of disease and who, for some reason, are not eligible for or not able to participate in a clinical trial” (Mason, 1989).
Although initially parallel track seemed like a simple proposal, NAPO and the FDA soon realized that its formulation was going to be quite complex, and that nailing down the details of the policy was going to take some time. At the hearing on July 19, Mason announced that the many unresolved questions around parallel track—which ranged from safety monitoring of released drugs, informed consent, and specific patient eligibility criteria to liability and reimbursement issues —would be presented to an expanded meeting of the FDA's Anti-Infective Drugs Advisory Committee on August 17. In preparatory meetings with PHS,
activist, and research community representatives, however, it became clear to NAPO officials that a much lengthier process would be needed. The August 17 meeting became, instead, a public forum for the discussion of the various options and positions, pro and con, on the parallel track concept. At that meeting, NAPO Director James Allen announced that a smaller working group would be established to develop guidelines for parallel track that included the PHS agencies as well as researchers, activists, care providers, and pharmaceutical company representatives.
The concepts of parallel track and ddI were first linked publicly after Fauci's San Francisco speech. When asked what would be a likely first candidate for parallel track, Fauci suggested ddI, although at the time only very preliminary discussions were taking place at FDA and with the drug's manufacturer, Bristol-Myers, regarding some form of early release. While ddI was always closely linked with the concept of parallel track, once it became clear that it would take time to develop the concept into a working system, ddI and parallel track decoupled, at least in terms of a bureaucratic response. As early as the July 19 congressional hearing, Assistant Secretary Mason stated that efforts to release ddI would not be constrained by any delays in the parallel track discussions. In a sense, ddI and parallel track were on parallel tracks of their own. As discussions were held at the NAPO level regarding guidelines for the parallel track policy, the FDA, NIH, Bristol-Myers, and AIDS activists were negotiating protocols for the early release of ddI. In these talks everyone took pains to make it clear that the ddI package was not meant to be the prototype for parallel track. Nevertheless, everyone also was very conscious that the success or failure of early release of ddI would greatly influence the outcome of the overall parallel track policy development.
PARALLEL TRACK: PROS AND CONS
Because attitudes toward parallel track are so closely tied to attitudes about early release of ddI, it is important to consider how the various players perceived parallel track before looking at some of the specifics of the decision making regarding ddI.
The Food and Drug Administration
The push for support of parallel track as a separate, new mechanism came from the commissioner's office. Frank Young's tenure at FDA had been marked by efforts to speed up the drug review process and reduce the level of regulation without compromising the safety and ef-
ficacy standards required by law. His 1987 effort to create greater flexibility in the use of treatment INDs, as discussed earlier, was met with skepticism by some patient groups who felt that FDA was merely repackaging old ineffective concepts and with criticism by some members of Congress and consumer advocates who believed FDA was allowing questionable reductions in safety and efficacy.
Young clearly communicated to career officials in charge of the drug review process that he wanted them to act expeditiously on parallel track and early release of ddI. At this juncture, Young was increasingly an embattled commissioner, with his administration of the FDA under close scrutiny by congressional oversight committees. He perceived the AIDS activist community as a source of strong support, perhaps the last stronghold, and was eager, if not anxious, to be seen moving the bureaucracy along on this issue.8
Alone, however, Young could only do so much. Considerable power in the federal bureaucracy is held by career officials, who make most of the regulatory decisions. In the minds of many activists, as well as some at NIH, one of the stumbling blocks to quicker approval or access to promising drugs was Ellen Cooper, then head of the Anti-Viral Drug Products Division. Cooper was perceived by some to be rigid in her adherence to traditional standards and inflexible in adapting to the new pressures associated with the AIDS epidemic. This perception changed somewhat during the course of the negotiations around release of ddI. Activists sensed a change of heart, and Cooper felt the activists were finally listening to what she had been saying all along.
Certainly, career officials at the FDA were angry with Fauci at his refusal to consult with them before his speech on parallel track. They had heard of the proposal and had actually sought out discussions with Fauci but were ignored. Yet despite some lingering frustration and anger, the bureaucrats were pushed forward by deadlines on both the parallel track and ddI fronts set by the political appointees, including their own commissioner. This haste was a source of annoyance, as some felt that their responsibility to ensure that all decisions were based on sound data or policy was being undermined.
To some career FDA officials, parallel track was simply the articulation of what had been their intent all along. In their view, the mechanisms for early release were there but had not been formalized, a condition that in some ways made the system potentially more flexible. In fact, they traced some of the difficulties in early access to Young's formalization of the treatment IND rules. This policy, they said, created false expectations among the AIDS activists and resulted in pressures to totally overhaul the system.
These officials were concerned that parallel track might be hard to
contain; that in implementing policies that showed sympathy for the dying, the FDA still needed to maintain an orderly system of access to experimental agents. Some officials felt that activists, in their frustration with some of the kinks in the established system, wanted to throw it out entirely rather than devise ways to make it work better.
Career officials were also concerned about a “drug of the month” mentality on the part of activists. Because activists had been correct in arguing that AZT and aerosolized pentamidine were effective therapies, many believed they should be able to obtain early access for any drug they considered promising. Scientists, on the other hand, pointed to the many instances, even in the context of the AIDS crisis, in which the popular perception that a drug worked simply was not borne out in clinical trials.
In the end, the bureaucracy has supported, at least on its face, the parallel track concept. It is less willing to concede that parallel track and the early release of ddI are at all different. The experience of working ddI through the system, however, as discussed below, has reassured them to some degree.
The Research Community
As leader of the AIDS research community, Fauci committed the research agency with the largest research program and the greatest level of AIDS research funding to the parallel track concept. However, as noted earlier, his was a solitary decision within NIAID that did not involve much prior consultation—especially not with the principal investigators of the ACTG, the primary vehicle for academic AIDS drug trials. Although dependent on NIH for their funding, these investigators were known for their independence, a source of some frustration to Fauci on this and other matters.
Prior consultation may well have bogged down the process and resulted in a proposal so watered down as to be meaningless—as well as insufficient to meet the demands of the activist community. On the other hand, the lack of prior consultation also resulted in an initially tentative and later mixed reaction from the research community, which resigned itself to the inevitability of parallel track but over time became more vocal about its doubts.
