Clinical Research and Drug Regulation
In perhaps no other area has the impact of AIDS been more clear than in the identification, clinical testing, and regulation of new drugs. The effects have been so extensive and the pace of change so rapid that it is difficult to describe all the important events within the confines of a single chapter. Yet it is still too soon to identify changes that are merely transient phenomena versus those that will endure (Henninger, 1990). Although there are early indications that the changes witnessed in AIDS research may be affecting pharmaceutical development for other diseases, it remains unclear whether AIDS will appreciably alter the development of drugs in general, especially for conditions that are less lethal or less common.
Perhaps the most profound change wrought by AIDS in the area of drug development is simply the dramatic increase in public awareness, especially regarding the very existence, as well as the structure and purpose, of clinical trials. Terms such as "randomized," "placebo," and ''double-blind," although perhaps not household words, are nonetheless regularly used in the mass media. Debate about the ethics and scientific validity of clinical trials occurs not only among physicians, statisticians, and ethicists, but also among patients, activists, and politicians. AIDS has opened the arena of clinical investigation—the organization, ethics, and politics of research—to media and public scrutiny.
The impact of the HIV/AIDS epidemic on clinical research poses a series of important questions with broad implications for the practice of science and medicine. For example, to what extent have changes been
driven by scientific, ethical, economic, or political considerations? Has AIDS merely accelerated changes that were already occurring in cancer research, or do they represent a clear departure from the practices of the last two decades? When changes have occurred, have they been defined narrowly for AIDS or more broadly constructed? How can change be monitored in the future, and can the likely long-term critical impact of such change be predicted?
This chapter begins with a brief historical perspective on clinical research and drug development prior to the advent of the HIV/AIDS epidemic. It is important that the interplay of politics, science, and ethics in this area be recognized as an ongoing dynamic during the twentieth century; it has been shaped by AIDS but it was not created by it. The chapter then describes drug development from 1981 through late 1991—the age of AIDS—and analyzes the impact of social institutions and events on the process and, conversely, the impact of the process on such institutions as the Food and Drug Administration and the National Institutes of Health. The development of social and political activism is examined as a major force for generating change. Finally, the chapter summarizes how key areas in clinical investigation and drug regulation have been shaped by the HIV/AIDS epidemic.
Several developments of the post-World War II era establish a baseline for evaluating changes driven by the HIV/AIDS epidemic. Those developments relate to four major historical shifts: (1) the development of randomized clinical trials as orthodox research methodology, (2) changes in federal drug regulation, (3) the evolution of protections for human research subjects and the growing recognition of ethical dilemmas inherent in research, and (4) the rise of patient advocacy and the changing dynamics of patient-physician relationships. In the postwar era, and especially since 1960, considerable change has taken place in each of these areas. The AIDS epidemic has had powerful effects on clinical research and the culture of science.
Randomized Clinical Trials
Although experimental approaches to clinical medicine date to antiquity, it was only in the 1940s and 1950s that contemporary approaches to the evaluation of medical treatments and technologies were fully articulated and began to be put into practice. Randomized clinical trials—experiments in which human subjects are randomly assigned to experimental and nonexperimental (control) groups for purposes of comparison—offer researchers
considerable advantages in evaluating new drugs and other medical technologies. First, variability among subjects will, on average, be equally distributed, thereby reducing the potential for selection bias. Second, because such studies are generally conducted on a double-blind basis (i.e., neither the researcher nor subject knows which group a given subject is in), randomized clinical trials offer a mechanism for reducing investigator bias (W.A. Silverman, 1985). Third, randomized clinical trials permit the use of sophisticated statistical tests of significance in the comparison of treatments. According to their proponents, prospective, double-blind trials offer the potential to place clinical medicine, at long last, on a truly scientific base (Marks, 1990). As biostatisticians have recognized, the development of randomized clinical trials was perhaps the most important methodologic advance associated with the scientific basis of therapeutics (Zelen, 1990).
Despite these advantages, randomized clinical trials did not immediately become the basic rule of clinical investigation. Such studies were expensive to conduct, administratively complex, and required new skills that many investigators lacked. Furthermore, investigators expressed a critical awareness of the difficult ethical dilemmas posed by randomized trials. As statistician A. Bradford Hill (1951:279) explained:
The first step in such a trial is to decide precisely what it hopes to prove, and secondly to consider whether these aims can be ethically fulfilled. It need hardly be said that the latter consideration is paramount and must never, on any scientific grounds whatsoever, be lost sight of. If a treatment cannot ethically be withheld then clearly no controlled trial can be instituted.
Researchers, attuned to this question, attempted to specify those conditions under which a particular treatment could be withheld for the purposes of a randomized trial (Chalmers, Block, and Lee, 1972). According to most researchers, randomization could be justified only in instances in which there was genuine ignorance concerning the advantages and disadvantages of an experimental drug; only then could a researcher claim "therapeutic indifference" (Hill, 1963:1047):
It must be possible ethically to give every patient admitted to a trial any of the treatments involved. The doctor accepts, in other words, that he really has no knowledge at all that one treatment will be better or worse, safer or more dangerous, than another. … If the doctor does not believe that, if he thinks even in the absence of any evidence that for the patient's benefit he ought to give one treatment rather than another, then that patient should not be admitted to the trial. Only if, in his state of ignorance, he believes the treatment given to be a matter of indifference can he accept a random distribution of patients to the different groups. … By certain omissions from a trial we may limit the generality of the answer given by it, but on
ethical grounds that, in my experience, must be accepted [emphasis in original].
In order to retain this state of ignorance as the process of collecting data began, trials were designed to isolate researchers from findings. This was accomplished by establishing an advisory panel, usually including expert biostatisticians, to monitor the data as they accrued. Such panels were empowered to halt a trial at the earliest time that a scientific conclusion, usually defined as statistically significant, could be reached. In practice, despite this mechanism, such determinations about when a trial should be ended proved to be complex (Chalmers, Block, and Lee, 1972).
In addition to the justification of therapeutic indifference, proponents of randomized trials also cited the fact that experimental preparations were sometimes scarce and that randomization offered a means of allocating drugs in a socially constructive manner (Rutstein, 1970). Such was the case with the Medical Research Council's streptomycin trials of 1946 (Hill, 1963:1043):
When, in 1946, the Medical Research Council's Streptomycin in Tuberculosis Trials Committee set out to investigate the effect of that drug in pulmonary tuberculosis it was faced with no serious ethical problem. The antibiotic had been discovered two years previously … the published clinical results were distinctly encouraging though not conclusive. Yet overriding all this evidence in favour of the drug was the fact that at that time exceedingly little of it was available in Great Britain. … Except for that situation it would certainly on ethical grounds have been impossible to withhold the drug from desperately ill patients. With that situation, however, it would, the Committee believed, have been unethical not to have seized the opportunity to design a strictly controlled trial [emphasis in original].
Despite these early attempts to define clear ethical criteria for randomization, withholding unproven but potentially beneficial treatments has remained controversial. So-called treatment indifference is far more simple to establish in theory than in practice (Johnson, Lilford, and Brazier, 1991). Drugs are subjected to controlled trials because there is already some evidence that they may be effective. This hope, which may or may not be borne out through further investigation, nevertheless shapes the research environment, especially in instances in which there is little to offer patients with serious, life-threatening disease. Jonas Salk, for example, opposed double-blind trials of the vaccine he developed against polio. In Salk's mind, the research he had already conducted demonstrated the effectiveness of the vaccine; to conduct a full-fledged randomized trial, he concluded, would be merely a "fetish of orthodoxy" and would lead to more new and unnecessary cases of paralytic polio among the group receiving the placebo. Other researchers countered that without a full randomized clinical trial the
effectiveness and safety of the vaccine would never have scientific and medical credibility, a requirement if doctors, parents, and children were to accept immunization (Carter, 1966; Brandt, 1978a). In this respect, it was argued that only through rigorous trials would it be possible to avoid "therapeutic anarchy," in which clinicians, based on experience and anecdote, would do as they saw fit. Randomized trials would add scientific evidence and legitimacy to clinical judgment (Feinstein, 1967). Proponents also pointed out that only randomization could adequately identify not only the beneficial impacts of experimental drugs but also their significant dangers and side effects (Chalmers, Block, and Lee, 1972).