Some in the medical community also raised objections to parallel track because drugs were being made available on the basis of very little data and might later prove to be unsafe. As Robert T. Schooley of Massachusetts General Hospital explained, potential toxicity “might be missed” in Phase I studies, a situation he said had occurred “with several AIDS drugs” (Zonana and Cimons, 1989).
The first full discussion of Fauci's proposal with principal investigators took place at a regularly scheduled meeting of the ACTG executive committee in July 1989. The principal investigators, who believed that people enrolled in clinical trials mainly to gain access to experimental drugs, challenged Fauci 's assumption that clinical trials could maintain participant levels even with parallel track. The investigators argued that an alternative source for experimental drugs would mean fewer enrollees in randomized trials, the backbone of the ACTG approach to research. Already under attack for low levels of enrollment in ACTG clinical trials, the researchers were concerned that parallel track would further exacerbate their problems.
The chairman of the ACTG executive committee, University of Washington researcher Larry Corey, gave a lukewarm, often rambling endorsement of parallel track at the FDA's Anti-Infectives Advisory Committee hearing. The bulk of his statement focused on the greater value of good clinical trials in improving access in the long run: “I believe in expanded access in selected areas. I guess as a researcher I believe more passionately that the goal of clinical research is to define if a drug is effective and how to use it for the practicing physician and that truly is increasing access. To turn off the clinical research program in any way, shape or form, I think, will end up being a detriment to the overall good rather than benefit. But I am a true believer that if designed correctly, that these programs could actually enhance each other” (FDA, 1989). In addition to such statements within the scientific community, researchers also began expressing their concerns about recruitment to the press. Newton E. Hyslop, Jr., a principal investigator for the TulaneLouisiana State University AIDS Clinical Trials Unit, told the Los Angeles Times that he was “concerned that the establishment of a parallel track process would affect the selection process for formal trials in a way that could bias the outcome because participants would no longer represent a statistically valid sample group” (Cimons, 1989).
Although much has been said about concerns within the research community regarding recruitment for ddI trials, others have suggested that the real problems may occur in recruiting participants for trials of different drugs. Clear distinctions are possible between parallel track and research eligibility for ddI alone; what happens, however, when trials for another drug seek participants and discover that many individuals in their potential recruitment pool are receiving a drug under the parallel track mechanism? Should or can they be forced to leave parallel track participation when a new clinical trial of a different drug, for which they meet the eligibility requirements, begins? Another concern among researchers and regulators—as well as among activists—is that there is no impartial monitoring mechanism within the context
of parallel track or early release of ddI. Without such a component, proponents and opponents of parallel track may seek to fit the data to their arguments if there are any questions at all about its interfering with ongoing research.
The AIDS Activists
The parallel track issue brought an unusual level of unanimity to the AIDS constituency. Although all elements in the community were supportive of expanded access to clinical trials, early release of drugs outside the research setting was an issue pushed initially by the ACT UP and Project Inform sector, which then brought along, in growing numbers, more of the mainstream groups. Indeed, parallel track was considered to be of such importance to the AIDS constituency that a wide range of organizations signed a consensus statement presented to the August FDA Anti-Infectives Advisory Committee meeting that became the basis for much of the debate at the meeting. The document defined parallel track as follows:
Parallel Track should encompass post-Phase I open-label treatment protocols for people unable to participate in controlled clinical trials for AIDS and HIV-related conditions. Drugs should be eligible for Parallel Track as soon as a tolerably safe dose range has been defined and preliminary evidence of efficacy has been obtained. 9
The document went on to offer the following eligibility criteria for early access:
People with a condition for which there is no standard treatment.
People who cannot tolerate the standard treatment for their condition.
People who are failing on standard treatment.
People who must stay on concomitant medications forbidden, but not expressly contraindicated, in trials of new experimental treatments.
People who live too far from the site of an appropriate controlled trial.
People who are too sick to participate in an appropriate controlled trial.10
Yet despite the apparent solidarity this consensus statement seemed to symbolize, disagreements continue to surface from time to time within the activist community. Some groups have adopted an essentially libertarian approach: that individuals have the right to make decisions about what drugs to take, even before clinical trials have been completed. They argue that even if people are eligible for trials, they should not be prevented from taking advantage of the parallel track.
Others are more cautious. Representative of that viewpoint is Neil Schram, chairman of the AIDS Task Force of the American Association of Physicians for Human Rights, the gay/lesbian physicians' group. Schram, in a Los Angeles Times op-ed piece asked: “Why shouldn't people who want the drug be allowed to have it? Shouldn't people with a life threatening illness be given any hope they want? ” He responded to his own question by saying, “There is no doubt that people who are sick and cannot tolerate AZT, or are getting sick in spite of AZT, need ddI as their only current hope. But for others, a proven alternative is available.” In more general terms, Schram made the argument common to more mainstream AIDS groups: “To completely do away with regulations that protect people from unproven drugs would lead to chaos and many unnecessary deaths. Even if the right to choose were granted, people would not have sufficient information to make an intelligent choice among untested drugs. So research and appropriate regulations must be protected ” (Schram, 1989).
Other Consumer Interests
To a large degree, the discussion around parallel track took place in a vacuum. Other consumer or patient interests were not well represented in the discussion, and it was assumed that the guidelines being developed for parallel track would be limited to AIDS drugs. But William Schultz, then with the Public Citizen Health Litigation Project, told the FDA's Anti-Infectives Advisory Committee, “I do not see how this can be formulated without talking about the impact on drugs for other diseases. Whatever program the FDA adopts, it is going to be argued that it ought to be applied to other diseases” (FDA, 1989).
Schultz, while supporting the parallel track concept, urged greater attention to both safety and efficacy issues. He warned that “not every drug is going to turn out to be sufficiently safe to be available at the beginning of Phase II or right after Phase I testing. I do not think it should be advertised as a program where all AIDS drugs are going to be available at the beginning of Phase II” (FDA, 1989). Schultz also said, “I do not think it can be emphasized enough that there has to be some evidence of efficacy before such drugs can be made available. One sort of test to apply to any program adopted would be to ask the question, would this program allow the drug laetrile to be made available during an investigational test? Laetrile, presumably, could pass the safety test. There is no evidence of efficacy and if the program would make laetrile available, then I would argue that there is a serious problem with it.” Schultz did not claim that the parallel track program as outlined would, indeed, have such a result (FDA, 1989).
Earlier, Schultz's boss, Sidney Wolfe, had also urged caution and warned the presidentially appointed National Committee to Review Current Procedures for Approval of New Drugs for Cancer and AIDS that parallel track could create “an extraordinary conflict between researchers and patients.” He expressed concern that parallel track could place clinical trials at risk. “Nothing will happen if science isn't applied,” Wolfe said. “The parallel track is fraught with that possibility” (Science, 1989).
EARLY RELEASE OF ddI
Although many of the players, particularly the activists, insist that the early release of ddI was not carried out through the parallel track mechanism, the two were closely linked in the minds of the regulators, the public, and the press. It is hard to believe that, in formulating the structure for early release of ddI, the arguments regarding parallel track were not part of the thinking of the responsible actors in the process and that there was not some awareness that the success or failure of the ddI early release would directly affect support for parallel track. The stakes were clearly higher than simple negotiation of a treatment IND for one drug.
After several years of experience with AZT as the only approved antiretroviral drug in the arsenal against AIDS, it was clear that something better was needed, ddI was the first new antiretroviral to clear Phase I studies successfully, and thus, in the spring of 1989, it was finally possible to talk of another potential therapy besides AZT. Once that was clear, according to the FDA, some form of early access or compassionate use proposal was immediately put on the table—and supported by the FDA and the drug's sponsor, Bristol-Myers.
The activist community seized on ddI after the June 1989 Fifth International AIDS Conference in Montreal, at which promising early data were presented. It appeared, as several people observed later, that ddI would be AZT “without tears,” that is, without some of the severe toxicities experienced by people on AZT. (It took lengthier trials for some of the toxicities associated with ddI to be revealed.) ddI's image was certainly enhanced by what some perceived as a ringing endorsement of the drug from National Cancer Institute Director Samuel Broder during a plenary speech in Montreal.
The activist community quickly began a push to gain early access to the drug on a broad scale, and in June, ACT UP/New York initiated discussions with Bristol-Myers and the FDA. As the sponsor of the drug, Bristol-Myers had no legal obligation to release it before the entire licensing process was completed (especially since it might not be
able to charge for the treatment). The company's participation in these conversations and its commitment to negotiate early release of the drug with FDA and the activists was a remarkable breakthrough.
Although some form of compassionate availability might have occurred in any event, Fauci's speech on parallel track and his immediate suggestion that ddI might be the first candidate for this new initiative increased public attention and bureaucratic pressure on both the FDA regulators and the drug company. From mid-July on, there were regular meetings of all the involved parties to discuss what form early release would take. The fundamental disagreement was on the extent to which early release would be offered. Initially, both Bristol-Myers and FDA advocated a fairly narrow release, whereas the activists argued for full individual choice: anyone who wanted to take an informed risk should have access to ddI. By the end of the negotiations, the parties had agreed on a middle ground. A meeting held in New York in late July was a major turning point, in large part because it was the first time all the players had been in one room together (rather than participating in two-way conversations that often became scrambled when related to third parties). FDA, Bristol-Myers, and NIH officials sat down with activists and community physicians to determine just what level of access should be permitted. They finally agreed that ddI should be made available to those with no other options. From then until the September release of the drug, the discussions focused on the definition of “no other options” and on the research protocols associated with early release.
This meeting was a dramatic departure from standard operating procedure for the FDA. Normally, such discussions were limited to the FDA and the drug's sponsor, and researchers were only occasionally involved. Time pressures were also very different in that normally the drug company set the pace for negotiations. In this instance, the players were under pressure from both Commissioner Young and the NIH. ACT UP also claimed it had a commitment from Bristol-Myers that the drug would be available by the end of August.
Above all, the meeting represented an unusual level of involvement by activists and nonscientist patient representatives in issues normally reserved to the scientific community. Many activists felt the New York meeting marked a breakthrough in relations with Ellen Cooper, who led the FDA team in the ddI negotiations. They had arranged for community physicians to attend to introduce a real-world perspective, and the activists felt that this contact with the dilemmas facing physicians contributed to Cooper's more flexible approach. Cooper, on the other hand, argues that the meeting was not as critical a juncture in her attitude as the activists think. She said she had always been open to input but that this meeting, in contrast to more confrontational ap-
proaches in the past from the activists, was a very constructive one. When she asked to be invited to the New York meeting, she said she felt the activists were shocked to find that a bureaucrat was willing to leave Washington.
The end result was early release of ddI a little later than the activists had hoped for but a little sooner than some of the regulators felt was wise. Data cannot be rushed, they argued, and a little more time would have avoided some of the protocol changes that had to be made after they were announced.
Commissioner Young does not agree with those who worked under him. He blames NIH for the delays. He says that the FDA's part—the treatment IND—was held up because it was hoped that the patient access and research arms could be announced simultaneously. By having the treatment IND ready, Young felt the FDA placed needed pressure on the NIH to resolve its end of the discussions sooner.
Throughout the early release process, the assistant secretary for health's office had minimal involvement regarding ddI and was only involved through NAPO in the parallel track discussion. Toward the end, however, the office did ensure that all relevant pieces of the release were coordinated and in place.
On September 28, 1989, the secretary of health and human services announced a three-part program for early release of ddI.
First, research would continue on the longer-term safety and efficacy of ddI through three Phase II clinical trials in the NIH's ACTG units among 2,600 patients: (1) comparing AZT and ddI among those with little or no exposure to AZT; (2) comparing AZT and ddI among those with more than a year's experience with AZT; and (3) comparing various doses of ddI among those who were intolerant of AZT.
Second, the Phase I studies were used as the basis for issuing a treatment IND for ddI, allowing those who were intolerant of AZT to receive ddI.
Third, an open safety protocol was permitted for those who had shown evidence of failing to respond to AZT.
The criterion for access to both the treatment IND and the open safety protocol was the inaccessibility of clinical trials: either because the individual did not meet the entry criteria or because there was no geographically accessible trial for the individual to join.