The difficult ethical considerations that attend randomized clinical trials remain a critical part of the debate concerning "therapeutic research," a term used to denote research conducted on subjects who, it is hoped, will benefit from the experimental preparation. When are randomized trials justified? How should they be organized? Is it ethical, in effect, to toss a coin to assign patients to a particular treatment? How can the conflicts between the desire to further scientific knowledge and the commitment to act in the interest of individual patients be resolved? Does the very act of randomization violate basic norms of the doctor-patient relationship (Fried, 1974; Schaefer, 1982)?
In randomized clinical trials, conflicts inevitably arise between the group and the individual, between the desire for scientific advancement and an individual patient's welfare. Attempts have been made to moderate such conflicts, but they can never be absolutely and categorically resolved. All these issues were debated prior to the AIDS epidemic, but the epidemic gave them a new immediacy. What would be the nature of clinical research in the midst of the epidemic (Eckholm, 1986)? With the onset of the epidemic and the establishment of a research program, the techniques of drug evaluation and experimentation and the ethics of randomized clinical trials came under intense scrutiny.
The Food and Drug Administration and the Politics of Drug Regulation
Since the early twentieth century, federal regulation of food and drugs had been largely reactive to scandals and tragedies in which consumers of unsafe products were harmed. Upton Sinclair's The Jungle (1906) helped to spur Congress to pass the first major piece of protective legislation in 1906, which prohibited false and deceptive labeling of food products and medications (Pure Food and Drug Act, P.L. 59-384). Prior to that time, a vigorous market had existed for nostrums and patent medicines, many of which made sweeping claims for miracle cures. Nevertheless, the act had no provisions requiring evidence that a drug be safe or effective. The
impact of the act was undermined by a Supreme Court ruling in 1912 requiring the Food and Drug Administration (FDA) to demonstrate not only that a particular claim was false, but also that it had been made with the intent to deceive. Promotional claims for useless and dangerous products, not surprisingly, continued to flourish (Quirk, 1980; Okun, 1986; Young, 1990).
In 1938 Congress again took action, this time enacting the Food, Drug, and Cosmetic Act (P.L. 75-717), which required that drugs be established as safe through "adequate tests" prior to marketing. Unfortunately, a tragedy helped to break the legislative deadlock that preceded passage of this act. Elixir sulfanilamide, a new sulfa drug marketed for children, was suspended in a solvent that could lead to immediate death; by the time the drug was withdrawn, more than 100 deaths had occurred. With the new legislation, the FDA was mandated to review safety data on drugs and remove dangerous products from the market.
Despite growing methodologic sophistication and rigor in the design and execution of drug trials during the 1950s and 1960s, it soon became clear that the public was still not always protected from pharmaceuticals that were either inadequately tested or produced. As late as the early 1960s, randomized clinical trials continued to be the exception rather than the rule in testing new therapeutic interventions.
Spurred by another series of scandals, Congress in 1962 enacted the Kefauver-Harris amendments to the Food, Drug, and Cosmetic Act of 1938 (Harris, 1964; Quirk, 1980). The notorious "Cutter incident" of 1956, in which 11 children died as a result of being inoculated with improperly produced Salk polio vaccine, was a powerful indicator of the need for better governmental regulation of biologic and pharmaceutical products. This tragedy was soon followed by the thalidomide disaster; children born to mothers who had taken this sedative had severe congenital malformations (Insight Team, 1979). Although thalidomide had not been marketed in the United States, it had been made available in loosely conducted premarketing studies. Both incidents pointed to weaknesses in the federal regulatory process. The Kefauver-Harris amendments to the Food, Drug, and Cosmetic Act required for the first time that before FDA approval was granted, a company had to categorically demonstrate through animal and human studies that a drug was safe and efficacious. (Prior to 1962 federal law had required that drugs be safe, but no specification regarding efficacy was mandated.) The amendments were a major impetus for the conduct of randomized clinical trials, which soon became the only accepted criterion for evaluating new drugs.
By 1970, with the growth in influence of the National Institutes of Health and the rise of biostatistics as a distinct discipline (Marks, 1990), the nature and methods of drug evaluation had achieved a form of scientific and
bureaucratic orthodoxy. Moreover, as medical costs rose precipitously during these years, interest in comparing treatments, not just on the basis of efficacy, but also on the basis of costs, was heightened and spurred more trials. A report by the Office of Technology Assessment (1983) concluded that randomized clinical trials offered a particularly effective methodology for evaluating a full range of medical interventions and technologies. Furthermore, the report concluded, such trials should become the basis for evaluating standards of medical practice, as well as for public policy.
The regulatory ethos established in the wake of the Kefauver-Harris amendments emphasized caution. The repercussions of approving a dangerous drug were typically perceived as being greater than those that would attend restricting the marketing of a potentially beneficial drug. Although some critics have recently called this view of regulation "paternalistic" (Delaney, 1989a,b), it reflected a growing recognition that the evaluation of pharmaceuticals was a complex scientific, political, and economic process. Individual patients and individual physicians, it was argued, could not make decisions about safety and effectiveness in a clear or objective manner. Any substantive evaluation required the aggregation and evaluation of large data sets. Thus, the government should mediate by establishing clear criteria, mandating trials, and evaluating industry claims before allowing drugs to be marketed. Rather than being seen as paternalistic, this approach was more often viewed as a significant role of the government in a consumer-oriented, government-regulated economy. In addition, as skepticism about medical technologies and interventions grew during the 1970s and 1980s, the FDA was often criticized for not being aggressive enough in its regulatory mission (Mintz, 1967; Silverman and Lee, 1974; Wolfe and Coley, 1980; U.S. Department of Health and Human Services, Advisory Committee on the FDA, 1991).
The Kefauver-Harris amendments shifted the regulatory burden from premarket notification of a new drug to premarket approval. The legislation required "substantial evidence" of efficacy, including "adequate and well-controlled investigations, by experts qualified by scientific training and experience to evaluate the effectiveness of the drug involved." This change obviously required the development of detailed criteria and bureaucratic mechanisms to evaluate the safety and efficacy of drugs. Under the regulations established to meet the requirements of the amendments, companies or other researchers seeking to have a product approved must first submit an investigational new drug application. The application had to show that initial screening and animal testing for toxicity had been completed and offer a thorough justification and plan for testing human subjects.
Generally, randomized clinical trials are divided into three phases: phase I trials typically assess data on safety; phase II trials evaluate efficacy; and
phase III trials compare the prospective drug with standard therapy, usually on a randomized, double-blind basis. This process can take considerable time, and a variety of approaches to streamlining the procedures have been proposed. In the view of the pharmaceutical industry, which typically bears the costs of clinical trials, the Kefauver-Harris amendments raised the cost of researching, developing, and marketing new products.
In the face of the AIDS epidemic, FDA's policies of cautious protection came under fire. Given attempts to modify regulatory practices, it will be critical in the future to monitor the potential effects of changes on new drug development, both for HIV and other diseases. How, in the context of the AIDS epidemic, has the FDA balanced its duties to ensure safe and effective drugs against its duty to make potentially beneficial agents available on a timely basis? (Edgar and Rothman, 1991). And how are changes in the regulatory environment likely to affect research and development in the pharmaceutical industry?