The details of these protocols had been the subject of substantial negotiation. Often, individual activists brought their personal experiences to the table—and were surprised to find them accommodated in the discussion. For many activists this was the first time they had been on the “inside,” negotiating with the decision makers, and the shift
in roles sometimes made them uncomfortable. In addition, those activists who remained on the outside soon questioned some of the protocols, as well as the activists who had participated in their design.
Questions also arose in short order from some in the traditional academic community. As mentioned above, it was a deliberate policy decision to issue all three parts of the early release program simultaneously because of the commitment to continuing the research process. Everyone involved in the early release decision was painfully aware of the criticisms of parallel track—that it had the potential for undermining clinical trials, thus preventing clear answers as to the usefulness of a therapy. But what had not been considered was that enrollment in the treatment IND and the open safety protocol—a relatively simple process of physicians contacting Bristol-Myers—would move much faster than enrollment of patients in the ACTG trials. The ACTG system was already under heavy criticism for its sluggishness in enrolling patients. By late November, when enrollment in ddI clinical trials seemed to be foundering while enrollment in the Bristol-Myers open safety program was booming, some scientists began to criticize the early release program openly. For example, Jerome Groopman of Boston's New England Deaconess Hospital told the New York Times, “People talked about and tried to reassure the academic community that, yes, parallel track will not dismantle our ability to do organized studies. But we have to face this head on. There really are conflicting issues here. If the philosophy is that anyone can decide at any point what drugs he or she wants to take, then you will not be able to do a clinical trial.” Donald Abrams of San Francisco General Hospital said in the same article, “Nobody ever said it was logical. It was a matter of acquiescing to political pressure. In our effort to get things to go a little faster, we can only hope that we are not slowed down” (Kolata, 1989b).
NIH officials were significantly less concerned: they felt it was too early to know the impact of early release. The activist community suggested that parallel track might become the whipping boy for larger recruitment problems within the ACTG system. Bristol-Myers, for its part, also was not concerned. They were keeping close tabs on applications, and their data showed that those receiving the drug through early access were following the established criterion; that is, they were not able to join trials for one reason or another.
At this writing, it is too soon to determine the impact of early release on clinical trials and on the trials' ultimate goal of learning whether ddI does, indeed, work. One thing, however, is clear: despite the claims of many of the players, including Young, Fauci, and the activists, that early release of ddI is not necessarily parallel track in its pure form, the success or failure of the ddI early release program will determine
the willingness of the FDA, the NIH, and drug companies to go down this path again.
Indeed, Ellen Cooper argues that if the ddI early release is not parallel track, she does not know what would be. The draft guidelines for parallel track now making their way through final bureaucratic review do not in any way create a new regulatory setup but instead use systems quite similar to those used for ddI. What makes the guidelines different—and perhaps therefore necessary from a policy standpoint —is that they bring together in a philosophical construct a commitment by the government to early release of promising AIDS therapies and make clear to all what methods will be available for accomplishing that end. Cooper admits that the FDA's commitment to early release, which she believes has always been present, has never before been clearly articulated, and these guidelines might solve that problem. But ultimately, she is correct. Because each drug will pose different issues, the exact nature of parallel track in the future will vary from drug to drug—and the ddI example will probably fit as closely as any successor drug might to a strict development structure.
One of the missing components of the ddI experiment with early access is a systematic review mechanism. With so much riding on the success of this experiment, it is unfortunate that no objective system of evaluation has been established. Ex post facto review is possible, but the fact remains that the need for ongoing evaluation should have been recognized by those responsible for developing the early release program, if only to prevent evaluation by press comment. In fact, there is very little in the way of hard data to support either side's claim regarding the impact of early release on clinical trial enrollment. The significance of parallel track or early release as a break with traditional practice and its potential importance to those with life-threatening illnesses call for systematic review and evaluation to supply the answer to this question.
The story for the moment must necessarily stop here, but it might be useful at this point to step back and assess some of the motivations and views of the key actors in this process. In a sense, all were thrust into new or dramatically altered roles. A consensus was reached through a confluence of self-interest, but some of that self-interest, while supportive of the immediate outcome, could well lead to future discord.
Bristol-Myers' support and role in early release of ddI is perhaps the
most interesting because in many respects company officials had the greatest capacity to insulate themselves from the political pressures being placed on them by the activist community. Instead, they embraced the efforts of the activists, even if they were not always in total agreement. In a process that depends heavily on the support of the drug sponsor, Bristol-Myers actively moved the discussions forward. In fact, all of the players in and observers of the ddI early release negotiations agree that Bristol-Myers believed in their drug, had actually sought out expedited processes within the FDA for review of AIDS drugs, and was willing, even before some of the publicity, to consider expedited release. Bristol-Myers was known as a cautious company; this history reassured some, especially in Congress, who might otherwise have been suspicious of a drug company's desire to expedite release of one of its products.
The activists perceived the company as acting sincerely during the negotiations. According to one company official, the firm understood the activists' message that people would be dying while waiting for approval of ddI. But seeking early release after only Phase I trials was a “quantum leap” from the past experience of the company (Paul Worrell, Bristol-Myers, personal communication, 1989). In addition, the scope of the early release undertaking may also have been beyond what Bristol-Myers initially anticipated. There is, apparently, some concern within the company about the costs associated with such a large program of distribution. The ability and willingness to set into motion the manufacture and distribution, at no cost, of an unapproved drug is a risk that probably only a larger company would be willing to take. A smaller company often waits until approval to gear up for manufacture and distribution—or to align itself with a larger company for this purpose. A small company would not have had the resources to undertake what the activist community wanted in terms of the scope of early access.
But the early release of ddI also served some of the corporate interests of the company. It certainly generated good will and positive media coverage for Bristol-Myers, a stark contrast to the feelings among AIDS activists toward Burroughs Wellcome, the manufacturer of AZT. Burroughs is under almost constant attack from the very same activists for its pricing structure. (Interestingly, no one who participated in the discussions with Bristol-Myers seems to have raised the issue of pricing postlicensure—and whether the cost of the early release program would be factored into that market price.)