Protecting Human Research Subjects
Just as changes in the regulation of new drugs were spurred by a series of scandals, so, too, were attempts to protect human subjects driven by reports of serious abuses. Efforts to improve the regulatory capacities of the FDA were followed by a heightened public concern about the ethics of human subject research. Widespread reports of experiments in which research subjects were unaware of their involvement led to calls for more vigorous protections (Gray, 1975; Barber, 1976; Katz, 1984). Henry Beecher's (1966) review of unethical experimental protocols in the New England Journal of Medicine (see also Rothman, 1987b) was soon followed by exposes of the Public Health Service's Tuskegee syphilis experiment, in which some 400 African American sharecroppers were denied treatment in a study of the natural history of the disease (Brandt, 1978b; Jones, 1981; Thomas and Quinn, 1991).
As a result of these reports, Congress in 1974 established the National Commission for the Protection of Human Subjects (Rothman, 1991; Jonsen, 1984). The commission devised general ethical principles for the conduct of human experimentation, which it issued in its 1978 "Belmont report" (National Commission for the Protection of Human Subjects, 1978). Three principles were to form the basis for the evaluation of experimental protocols and provide a guide for obtaining informed consent, establishing favorable risk-benefit ratios, and equitably selecting test subjects: respect for persons, beneficence, and justice. A central tenet of the Belmont report was the need to be rigorous in differentiating experimentation from therapy. The commission argued that therapeutic research had the potential of subjecting patients to risky and dangerous experimentation offered in the hope
of therapeutic benefits. The heavy emphasis implicit in the work of the commission, as well as its formal title, was the need to protect individuals from potentially dangerous research—that is, that government must set limits on researchers and their prospective subjects. In this context, taking part in an experiment was typically viewed as a sacrifice that citizens, in the name of the public good, might periodically be called on to perform. Most commentators emphasized, however, that all too often the burden of medical progress fell inadequately on vulnerable populations—the uneducated and minorities (Katz and Capron, 1975). The commission reiterated its support for clinical investigation and scientific research, but it sought to establish clear principles on which to evaluate research on human subjects and specific mechanisms to protect human subjects from possible harm.
The findings of the commission were supported by the federal government in the subsequent requirements for institutional review boards to evaluate research protocols in advance of their implementation. Any institution that received federal funds was required by law to establish such a board to review all proposed research on human subjects and to assess the procedures for obtaining each subject's informed consent. In fact, most research institutes, regardless of funding sources, did establish institutional review boards. The effect of the boards was to create a somewhat more restrictive environment for clinical investigation by establishing clear requirements for consultation and oversight. The era of autonomous research had come to an end.
The HIV/AIDS epidemic suddenly and dramatically fractured the risk-aversive ethic of human experimentation, raising again fundamental questions regarding the nature of clinical investigation and research ethics (Levine, 1988). The epidemic created a constituency that would find the protections of the past too restrictive. It created a constituency of individuals eager to take risks with unproven therapies. In the midst of the AIDS epidemic, when access to clinical trials has become a bitterly contested question, it is worth remembering that only a short time ago the focus of discussion was the protection of research subjects from potentially dangerous experimental protocols. What the regulatory process had failed to recognize was that in certain specific situations individuals might be eager to have access to experimental drugs even if their safety and efficacy had yet to be proven by scientific criteria. The AIDS epidemic made clear that risk can only be defined in a very specific personal and social context. In this respect, an appropriate margin of safety for any set of clinical trials cannot be uniformly set.
Patient Advocacy and Activism
The critiques of medical experimentation articulated in the 1970s were not new, but they were being expressed with considerable vigor. Efforts to protect human research subjects were part of a larger social movement that reflected a shift in the medical balance of power. Many have called AIDS activism unique—and certainly in some ways it has been unprecedented—but it can be understood in the context of changes in medical relationships and authority that have characterized American medicine since the late 1960s (Jonsen, 1991; Rothman, 1991).
Although there is no adequate history of the patient-doctor relationship, it is widely recognized that fundamental cultural changes in the nature of this relationship have occurred since the mid-1960s. Since that time, there have been several changes that reflect, if not characterize, a shift away from "medical paternalism": a growing skepticism by patients of medical and scientific authority; a deeper commitment to disclosure and consent by physicians; the legal codification of informed consent, as well as clearer promulgation of patients' "rights"; and the establishment of institutional mechanisms, such as internal review boards to protect human subjects. All these developments reflected significant change in the general practice of medicine and the nature of clinical investigation (Katz, 1984; Faden and Beauchamp, 1986).
The rise of AIDS activism can be understood only in the context of the critical shifts in the medical culture just noted and the gay rights movement of the 1970s and 1980s (D'Emilio, 1983). The development of organized activist groups provided an institutional vehicle for articulating critiques of traditional research approaches and federal regulatory mechanisms (see Chapter 6). Those critiques reflected the powerful conflicts in values and priorities revealed by the epidemic, as well as a more general erosion of medical and scientific authority. This social process, as noted, was already under way before the onset of the epidemic; the sacrosanct world of scientific and medical investigation had been opened to public scrutiny, regulation, and criticism. Research, in this new context, would fundamentally require active participation and negotiation between researchers and subjects.
The AIDS epidemic would severely strain virtually every assumption of what had come—over a 20-year period—to be the guiding, orthodox assumptions regarding clinical research and the regulation of new drugs. Every aspect of the process by which new pharmacologic agents were identified, evaluated, regulated, and allocated would be tested by the exigencies of epidemic disease. Questions basic to the epistemologic foundations of biomedicine—questions of verifiability, reproductibility, proof, variability, safety, and efficacy—would all be subject to debate and reevaluation. In
this respect, the epidemic has already had a profound social impact on science and medicine.
THE FIRST DECADE OF AIDS
The recognition of the first case of AIDS in the summer of 1981 was an abrupt event, but a basic understanding of its cause, epidemiology, and natural history evolved over a period of 5 years. During that time, despite great limitations of knowledge, many treatments were pursued and some therapeutic successes claimed.
Kaposi's sarcoma and Pneumocystis carinii pneumonia (PCP) were the two diseases that heralded the arrival of the epidemic. Both diseases had been seen before (in elderly and cancer patients, respectively), and drugs already licensed were commonly used to treat the disorders, generally with good results. Indeed, it was the sudden marked increase in requests for a drug long used to treat PCP, pentamidine, that provided one of the first pieces of evidence that AIDS was an epidemic disease.
Frustration that PCP and other complications in AIDS patients were not amenable to conventional therapies led to a host of alternative approaches, including vitamins, holistic therapies, and imported, non-FDA-approved drugs. At the same time, the failures led scientists to consider the infectious and neoplastic aftermath of AIDS's destruction of the immune system as essentially insurmountable and to concentrate on the central immune defect underlying the disease as the proper focus for drug research. Heroic interventions, such as bone marrow transplants in twins (one infected with HIV and one not) and aggressive chemotherapy for Kaposi's sarcoma, were attempted. As patients continued to die despite these measures, AIDS increasingly came to be viewed as an untreatable disease, and medical and lay literature emphasized the need to respect the rights and needs of individuals with a fatal disease of short duration, including their right to refuse intubation and resuscitation, their need for companionship and personal assistance, and the need to create low-cost, compassionate, out-of-hospital care.
The year 1985 was a landmark year in the epidemic. First, it was revealed that movie star Rock Hudson had AIDS and that he had traveled to France to obtain HPA-23, an antiviral agent that French researchers had hailed at a press conference as curing several patients with the disease. Patients in the United States, desperate for therapy, questioned why a potentially useful drug could not be brought into this country. Affected individuals became increasingly aware of the often painfully slow process of drug evaluation and licensing regulated by the FDA. When additional experience with HPA-23 did not confirm early promising results, however, the wisdom of traditional methods of drug testing seemed to be substantiated.
Second, suramin, an agent that had been licensed for over 50 years and
that had been used worldwide to treat sleeping sickness, was identified as having strong in vivo anti-HIV activity and was quickly put into a small, phase I trial in several institutions. Surprisingly, the drug produced severe and apparently new side effects in many recipients and led to at least two deaths. The experience with suramin again appeared to validate the preeminence of orthodox clinical trials.