There was also an important opportunity, through this program, for Bristol-Myers to educate the overwhelming majority of physicians treating the nation's AIDS patients about their drug, something normally forbidden before a drug is licensed. Early release resulted in a
great deal of free advertising and marketing, all with the FDA's blessing and encouragement. From the regulatory perspective there was also the belief, probably not unfounded, that if Bristol-Myers went along with a process that seemed important to the FDA, it would be hard for the FDA to turn down the firm's new drug application for ddI unless the data proved to be very negative. Judgment calls are always made at critical points in the licensing process; had Bristol-Myers not gone along with the FDA on early release, some of those calls might not have been favorable.
The Food and Drug Administration
There were two levels of participants at the FDA. Commissioner Young, who was described by one activist as “tired of being out-Fauci-ed ” in terms of favorable publicity among AIDS activists, wanted to use ddI as an example of how the FDA could make things work. He also saw the ddI case as an opportunity to vindicate his advocacy for the changes in the treatment IND regulations that allowed earlier release of drugs for life-threatening illnesses. Young believes that the ddI experience is a demonstration of the success of those changes, even though they were attacked by some activists and congressional leaders when they were announced, many of whom later supported early release of ddI. Under the regulations, the commissioner personally approves issuance of the treatment IND—the only FDA regulatory decision made in the commissioner's office. At a time when he was under increasing attack for his stewardship at the FDA, Young felt that his support of such early release mechanisms gave him a chance to exercise power that would generate good press.
Yet much greater regulatory power rests with Ellen Cooper as head of the antiviral division. She, too, saw the ddI experience as a vindication of her role and the positions she had taken—but in a very different way from Commissioner Young. She saw no need for the codification of the treatment IND regulations; she strongly believed that early release had always been an option for the agency and that codification created false expectations when no drugs were ready for treatment INDs. She saw parallel track in a similar vein. Accomplishing early release of ddI within existing constructs proved the flexibility of the system, and there was no need to codify the process. She felt strongly that there had been too little experience with early access after only one drug to try to write firm guidelines and that each drug would pose very different issues. Indeed, she felt the ddI experience was parallel track as it was meant to be structured.
The National Institutes of Health
Technically, the NIH role in the early release process was somewhat limited. NIH was responsible for negotiating and initiating the protocols for Phase II trials, which would have occurred even without early access to ddI.11 In fact, however, the whole process was set in motion by Fauci's proposal for parallel track and his immediate linking of it to ddI. In his effort to gain support within the AIDS activist community, as well as out of sincere support for the concept, Fauci began something of which the ultimate outcome was not at all clear to him—and whose far-reaching implications were probably not considered and thrashed out within NIH before the concept was introduced. This kind of process can be effective if one has faith that the events will play out as one would like them to. Given the number of competing players, however —and the fact that much of the ultimate authority for the success of parallel track rested with his competitors at the FDA—if Fauci 's initiation of the process was deliberate, it was an unusual display of trust because the entire exercise ran the risk of turning out very differently from what Fauci and the original proponents intended.
One of the most important roles the NIH could play in this process was to gain the support of the academic research community for early access, whether as parallel track or for ddI specifically. One of the recurrent criticisms from the other players in the early release effort was that they assumed that Fauci, as head of the NIH's AIDS research effort, had consulted with or would in some way limit criticism from the academic community. In fact, most NIH officials and the NIH principal AIDS investigators were in the dark about his proposal until it was made public. This lack of consultation and briefing raised the specter, which is still present, that the academic researchers would actively oppose early access or otherwise try to undermine its success.
The philosophical imperative behind the work of the activists was simple: to the extent possible, patient autonomy in the drug availability process should be respected. Government interference should be limited in these decisions, and the sooner the opportunity for informed choice was permitted, the better.
The push for early access to ddI was motivated by the belief that ddI was, indeed, going to prove successful as an antiretroviral therapy and also by a more personal imperative: many of the activists and their friends needed an alternative to AZT. The activists knew that the proc-
ess that had been agreed to for ddI release was going to be a prototype against which the ultimate parallel track initiative would be measured. It may not have been the systematic structure they envisioned for parallel track—and they felt very strongly that a systematic structure was needed—but it at least provided an opportunity to pilot-test some of the structure's components.
In a certain sense, the activists helped create something new by their participation in the early release negotiations. The system may always have been open to this kind of input or may always have had this level of flexibility, as government and company officials often claimed; but in the case of ddI the activists achieved an unprecedented level of involvement by consumer and patient advocates in the drug regulation process, normally a very tightly held series of negotiations, and forced real-world considerations into the decision making of the regulators and the sponsor.
Several fortunate political circumstances enhanced the activists' influence. There was a bidding war for community support between Young and Fauci that gave the activists an advantage, even if they were not conscious manipulators of that competition. In addition, ddI had gotten good press, which created public and media interest in how negotiations were faring. And above all, Bristol-Myers was a relatively open company.
From a policymaking perspective, some questions can certainly be raised. The personal imperative that motivated many of the activists involved in the process added urgency to the discussions. But some activists also recognized that there may have been a fuzzy line between good policy and what they as individuals needed; for example, were accommodations being made, one activist asked himself, because they met his personal medical needs or because they also represented appropriate modifications in the protocols?
Major precedent-setting changes were occurring in drug access policy. Both the ddI and parallel track decisions were made as though there was no carryover to diseases other than AIDS. The limited experience of the activists in broader health advocacy issues may have closed off opportunities for broader reform—or created problems for other disease groups. Even within the context of AIDS, decisions were made based on the immediate urgency of finding an alternative to AZT, while at the same time precedents were being set for future AIDS-related decisions.
It is not entirely clear whether the activists were aware of the situational nature of the alliances they were making. Bristol-Myers may appear to be aligned with their concerns at this stage of the drug development process, but the company's decision was based on sound business imperatives and not on moral factors alone. It is not incon-
ceivable that those same imperatives may drive a wedge between Bristol-Myers and the activists later in the process.
Similarly, the activists seemed dependent on the support of FDA Commissioner Frank Young, particularly to push the career officials to accept earlier access. Young's sudden departure from the scene startled them—and taught them what many advocacy groups and career bureaucrats with longer experience have known for some time: political appointees come and go, and their value is of limited duration. Educating and building alliances with career officials may be more painstaking and less public, but they can, over the long term, provide lasting support. Given the substantial discretion that career officials have at a regulatory agency such as the FDA, it is not at all clear what impact Young's departure and a long transition period at the agency will have on the influence of the activists on future drug policy decisions.