The Emergence of AZT
In the winter of 1985-1986, preliminary studies of a previously discarded anticancer agent, zidovudine (AZT), were begun by Drs. Samuel Broder and Robert Yarchoan at the National Cancer Institute (Wastila and Lasagna, 1990). The work of Broder and Yarchoan was carried out under the drug development program initiated by the National Cancer Institute (NCI) under the authority provided by the 1971 National Cancer Act. AZT appeared to increase patient well-being and possibly to slow progression of the disease, although it was not obviously curative. Because of acceptable toxicity, a phase II clinical trial was mounted by Burroughs-Wellcome, Ltd., in the late winter of 1986. Originally planned to last for 27 weeks of observation per patient, the trial was halted prematurely in September 1986 because of significantly greater survival rates in those receiving AZT in comparison with those receiving a placebo. At the time the trial was ended, most patients had been observed for only a short period of time. With this dramatic finding, the world of clinical research was turned upside down.
Virtually overnight, AIDS patients and others demanded that AZT be made immediately available. Burroughs-Wellcome established a compassionate-plea mechanism for drug distribution, modeled on the compassionate, investigational new drug (IND) mechanisms developed by NCI in the early years of its drug development program (see below). Physicians were required to document that patients had AIDS as evidenced by PCP or advanced AIDS-related complex (ARC), the characteristics of the trial participants. Generally used only for small numbers of patients with life-threatening conditions, the compassionate IND mechanism was a cumbersome, ad hoc process requiring active participation by individual physicians to obtain a drug and document its use. It had never before been applied in the wide manner an epidemic required. Although immediate problems of access were thus solved, it was clear that this mechanism would be quickly overwhelmed. Faced with urgent demands, the FDA hastened AZT through the regulatory process by easing requirements for animal and preclinical data. On January 16, 1987, the FDA convened a meeting of its Anti-Infective Advisory Committee to consider FDA approval. The advisory committee voted 10 to 1 for licensing AZT despite the expedited procedures (Nussbaum, 1990).
Dissent in the research community regarding this decision was muted but real. Concerns were raised that despite specific labeling that identified the drug as useful only in patients who had AIDS as evidenced by PCP or severe ARC, even asymptomatic patients might demand and physicians acquiesce to treatment with AZT, thereby risking major adverse effects in patients who might expect at least several healthy years if untreated. In contrast, others believed that all patients with AIDS, not only those who had PCP, were equally in danger of death and should receive the drug and that the labeling was thus too narrow. In an unprecedented program, Burroughs-Wellcome announced plans to collect postlicensing data on 20,000 patients receiving AZT to lessen the possible danger of widespread use of a drug on the basis of minimal data on toxicity and efficacy.
The Rise of Advocacy
As AZT quickly made its way from test tube to pharmacy in 1987, two other major developments occurred: AIDS activists, most notably represented by the AIDS Coalition to Unleash Power (ACT-UP), became increasingly visible, and the primary federal mechanism for the conduct of clinical trials, the AIDS Clinical Trials Group of the National Institute of Allergy and Infectious Diseases, expanded rapidly. Advocacy by grass-roots organizations in the gay community had been strong almost from the beginning of the epidemic. Initially focusing on issues of discrimination and access to health care, such groups became increasingly involved with issues of experimental treatments.
Advocacy groups made the FDA their prime focus, and from 1987 to 1989 they steadily and forcefully challenged the FDA to widen and speed the availability of drugs for AIDS. In addition, the Reagan administration, with its general philosophy of deregulation, became a sympathetic (if ironic) partner in the push toward greater consumer choice in AIDS drugs. Together, patient activism, a political climate favoring deregulation, and the exigency of an epidemic disease drove the most dramatic changes in drug regulation since the Kefauver-Harris legislation of 1962.
Initially, the changes were largely modifications of existing programs. As noted above, prior to 1987 physicians treating patients with life-threatening disease for which they had exhausted all licensed therapeutic options, could (on a case-by-case basis) obtain therapeutic agents that were still under investigation. Several variations on the basic process evolved, but terminology and procedures were ambiguous, and use of the various processes was sporadic and not well known in the medical community. Several attempts had been made to codify the ''compassionate use of drugs," but regulation by the FDA only occurred in 1987 under pressure from AIDS activists.
The 1987 IND regulations require that the disease for which the drug is to be used be "severe or immediately life-threatening," although specific diseases meeting this criterion are not specified. They also require that there is no licensed therapeutic option for the disease, that clinical trials or studies be ongoing or completed, and that the sponsor is in fact seeking marketing approval. In most instances, companies can charge for the drug. The intent of the 1987 regulations was to retain close control of treatment IND use but to enable it to be granted in a uniform way and earlier in the drug-approval process, as early as the completion of promising phase I studies.
Another landmark event occurred in 1988 when the Presidential Committee on Regulatory Relief, chaired by (then) Vice President George Bush, approved importation of limited supplies of foreign drugs for personal use. The FDA also developed the "expedited review procedure for drugs to treat life-threatening and severely debilitating diseases," through which drug sponsors could trade early involvement of the FDA in the drug development process for the promise of more efficient approval later. The FDA also demonstrated increasing willingness to provide treatment IND distribution. The 1987 regulations were viewed by the FDA as variously defined for only a few drugs and unlikely to be highly attractive to drug sponsors because of the cost of early FDA involvement. Others have noted that the regulations seemed to change the definition of "severe" in such a way that a relatively large number of drugs might in fact be included.
Despite these changes, which permitted patients a greater degree of personal choice in deciding which drugs to take, activists charged that the FDA interpreted the 1987 regulations too conservatively, only approving drugs for which phase II trials were completed or nearly completed, as a bridge between trials and FDA approval. Activists pushed for the creation of a novel mechanism to widen patient access, a concept that came to be called the "parallel track."
The parallel track concept was loosely modeled on an approach to clinical trials common in Europe to obtain definitive information about treatment efficay. As ultimately implemented in the United States, however, the parallel track became primarily a drug distribution mechanism that foreshortens phase II and virtually eliminates phase III of clinical trials. Such an approach is scientifically valid only if joined to intensive postmarketing surveillance. Although the parallel track was described as an alternative phase II model, permitting it to fit within the regulatory framework of the standard FDA procedures, it was in reality an ad hoc concession to public pressure that demanded wide distribution of clinically unproven new drugs. By contrast, the compassionate IND offered limited availability of drugs that had demonstrated clinical efficacy, but were not yet fully approved. AIDS activists argued that as soon as drugs had been shown to be safe and
at all effective in phase I trials, they should be made available to persons who could not enter phase II trials because of ineligibility or geographic barriers. The parallel track, then, would be a large-scale release of an unapproved drug in lieu of the physician-initiated, case-by-case treatment IND request. It was designed to broaden dramatically the availability of experimental drugs at the earliest possible time. New York's ACT-UP, which had become increasingly knowledgeable and sophisticated about drug approval, did the most to popularize this concept.
The first public discussion of the parallel track concept occurred in April 1988 under the leadership of the AIDS Clinical Trials Group. Dr. Anthony Fauci, associate director of the National Institute of Allergy and Infectious Diseases and AIDS coordinator for NIH, ultimately became a defender of the process, which he argued could be accomplished without sacrificing the integrity of concurrent formal clinical trials. Others, including the FDA's then Chief of Antiviral Drug Products, Dr. Ellen Cooper, contended that the 1987 treatment IND regulations were themselves sufficiently flexible to permit very early access and that no new mechanism was needed.
As the parallel track concept became widely discussed in the summer of 1989, enthusiasm was mounting for dideoxyinosine (ddI), a drug similar in composition and mechanism of action to AZT. Dideoxyinosine had shown dramatic in vitro inhibition of HIV replication and, in phase I studies, acceptable toxicity and some evidence of possible efficacy. Phase II trials of the drug were being quickly designed by the sponsor, Bristol-Myers Squibb, and activists pushed for wide access to the drug for those who were ineligible for trials and for concurrent formal trials. Thus, before the parallel track was fully developed as a concept, the White House decided to authorize its implementation, without the usual normal consultation with the scientific community. Because the decision to implement a parallel track circumvented normal procedures, both within the federal government and the larger scientific community, it emerged without the essential component of postmarketing scientific review. Indeed, it is questionable that such a review could be either designed or implemented for the HIV/AIDS patient group due to problems such as compliance with treatment protocols and extensive self-medication. A program was set up by the FDA that, for all intents and purposes, represented a parallel track. Perhaps because it was not officially viewed as such, or because of the great pressure to begin, no evaluation component was included in the program. Moreover, the formal clinical trial of the drug had, in retrospect, especially stringent criteria for entry: a large percentage of people with severe HIV disease were excluded and thus had no alternative for getting ddI other than enrolling in the expanded access program. Finally, initiation of expanded access actually preceded the beginning of the phase II trials, which resulted in a situation in
which, within months, thousands of patients had enrolled in an expanded access program, but the number of participants in the formal trial was significantly less than initial projections.
Many observers have cautioned that the ddI experience may be unique and cannot be used to determine the likely success of the parallel track mechanism. Indeed, debate continues as to what the exact purpose of the parallel track should be. Some argue that through such an approach valuable data could be obtained concerning safety and efficacy, data of sufficient quality even for the purposes of drug licensing. Others have argued that such data would be flawed and that expending the resources required to collect and analyze the data would be counterproductive to the central goal of increasing access.
Although the intention of the parallel track concept was that drugs would be supplied free to participants, no consideration was given to how the costs of laboratory monitoring and physician visits would be handled. Thus, those costs, traditionally covered by sponsors of research in formal clinical trials, have fallen on the patients themselves and, in some cases, third-party payers. This situation has led to the virtual exclusion of poor people from the expanded access program, despite the fact that one of the goals of the parallel track concept was to increase access. Lack of access for poor people has played an important part in recent demands by activists groups for very early licensing of drugs—immediately after phase I trials if any efficacy is demonstrated.
In July 1991, an advisory committee of the FDA recommended the approval of ddI (Kolata, 1991), and the FDA commissioner completed the final steps for approval. Thus, in the space of a very few years, radical changes in drug regulation have occurred, and for the first time, they have been due not to public reaction to a fatal drug toxicity, but to strong consumer activism by a group of affected patients and their advocates.
The Integration of Activists in the Clinical Trials Process
As activists achieved success in altering the regulation of drugs from 1987 to 1989, they began to turn their attention to the actual design and conduct of clinical trials, criticizing what they perceived as the failure of the AIDS Clinical Trials Group to test and deliver effective therapies rapidly. Although the group did not control many of the clinical trials being conducted in the United States, it had a critical mass of investigators, and its budget had grown considerably. Thus, it offered an identifiable and powerful target against which activists could organize their efforts to reform clinical trials. Over the next 3 years, activism exerted an unprecedented influence over the clinical trials process. The fifth annual International AIDS Conference, which met in Montreal in June 1989, proved to be
pivotal in this regard. At the time, most activism was still being directed at the FDA and the regulatory process, but activists were also openly critical of the AIDS Clinical Trials Group, which they accused of having produced no tangible results despite an investment of hundreds of millions of dollars. In Montreal, ACT-UP New York circulated a highly specific critique of the group's programs and protocols and called for sweeping changes in trial design (ACT-UP, 1989).
Investigators and statisticians initially offered predictable defenses of the traditional clinical trials process, but shortly after the Montreal meeting, as attention was focused on the surprising adoption of the parallel track concept by Dr. Anthony Fauci, an equally surprising event occurred. Dr. Susan Ellenberg, chief of the Biostatistics Branch of National Institute of Allergy and Infectious Diseases (NIAID) told activists and others that she found real merit in proposals for alternative designs (Green et al., 1990). She organized an NIH-FDA conference on clinical trial design in November 1989 and subsequently formed a statistical working group within the AIDS Clinical Trials Group as a forum in which novel trial designs could be debated and developed by statisticians, clinical investigators, activists, and persons with AIDS. This working group, which continues to meet, provided the first inroad for activists into the mainstream process of the AIDS Clinical Trials Group.
In September 1989, NIAID had awarded a large contract for statistical support to the Statistical and Data Analysis Center at the Harvard School of Public Health. This group, which had been providing limited statistical consultation to NIAID during 1989, began to provide increasing expertise in the design and analysis of trials. Senior statisticians at the center brought their significant experience in clinical trials in cancer to AIDS and formed a community advisory board to keep local community groups aware of their activities and to explain statistical issues in trials (Valdiserri, Tama, and Ho, 1988).
Later, the NIAID Division of AIDS and the leadership of the AIDS Clinical Trials Group (at least as a group) also became more amenable to inclusion of activists after several came uninvited to a meeting of the group in November 1989. Whether this cooperation represented the activists' banging down the door or clinical researchers' opening it is unclear. (To date, few of the principal individuals involved have written about or publicly discussed their recollection of these events.) Several individual trial sites (called AIDS clinical trial units) formed local community advisory boards to provide two-way communication of issues and concerns. Increasing interaction occurred among members of the AIDS Division staff and activists. A series of discussions and planning meetings ensued, and by July of 1990 Dr. Anthony Fauci announced that activists would have representation on all committees and in all activities of the AIDS Clinical Trials
Group (ACTG). The Community Constituency Group (CCG) was established as a formal committee of the ACTG.
After several years of growth, the CCG now consists of 24 members with diverse representation from the AIDS community. The CCG has come to play an active role in the ACTG: two positions on the ACTG Executive Committee are allotted to the CCG, and CCG members serve on each of the core committees. The CCG also works closely with the individual AIDS Clinical Trials Units (ACTUs), each of which is required to establish a local Community Advisory Board (CAB). Communication between the staff at the ACTU, CCG, and CAB has ensured broader input and enables the trial sites to effectively address patients' concerns, and other pertinent issues, such as recruitment, compliance, and retention.
It is perhaps too early to evaluate the full impact of these developments on the conduct of clinical trials. The process has certainly changed, but it is still unclear how decisions concerning which trials to mount and how to conduct and interpret them will be made. Clearly, the entire process will be open and highly visible. It is likely that studies will take longer to design because of the number of people who are providing input, but it is also more likely that the final trial designs will be more agreeable to patients and activists, which might lead to faster identification of participants and improved retention of them in studies (Arras, 1990).
Activists are also turning their attention back to the FDA. As discussed above, activists argue that expanded access, no matter how designed, will never achieve widespread distribution of drugs because of the cost of patient monitoring. Thus, they argue that the earliest possible licensing should be the goal. To that end, San Francisco's Community Consortium of Bay Area HIV Care Providers used a little-known legal mechanism of a citizen petition of the FDA to seek the release of ddI (and ddC) on the basis of short-term effects on laboratory markers (Kolata, 1990).
CURRENT ISSUES AND PRACTICES
Drugs and Alternative Therapies
Drugs used in the treatment of AIDS and related disorders have come from a variety of sources. Encouraged by the efficacy of AZT, the National Cancer Institute launched an ambitious program of in vivo drug screening in 1986, as part of which it has evaluated hundreds of drugs through the same simple assay that first identified AZT as a candidate for treatment. Antiviral drugs, such as ddI and ddC, have resulted from this process. As more became known about the early pathogenesis of HIV disease and its effects on the immune system, the first of the so-called designer AIDS drugs, soluble CD4, was put into trial.