There is also a question of accountability with regard to the activists who participated in the early release negotiations. Although nominally affiliated with specific organizations, they were very much a part of the negotiating process as individuals who were respected for their expertise on the issues and for their ties to certain parts of the AIDS constituency. Through force of intelligence and personality they won their way into the decision-making process—a measure of their capabilities. But it left them open to second-guessing from other activists and to questions from players on the other side of the negotiating table as to how much political strength they were bringing to the process. In a sense, whatever naivete the activists brought to the negotiations regarding the drug industry and bureaucratic politics was matched by the other participants' assumption that the activists were responsible to an organizational structure similar to that of their own employers.
All of the key players in the decision-making process publicly express support for the final outcome—early access to ddI. The process by which this decision was constructed lends itself to several general observations.
Competition and closely controlled decision making may move an agenda in the short term, but they may undermine long-term support. The competition between Fauci and Young caused Fauci to seek relatively little counsel beyond his inner circle regarding parallel track for fear Young would jump ahead of him. Young, for his part, pushed the FDA bureaucracy to move as quickly as possible on the ddI treatment IND so he could prove that he could make parallel track work before there was even a parallel track structure. In a sense, the strength of support
from these two top health officials prevented the process from becoming bogged down in haggling over details. But that support also had its negative side: the lack of stability that accompanies dependence on a political appointee (Young) and the result of Fauci's lack of consultation with the academic community, whose long-term backing of parallel track remains very much in doubt. A process that allowed the academic community in particular to buy into the concept early on—to allow them, perhaps, the opportunity to be educated by the activists, as Fauci was—might have caused some frustration among the activists over delays, but it might also have assured a longer-term commitment, once the media and the politicians were no longer focusing on the subject.
A broader definition of decision makers produces better policy outcomes The level of involvement in the ddI process by community physicians and advocates was unprecedented in FDA history, and all those who were part of the process agreed that the final outcome was the better for it. In particular, the New York meeting in late July that brought the FDA together with community physicians has been cited as introducing a critical and welcome real-world element to the discussion. FDA officials needed to hear from peers and fellow physicians, not just activists, to be convinced finally of the need for this level of expanded access. Yet no permanent mechanism currently exists to assure this level of involvement for future AIDS drugs or for drugs for any other disease.
Ellen Cooper, who has often been perceived by many as an obstacle to just this kind of involvement, is one who believes that there is a need for all relevant participants to meet on a more formal basis. By participants, she means the drug's sponsor, researchers, and patient advocates. At some point, the legally empowered decision makers need to make their choices, but those choices will be better if they are informed by the input of all interested parties. Logically, she feels, the FDA should call this group together. Yet because, at least in the context of AIDS, FDA is often perceived as part of the problem, an independent group may need to facilitate the process.
Even if the proposed parallel track guidelines are adopted, they would institutionalize advocacy involvement in only part of the drug access process. Broader solutions need to be proposed and debated, outside the context of a specific decision on one drug. The advocacy and activist groups, in facing the immediacy of the AIDS crisis, often neglect the longer-term value of institutionalizing their gains. Pressure for change needs to come from the outside, because even though regulators have been supportive of much that has happened in this individual case, the tendency is to fall back on familiar patterns that seem safer, even if not always better.
Evaluation of new initiatives is too often an afterthought. In this case,
a fairly far-reaching change in standard procedures has been implemented without setting up evaluation mechanisms. The success or failure of the ddI experiment will affect the future of early access, but there is currently no independent means of assessing and auditing what is occurring. Advocates for the various points of view within the research, regulatory, pharmaceutical, and activist constituencies will be left to argue over whether problems arising during this experiment (such as difficulties in recruiting patients) are the result of the early access program or other causes. This new early access process is too critical a change to be left to this kind of debate. It may be too soon to draw conclusions, but it is not too soon to identify the data that should be collected and how they ought to be evaluated.
EDITOR'S NOTE: As of January 1991 clinical trial results suggest that ddI has promising retroviral activity and is generally well-tolerated.
1. Not all AIDS activists come from the gay community, just as AIDS is not a gay disease. But within the AIDS constituency, much of the pressure around early access to experimental drugs has, indeed, come from the gay community. More fundamental issues of access—to primary care and to clinical trials—are also part of the AIDS constituency 's agenda and have been pushed by the needs and advocacy of those minority and poverty community representatives active in the AIDS effort. These are antecedents, in a sense, to access to the experimental drugs that are the subject of this paper.
2. Among the individuals interviewed for this paper were James Allen, director of the National AIDS Program Office, and his deputy, Bruce Artem; Ellen Cooper, director of the Anti-Viral Drug Products Division of the FDA; Margaret Hamburg, then special assistant to Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, and now with the New York City Health Commission; Jay Lipner, a New York activist attorney; Tim Westmoreland, counsel to the House Subcommittee on Health and the Environment; Paul Worrell, of Bristol-Myers; and Frank Young, former commissioner of the FDA. The author, as an AIDS consultant working with several groups, including the National Gay and Lesbian Task Force and Gay Men's Health Crisis, also attended a number of the key meetings in question and had numerous conversations with participants in the ddI early release and parallel track process over the period under discussion in this paper.
3. The history of federal drug regulation and description of the regulatory process in this section draw heavily on an issue brief by the National Health Policy Forum (NHPF), “The Pharmaceutical Industry, ” and a talk by Peter Barton Hutt, at an NHPF panel on January 11, 1990 in Washington, D.C.
4. The AIDS crisis, however, brought a shift in thinking about the state role in drug regulation in this regard. Reflecting frustration with the slow pace
of federal drug approval for AIDS treatments, in 1987 the California legislature created its own version of the FDA in the hope that California-based pharmaceutical companies might move their drugs through an expedited state process.
5. The FDA had not permitted distribution of thalidomide in the United States, even without the additional authority and requirements granted with the Kefauver amendments.
6. Randomized trials are considered the “gold standard” for drug research. They involve assignment of patients to different parts of a study, to compare a new drug with a placebo (i.e., no treatment) or to compare the new drug with standard or current therapy. Patients are randomly assigned to a particular group, and researchers are not supposed to know to which study a patient is assigned.