The relatively large number of antiviral agents that have been shown to have activity in the test tube was not anticipated even a few years ago, and their existence has raised new issues regarding priorities in drug testing. At present, simultaneous trials of several agents are ongoing. Whether it is most efficient to test drugs in this manner or whether it would be better to rank drugs and test them in sequence (i.e., starting a new trial only when a previous one has enough subjects) is a matter of ongoing but ill-defined debate. Arguments on the basis of personal choice, which have been predominant in the epidemic (Delaney, 1989a,b), would seem to support the current system of multiple trials. Some have argued that federally sponsored trials might be conducted sequentially, but not the myriad of trials sponsored by pharmaceutical companies and others whose research and development information is proprietary. For a brief time, the AIDS Clinical Trials Group considered requiring that pharmaceutical companies virtually hand over promising drugs they wanted put into the group's trials, but in 1990 the pendulum seemed to swing back to individual companies' dictating many of the details of drug design and monitoring. Thus, it is likely that multiple experimental studies will continue to be conducted concurrently.
Even more interesting is the issue of how resources should be divided between testing candidate antiviral drugs and evaluating drugs used to treat the myriad of complications associated with HIV disease. Critics of the use of resources for the investigation of such supportive therapies have drawn an analogy with the polio epidemic of the 1950s, likening testing of drugs for opportunistic infections to investing large resources in the development of a better iron lung. They argue that the emphasis should be on the underlying cause of AIDS. This viewpoint has been criticized by some clinical investigators, who point out that prophylaxis of PCP has been perhaps more influential in extending survival than the use of AZT. The argument in favor of focusing on causes has also been heavily criticized by activist groups, who interpret the approach as "writing off" the hundreds of thousands of patients with severely impaired immune systems who are perhaps unlikely to benefit from antiviral therapy, however effective it may be, but who still might benefit at least modestly from more effective therapies for opportunistic infections and tumors. Debate on this issue became so intense that congressional hearings were held on the subject, and budgets for opportunistic infection therapies were increased (U.S. Congress, 1991).
Although the debate has often been deemed a political one—between scientists with conflicting research agendas or between future patients and current ones—there remains fundamental scientific uncertainty about the most efficient approach to prolonging the life of AIDS patients. It is likely that interest and resources will continue to shift between the antiviral approach and the supportive (opportunistic infections) approach as new information
is obtained concerning the pathogenesis of the disease and as new drugs are identified.
Along with numerous antivirals that emerge from the National Cancer Institute, there have been agents whose proponents, often patients or their advocates, cite dramatic if anecdotal case histories to support their claims of efficacy. So-called alternative therapies include not only drugs identified in the laboratory, but also traditional (folk) remedies used by ethnic minorities and nonmedicinal therapies, such as acupuncture, visualization, meditation, and special diets (Hand, 1989).
The use of alternative therapies is widespread among persons with HIV infection, especially gay men; less is known about the therapeutic choices of ethnic minorities. Alternative therapies were initially prohibited for patients in federal trials, but it quickly became clear that many patients continued to use them without reporting such use (Arras, 1990). Concerned that the possible effects of such unknown therapies might adversely affect interpretation of the trial design and aware that some patients were refusing to enter or stay in trials because of these prohibitions, clinical researchers gradually came to tolerate their use.
Nonmedicinal healing regimens and approaches in which empowerment of patients is encouraged ("survivors" rather than "victims"), already used by cancer patients prior to the AIDS epidemic, have increased in popularity during the AIDS epidemic for all life-threatening diseases. Although the degree of activism that has arisen in response to AIDS has yet to be seen for cancer, there are indications that it may occur for breast cancer, a disease that, like AIDS, can be associated with a long period of clinical well-being after initial diagnosis. Breast cancer patients are beginning to move beyond individual empowerment to group empowerment as a means of fighting the disease. Breast cancer advocates are demanding increased research funding, better dissemination of information concerning treatment options, and a more patient-centered approach to decision making (Gross, 1991). Although it did not rise directly from AIDS activism, it is probable that some of the inspiration for the new activist approach to breast cancer came from media coverage of AIDS activism, as well as from the feminist and natural childbirth movements.
Who Performs Clinical Trials
Clinical trials have been traditionally performed almost exclusively in tertiary care (specialized teaching or research) hospitals and clinics by academic physicians. Virtually all federally sponsored research and most of the research sponsored by the pharmaceutical industry has been conducted in this way. The advantages of such a centralized approach have included the interest and expertise of physicians at these sites, many of whom are
conducting basic research in the disease; the ability to develop dedicated facilities that are highly efficient for the conduct of such research (clinical research centers); the availability of specialized laboratory testing and nursing personnel trained in the requirements of research; and the availability of people to constitute an institutional review board to provide approval and oversight of the study and ensure proper protection of subjects.
The first significant changes in the traditional approach to clinical trials resulted from the large expansion in clinical research funded through the National Cancer Institute. The increasing numbers of young oncologists trained in the cancer research centers saturated the positions available in cancer centers and resulted in significant numbers of fully trained clinical oncologists establishing practices in smaller cities and towns, distant from cancer research centers. In response to this development, NCI established the Community Clinical Oncology Program (CCOP), which enabled community-based oncologists to be tethered to cancer centers and to use state-of-the-art cancer drugs in research protocols that were fully integrated into center-based clinical trials. CCOP has been extremely successful, but critical to that success are the well-established links with cancer centers and the available pool of oncologists fully trained in clinical research methods.
AIDS has strained the traditional system for clinical trials in several ways. First, unlike the situation for other fatal diseases, in which primary care physicians did not feel competent to treat unusual and severe diseases (like cancer), many community physicians with practices comprising mostly gay men acquired extensive experience in the management of HIV disease. Those physicians were often unwilling to transfer even part of the primary care of their patients to clinical investigators. Patients, too, wanted to remain with their physicians, many of whom were more tolerant than the medical establishment of alternative therapies. In addition, many gay patients had a deep distrust of a medical system that they saw at best as unwelcoming and often openly condemning of their life-styles.
As previously noted, access to investigational therapy has increasingly been viewed as a right rather than a burden (Annas, 1988; Bayer, 1990). Even a large and multicentered system may be "unfair"—in not meeting the needs of infected persons if there are geographical and financial barriers to access. Thus, many AIDS patients and their advocates have argued that studies could be done more efficiently and equitably and with increased patient compliance if community physicians and their offices participated in the conduct of clinical trials (Merton, 1990). An early example of such a community network was the Community Consortium of Bay Area HIV Care Providers, headed by Donald Abrams, a San Francisco physician. This group performed a study of the efficacy of three different doses of aerosolized pentamidine in the prevention of PCP, a trial that ultimately provided the basis for the licensing of aerosolized pentamidine for this purpose by the FDA (Abrams, 1990).
Both the American Foundation for AIDS Research and the National Institute of Allergy and Infectious Diseases have established a formal network of community-based trials (and some sites receive funding from both sources). In addition to sponsorship of clinical trials, the network's program will include establishment of an observational database to collect information gathered during the routine care of patients by their primary physicians. It is hoped that the database will provide information that will be useful in studying the natural history of HIV disease, in assessing the use of particular therapies by people with HIV disease, and perhaps, in projecting those patients' resource needs. Such databases have been developed to a lesser extent for other diseases and have been suggested as a means of creating an accurate historical group and thereby lessening the need for randomized trials.
It will be important to monitor the types of community studies done and their success in improving patient participation and retention and to delineate any drawbacks to the approach, such as decreased ability to monitor or respond to drug toxicities. It will also be important to determine whether a national system can retain the flexibility of a more grass-roots structure while providing the capability to mount large trials.
Almost from the beginning of the AIDS epidemic, patient advocacy groups took on the role of medical advisers through the creation of patient information newsletters (Steinbrook and Lo, 1990; Bishop, 1991). In San Francisco, AIDS Treatment News began appearing in May 1986 and grew from a biweekly column in a local gay newspaper to a newsletter with a circulation of about 5,000 (James, 1989). Also in 1986 in New York City, the Gay Men's Health Crisis began to publish Treatment Issues, an update on research in progress, complete with extensive references to the medical literature. The American Foundation for AIDS Research publishes the AIDS/HIV Experimental Treatment Directory . Buyers' clubs were also formed to import and distribute drugs that were not approved for use in the United States or approved drugs that were available more cheaply in other countries. Such groups were instrumental in educating patients about therapeutic options and creating in many areas a climate of debate on AIDS care and research. These developments reflected a new "consultant" ethos among potential and actual research subjects.