7. This is not to say, however, that Fauci was caught by surprise at the attention his speech drew from the media. On the contrary, he personally called key AIDS reporters to make sure they would be covering the speech.
8. There was certainly some division within the AIDS activist community about Frank Young. New activists—such as those from ACT UP—had only relatively recent experience working with Young and saw him as trying sincerely to support their efforts and as responsive to the public demonstrations they undertook. Those with a longer history of AIDS lobbying were a bit more skeptical. They perceived Young as being supportive only as long as the interests of the AIDS community coincided with those of the pharmaceutical industry. And they had had some difficult experiences with Young in the early days of the AIDS epidemic around the licensing of the original antibody test for HIV.
9. “Guidelines for the Parallel Track Program for AIDS and HIV-Related Treatments,” memo to James Mason, M.D., Assistant Secretary for Health, August 17, 1989, from the AIDS Action Council, ACT UP/New York, ACT UP/San Francisco, AIDS Project Los Angeles, American Association of Physicians for Human Rights, American Civil Liberties Union AIDS Project, American Foundation for AIDS Research, Community Research Alliance, Gay Men's Health Crisis, Human Rights Campaign Fund, Lambda Legal Defense and Education Fund, Mobilization Against AIDS, National Association of People with AIDS, National Gay and Lesbian Task Force, National Gay Rights Advocates, Project Inform, and San Francisco AIDS Foundation.
11. The NIH had a role in research protocol design only because ddI Phase II trials are being done through the NIH's ACTG system. Had Bristol-Myers decided to do the trials privately, there might have been no NIH role.
Cimons, M. 1989. Coalition proposes criteria for AIDS drug trials. Los Angeles Times, August 17, 1989.
Eigo, J. 1989. Testimony before the Energy and Commerce Committee, Subcommittee on Health and the Environment, U.S. House of Representatives, July 19, 1989.
Food and Drug Administration Anti-Infectives Advisory Committee. 1989. Transcript of August 17, 1989 meeting.
Kolata, G. 1989a. AIDS studies head seeks wide access to drugs in tests. New York Times, June 26, 1989.
Kolata, G. 1989b. Innovative AIDS drug plan may be undermining testing. New York Times, November 21, 1989.
Mason, J. O. 1989. Testimony before the Energy and Commerce Committee, Subcommittee on Health and the Environment, U.S. House of Representatives, July 19, 1989.
Schram, N. R. 1989. Lives, research hinge on limiting new AIDS drug. Los Angeles Times, September 27, 1989.
Science. July 28, 1989. Quick release of AIDS drugs. Page 347.
Young, F. 1989. Testimony before the National Committee to Review Current Procedures for Approval of New Drugs for Cancer and AIDS, January 4, 1989.
Zonana, V. F., and M. Cimons. 1989. Ease AIDS drug rules, health chief urges. Los Angeles Times, June 24, 1989.
The decision by the FDA to make ddI available early in the drug evaluation process is unique in that the “user” community (patients and potential patients) played a decisive role in the outcome. For this discussion, the bureaucratic dispute about whether the decision was implemented by an existing mechanism that permitted release for “compassionate use” or whether it represented the de novo creation of a parallel clinical trial is irrelevant. The important point is that the FDA responded to patient-generated pressure for release of ddI with only meager safety data and before randomized controlled trials (RCTs) had been completed. The FDA action was strongly opposed by those committed to RCTs as the only scientific way to evaluate efficacy and safety in a chronic disease with a highly variable course and duration. Moreover, even many of those who had been reluctantly persuaded by the case for compassionate use because of the dire prognosis of AIDS were concerned about the precedent-setting consequences of this action would this step mean that the FDA would subsequently release other drugs before thorough testing and thereby expose patients to unwarranted risk?
The epidemiology of AIDS and the historical moment of its appearance combined to create a singular set of circumstances. The first cases were exclusively male homosexuals; thereafter the syndrome was successively recognized in other population groups (transfusion recipients, hemophiliacs, intravenous (IV) drug users and their heterosexual partners, children born to infected mothers). The most recent case data (January 1990) show that gays
Leon Eisenberg chairs the Department of Social Medicine at Harvard University.
are still overrepresented in the caseload; the Centers for Disease Control attributes 60 percent of the 121,000 U.S. AIDS cases to male homosexual or bisexual behavior.
What gave this situation particular relevance was that white middle-class gays had organized an effective political movement before the disease was first recognized. Consequently, they were in a position to speak out—and they had powerful reasons for doing so because the federal response to the epidemic was grossly inadequate. In contrast to self-declared, politically active gays, many of the other AIDS patient constituencies (e.g., IV drug addicts, bisexuals, and men who engage in homosexual behavior but do not identify themselves as gay) were not and are not represented in health care decisions. Almost all diseases are differentially distributed in populations. But there is no other instance in recent memory of an epidemic of an infectious disease that has been so narrowly focused on a self-conscious, politically experienced patient group, one that has spoken out forcefully in its own interests.
It is not that concern for patients has been absent from prior FDA deliberations. Indeed, concern for health is the raison d'être of the FDA. Nor have patients lacked advocates. To the contrary, physician specialists have spoken up on behalf of their patients. What is different in the case of ddI is that for the first time patients spoke for themselves in the decision-making process.
Traditionally in FDA deliberations the dialogue has been confined to civil servants, pharmacologists, clinicians, and pharmaceutical representatives. With ddI, patient spokespersons in effect forced their way to the negotiating table, where the authenticity of their experience gave them a special expertise. Permitting laypersons to participate in a technical decision, even when they are highly educated, raised questions in the regulatory community. Can dialogue be meaningful when the premises from which each group starts are not shared, and often not understood, by the others? Furthermore, the gay spokespersons had no mandate to represent other AIDS patients—for example, drug users and closet gays in the minority communities, whose lives are also at risk but who are largely excluded from access to standard treatments no less than to experimental ones.
If the decision to proceed with the parallel track can be hailed in one sense as a victory for a more democratic process, what are its potential costs? Three risks can be foreseen: (1) the parallel track may impede or even block completion of the RCT; (2) early access to ddI may expose more patients to the risk of severe toxicity; and (3) the precedent may return to haunt us if it is applied unwisely. Let us consider each in turn.