Such advocacy took on a whole new dimension, however, when one such group, San Francisco's Project Inform, under the direction of Martin Delaney, carried on its own study of compound Q, a cucumber derivative that had been used as an abortifacient in China and was reported to have significant in vitro activity against HIV. This study was conducted without oversight by FDA or an institutional review board, which sparked criticism from a number of lawyers and bioethicists. Project Inform argued that this was an ethical trial because patients were already taking the drug on their own, and the trial afforded at least some oversight by a participating physician.
When it was learned that two patients in the study had become comatose and died shortly after taking the drug, the FDA stepped in and halted the trial. Interestingly, Delaney was later permitted to reinstitute the trial after filing a treatment IND application and forming an institutional review board for oversight. This decision was especially noteworthy in light of the heavy hand of the FDA in past dealings with even minor clinical trial infractions: ''wrongdoing" investigators were often barred for life from conducting further drug trials. It remains to be seen whether the creation of community research consortiums will decrease the interest of advocacy groups in mounting their own trials and increase the level of cooperation and compliance among patients/subjects.
Activists have also become researchers in other less dramatic ways. As noted above, activist representatives now sit on many major scientific committees, including the Executive Committee of the AIDS Clinical Trials Group and several AIDS clinical trial units have community advisory boards. The mission of the latter is currently evolving, but they will likely at least informally review studies to be done at each site. Activists also routinely publish articles on AIDS therapies and have suggested novel approaches to trial design. Indeed, they have been more visible in strategic planning concerning trial design and conduct than the AIDS Clinical Trials Group or other academically based organizations.
The Design of Clinical Trials and Access to Them
AIDS has deepened and broadened critiques of clinical trials as the "gold standard" approach to the testing of new drugs. Such criticism certainly predated AIDS, but it focused largely on the ethics of trials (Beecher, 1966; Rothman, 1987a,b); criticism of AIDS-related trial design also includes methodologic issues.
Traditionally, clinical trials have been designed to ensure that answers are arrived at as quickly as possible and with no associated uncertainties as to whether the drug itself or confounding factors produced the observed effect. Thus, subjects in trials have no serious diseases other than the one under investigation and are not taking other medications. Definitive events are chosen for measurement of drug effect, such as survival, objective clinical improvement (such as disappearance or shrinkage of tumors), or in some cases, the return of an abnormal laboratory or clinical measurement to normal (Weiss and Mazade, 1990). When such laboratory measurements are chosen, they must have been previously shown to correlate to survival or clinical outcomes. For example, since it has been shown that lowering of blood pressure by drugs leads to a decrease in strokes and deaths, new drugs are licensed if they can be shown to lower blood pressure; it is not required that subjects be followed until the occurrence of stroke or death.
The smallest number of patients necessary to detect a predetermined effect of a drug are studied because, it is argued, it would be unethical to subject more people to experimental therapies than those needed to demonstrate an effect. To detect possible side effects of drugs, physical examinations and laboratory testing are generally performed frequently and thoroughly.
AIDS, however, is a complex disease. Its very name, acquired immune deficiency syndrome, reveals a key feature that makes traditional trial design problematic: HIV's destruction of the immune system manifests itself clinically through a host of secondary diseases, such as PCP, cerebral toxoplasmosis, atypical mycobacterial disease, and unusual tumors. Over 30 such diseases have been identified, and the majority of patients develop three or more of them during their lifetimes. Thus, it is not possible to find a population of HIV-infected patients in the late stages of disease who could be considered free of other serious disease. Patients entering a trial typically have several opportunistic infections, each of which may be treated by one or more drugs, each drug with its own side effects and interactions with other drugs. It is not uncommon for patients with AIDS to take ten or more such medications. Monitoring of toxicity due to an experimental drug is therefore difficult because observed side effects may be due to medications other than the experimental agent or because of interactions with that agent.
The use of definitive outcomes in AIDS research to determine drug efficacy is becoming increasingly problematic. AIDS is a chronic disease: using survival as an outcome, especially in trials of asymptomatic persons, would entail many years of treatment and observation before the value of a therapy could be ascertained. Choosing the development of CDC-defined AIDS as the outcome shortens the duration of a trial somewhat, but that outcome may be delayed by the simultaneous use of prophylactic drugs for opportunistic infections. For this reason, some people have proposed using surrogate markers, that is, laboratory indicators of disease progression that could give early evidence of drug effect. Of the markers proposed, the measurement of the absolute number of CD4 cells is currently considered the most promising. Indeed, at this time, many AIDS investigators favor basing trials and early drug licensing on CD4 counts in lieu of clinical outcomes. However, other researchers have pointed out that even in diseases for which some drugs have been shown to have the same effect on surrogate markers and clinical outcomes, other drugs may have a beneficial effect on markers but not show clinical benefit, or may even cause clinical harm. Conversely, some drugs may have clinical benefit without having any effect on surrogate markers. Two recent trials not involving AIDS illustrate this issue. In one study, interferon was of dramatic benefit in the treatment of chronic granulomatous disease (CGD), an inherited disorder leading to life-threatening infections. However, all laboratory tests of neutrophil
function (the affected cell in CGD) failed to show any improvement. In contrast, a recent large clinical trial (Cardiac Arrhythmia Suppression Trial) looking at prevention of sudden death due to cardiac arrhythmias after myocardial infarction (heart attack) was stopped when two agents (ecainide and flecanide) were found to suppress cardiac arrhythmias, but paradoxically, to lead to sudden death (Ruskin, 1989).
In addition to emphasizing the need to include in trials people whose clinical status is representative of all those with the disease, AIDS trials have also raised the issue of access to trials for groups that historically have been excluded (passively or actively) from participation, such as intravenous drug users, minorities, and women of childbearing age. Initially, AIDS trials almost exclusively enrolled white, middle-class gay men. These participants were, in general, better educated than average, usually well informed about treatment options, and able to enter and remain in trials. As AIDS increased among intravenous drug users and their sexual partners, advocacy groups began to argue not only for increased enrollment of these groups in trials, but also for the provision of social support services to enable participation, such as translators, transportation vouchers, and child care services. Advocates have argued that inclusion of a demographically representative group of patients is warranted not only for reasons of fairness (recognizing that for a disease with few options for licensed drug treatment, experimental trials are themselves rare), but also because drugs may have different efficacy or toxicity on demographically different patient groups.
Thus, arguments of justice and science have been put forth to support increased access. Trial sponsors and community leaders now publicly support this approach, but there is as yet little indication that groups other than gay men are entering trials in appreciable numbers. Indeed, the pragmatic obstacles to such participation are considerable. In a recent report, the National Commission on AIDS (1991:104) noted:
the difficulties in deciding who pays for health care associated with research vividly reveals the discontinuity in federal health programs. NIH-based researchers claim no jurisdiction over health care provision, and the federal agencies responsible for the reimbursement and delivery of health care, HCFA and HRSA, are not designing program strategies that would take these research-related issues in account.
NIH, charged with conducting drug trials, has no mandate or funding support to provide health care services.
There has been little research to explore the desire or capabilities of underrepresented groups to participate in trials. Some have argued that at least among African Americans, distrust of the medical system is so high and the collective memory of clinical trial abuses (as in the Tuskegee syphilis study) so strong that many people are actively opposed to participation
(Brandt, 1978a; Thomas and Quinn, 1991). Particular questions have been raised about children and women.