Although gay advocates do not speak with a single voice, many agree that RCTs are necessary to evaluate drug efficacy and safety. They share the concern of clinical investigators that the ability to obtain ddI outside of trials may markedly decrease enrollment in clinical trials. Why should a patient agree to a 50-50 chance of getting the highly touted new drug in the RCT
when he or she can be sure of receiving it through the parallel track? If enough RCT volunteers were to default, that would be a misfortune for all HIV-positive patients. Will that happen? Is it happening? As of the fall of 1990, there is no answer. Recruitment of participants for the ddI RCT has been slow; the same has been true, however, of earlier trials of other drugs.
What about undue toxicity? Recent reports indicate that the death rate among those in the parallel track is some 10 times higher than among those in the RCT. The explanation, however, is anything but certain. Criteria for admission into the RCT rule out patients with advanced disease, but such patients are allowed to receive the drug through the parallel track. The observed higher death rate in the parallel track may thus represent nothing more than the worse initial prognosis of the enrollees. The example of suramin, however, should remind policymakers that an RCT was necessary to discover that drug 's unacceptably high toxicity.
The death rate from pancreatitis has also been much higher among those in the parallel track and probably does represent ddI poisoning. Again, this may be a dose-related phenomenon or interaction with stage of disease. Even if ddI use has caused pancreatitis and subsequent deaths, the fundamental question remains: does the chance to reverse a fatal disease justify the risk? Some patient advocates contend that life with advanced AIDS has such little value that hastening death is no tragedy. Unfortunately, during the FDA discussions, there was no agreement to spell out in advance (1) what data would justify abandoning the parallel track and (2) what evidence from the parallel track could properly be admitted into the evaluation process. Lack of such agreement led to predictable responses to the first mortality data. Those who had been against the parallel track from the first contended that the high mortality warranted its discontinuation; gay activists argue that the hope for benefit justified the risk.
At present, whether ddI will prove to be a safe, effective drug is not clear. What is more important for public health in the long run is the way we evaluate the parallel track as a response to clamor for access to new drugs. Surely this mechanism should be neither adopted nor abandoned without a close reading of the ddI story. Did it allow very sick patients to obtain a useful agent months before its official release? Did it lead to avoidable deaths because toxicity exceeded utility? Could either outcome have been predicted from the data on hand when the decision was made?
Whether ddI leads to early salvage or undue toxicity, should patients be given access to an agent they want to use when the risk is considerable and efficacy is uncertain? The recent example of laetrile raises serious questions about the wisdom of such a strategy. The danger is less the confraternity of charlatans ready to peddle expensive nostrums to credulous patients than the risk that even nontoxic nostrums, through their illusory promise of ready cure at no discomfort, can lead to great harm by diverting patients from appropriate (but unpleasant) treatment protocols.
Peter F. Carpenter
This case study is a well-researched, balanced, unbiased review of the decision-making process that allowed release of the AIDS drug ddI for use prior to approval of the new drug application. The case is an excellent example of the difficulties involved in attempting to achieve an appropriate balance between the needs of individuals (people with AIDS), who desire access to an unproven therapy, and the needs of society, which must ensure proper testing of pharmaceutical products before they are made available for use outside of controlled clinical trials. There is no “right answer” to this dilemma: rather, the decision will reflect both the specific circumstances of the disease and of the unproven therapy. In the case of ddI, the expected outcome for patients who did not use the unproven therapy was sufficiently severe that none of the therapy's anticipated side effects would have been more serious than the outcome without therapy. In addition, more active involvement (than might have occurred even a few years ago) of patients in the decision-making process was a factor in shifting the balance toward making the drug available.
The case study also illustrates the importance of finding solutions within established rules, rather than forcing a change. Such an approach is not only face saving but, more importantly, avoids the often unexpected or unintended consequences of what was thought to be a simple change. However, the case
Peter F. Carpenter, a former pharmaceutical company (ALZA) and federal government (Office of Management and Budget) executive, is a visiting scholar at the Center for Biomedical Ethics at Stanford University.
does not carefully analyze the historical precedent of the availability of unapproved cancer drugs. These so-called Class C drugs and the controversy surrounding their use are, in many ways, quite similar to the ddI dilemma. It is unfortunate that we have failed to apply directly both the process and the substance of the lessons learned in developing the Class C system to ddI and other AIDS drugs.
Some key elements in the process examined by Jeffrey Levi were the identification of concerned constituencies, clarity regarding the issue, and the role of certain key players in making things happen. A previously weak constituency, the gay community, consolidated recent increases in its political strength and was highly motivated by the impact of AIDS on its members to represent itself forcefully. The general public and the general scientific community, however, have not been represented in the process to date, which may have important long-term implications. The willingness of individuals in positions of formal power to engage in dialogue with the advocates of change was a procedural step that was critical in this decision-making process. In addition, the recognition that existing rules could be interpreted to accommodate the desired outcome was essential in reaching a timely decision. These procedural steps can and should be generalized to other decisions involving biomedical innovations.
The two most important shortcomings of the ddI decision-making process were its failure to include representatives from all constituencies and to make provisions for evaluation. For example, the broader scientific community was not formally involved in the ddI decision and it is now beginning to question the apparent decision to move away from the “gold standard” of randomized controlled clinical trials, particularly in light of reports of adverse reactions among individuals receiving the drug outside of clinical trials. The FDA has published a formal proposal to revise its rules to permit the general application of the procedure used in the ddI case. The review process that the proposal will undergo could bring to light important constituencies that did not participate in the original ddI decision. Second, it is necessary and important to have a prospective evaluation mechanism to assess the soundness of the decision based on the actual results versus the expected results.
The following potential process guidelines emerge from this case:
Participants in the decision-making process should clearly identify the issue being addressed or, if they are starting from substantially different definitions, develop a convergent definition.
It is necessary to separate the specific from the general—realizing that the solution will later need to be generalized.
It is important to encourage dialogue among the broadest range of constituencies.
There is value in finding ways to use interpretations of existing policy as the solution rather than developing totally new policies.