The ethical issues inherent in including children in trials, long debated for other diseases, are even more complex in relation to HIV disease. In general, candidate AIDS drugs are moving into trials on the basis of relatively less preclinical and toxicologic data, which increases traditional concerns over possible adverse outcomes for children in trials. Most children with HIV disease have mothers who themselves have the disease, many of whom are unable to care for their children. For this reason, and in some cases because of ongoing maternal drug use, many HIV-infected children are in foster care (see Chapter 8). It has proven difficult to obtain consent for enrolling such children in trials (Martin and Sacks, 1990). Yet the natural history of HIV disease in children differs appreciably from that in adults, which is an argument for earlier and fuller trials involving children. Lacking such trials, therapy for children has lagged behind that for adults, and guidelines for translating the results of trials to clinical practice have been limited to adults, leaving practicing pediatricians uncertain as to how to treat children.
In the case of AZT, use in symptomatic children was not approved until April 1990—fully 3 years after the drug was licensed for adult use. Moreover, licensing was granted by virtue of a regulatory waiver because the limited trials conducted among children did not satisfy minimal FDA requirements. As a result of that experience, pediatric trials of ddI were started very soon after adult trials were initiated and were designed to satisfy FDA licensing requirements. In addition, it has recently been suggested that because opportunistic infections occur in children too infrequently to warrant separate pediatric trials, children should instead be included in adult trials, with appropriate modification of dosage (Nolan, 1990). Thus, it is likely that, at least for AIDS, and possibly for other life-threatening diseases, there will be a shift toward earlier, larger trials among affected children.
The issue of inclusion of women of childbearing age in trials, especially pregnant women, has been the subject of much debate (Levine, 1990). Early trials often explicitly excluded pregnant women. At a minimum, all trials have required a negative pregnancy test at entry and have stipulated that birth control must be practiced. Whether because of these restrictions or because women are less aware of trials and more likely to face numerous
barriers to participation, as of the end of 1991 fewer than 7 percent of participants in the AIDS Clinical Trials Group trials were women.
AIDS trials are not unique in having few women participants (Cotton, 1990a,b). Indeed, the norm in clinical trials for a wide variety of illnesses has been explicitly to exclude women. Despite NIH guidelines requiring that any research protocol not including women offer a justification for the exclusion, a study released by the U.S. General Accounting Office (1990) demonstrated that few federally sponsored studies complied with those guidelines. That report, widely covered by the media, resulted in congressional inquiry into the lack of inclusion of women in trials and the creation of a new post of assistant director of the NIH for women's research. Recent NIH directives require, not merely encourage, the participation of women (and minorities) in federally funded research (National Institutes of Health, 1990; Dresser, 1992). AIDS was not the cause of the debate concerning this issue, but it certainly sharpened that debate because AIDS is now a leading cause of death in women of childbearing age.
Inclusion of pregnant women in AIDS trials has created an arena for debate concerning larger social issues. Because most cases of pediatric AIDS are a result of transmission of the virus from mother to fetus, proposals for treatment of the mother in an attempt to prevent such transmission have been offered. At the time of this writing, a study designed to treat pregnant women with AZT to prevent such transmission has recently begun. The protocol for this study (the so-called AIDS Clinical Trials Group protocol 076) has been the subject of more debate and controversy than that for any other AIDS trial; that 2-year debate illustrates the changing climate regarding treating pregnant women with experimental therapies.
The trial was originally designed to treat newborn infants with either AZT or a placebo (chosen at random) for 6 weeks in an attempt to prevent or ameliorate HIV infection, because, at that time, most investigators were unwilling to treat pregnant women with AZT. Although some of this reluctance stemmed from concerns over maternal side effects (the drug causes anemia when given at high doses), even greater reservations were related to fetal toxicity, especially the possibility of inducing birth defects. Unfortunately, key information needed to assess the risk-benefit of treatment during pregnancy for mother and fetus was, and to a large degree still is, unavailable. Estimates of the overall risk of transmission from mother to fetus have decreased steadily from approximately 50 percent to 25-35 percent (Hardy, 1991), and preliminary data and extrapolation from other diseases have raised the possibility that the virus might be transmitted during the birth process itself. A registry of over 40 cases in which pregnant women had received AZT during at least part of their pregnancy revealed no cases of teratogenicity likely due to AZT. Armed with these admittedly limited data, the investigators for protocol 076 proposed including women in the
third trimester of pregnancy, on a randomized basis, to receive either AZT or a placebo until the birth of the child. The newborn would then receive the same medication as the mother for an additional 6 weeks after birth. This design was chosen to decrease the risk of fetal malformation by starting therapy well after organ development was complete. Study organizers believed that the risk of giving AZT to an uninfected fetus was acceptable because of the inability to identify such fetuses and the relatively high likelihood of transmission. Thus, the pediatricians who designed the trial believed the study to be timely and ethical.
Criticism of the design, however, was immediate and severe. Advocates of women with HIV disease and some AIDS Clinical Trials Group investigators argued that the focus of the protocol, including any risk-benefit analysis, should be the pregnant woman, not the fetus. Previously expressed concerns that women were considered of importance in the epidemic only in relation to their role as "vectors" of disease to their unborn children or as sexual partners seemed to them to be borne out by the design of protocol 076 (Mitchell, 1988). Critics also pointed out that AZT had been demonstrated to prolong life in patients with AIDS and to delay progression to AIDS in asymptomatic HIV-infected individuals with CD4 counts below 500. They argued that by randomizing pregnant women to receive AZT or a placebo, the design would deny some women essential therapy in order to determine the effects of the drug on preventing fetal transmission. To meet concerns that in this protocol, and others, decisions regarding the treatment of women were not being addressed by those qualified to make them, a working group composed of obstetrician-gynecologists, internists, and pediatricians was established within the AIDS Clinical Trials Group. The working group has since been raised to the level of a full scientific committee and presumably will be consulted regarding all studies in which pregnant women will be enrolled.
Ultimately, the FDA asked its Anti-Viral Advisory Committee to review protocol 076. That review, in September 1990, resulted in recommendations to radically redesign the study. In general, the recommendations supported the view that women should be the focus of the trial. Thus, the committee recommended that women with AIDS or severe immune dysfunction be excluded and given AZT. Women with moderate immune dysfunction, although eligible, would be told that if they were not pregnant, AZT would often be recommended, although the risk of delaying therapy for 6 months was believed to be small. Randomization much earlier, just after the first trimester, was recommended to maximize the possible benefit in preventing fetal transmission.
Advocates have recently voiced new doubts about the ethical propriety of the protocol. They point out that the carcinogenic effects of AZT in women are unknown because studies of AZT have included very few women.
(Female rodents given a high dose of AZT for prolonged periods have developed cervical dysplasia.) Advocates argue that the prime group to be recruited for this study (HIV-positive women with more than 500 CD4 cells) receive no known benefit from AZT and face a possible hazard in its use, genital carcinogenesis. Again, they argue that the protocol subsumes the interests of the mother to those of the fetus.
However this debate evolves and whatever final shape protocol 076 takes, it is clear that this experience has altered how pregnant women are viewed as trial subjects, and it has highlighted the lack of information on gender-specific treatment effects in women. Whether it will have far-reaching effects on how decisions are made regarding the use of experimental or standard therapies with pregnant women with other diseases is unclear, but given that the entire area of research in women's health is undergoing dramatic evolution, it is likely that developments regarding AIDS will be scrutinized and considered as possible models.
Dissemination of Information
Peer-Reviewed Reporting of Clinical Research
Dissemination of information regarding research results has traditionally occurred first through peer-reviewed scientific and medical publications. This process begins with an author's writing and rewriting of a manuscript and submitting it to a journal, initial editorial consideration, and, usually, referral to outside reviewers. Editors then compile and compare reviews and make a final decision regarding publication. Papers may be accepted outright, accepted under the condition that changes are made, returned to the author for possible resubmission, or rejected outright. Rejected manuscripts are almost always submitted to other journals, beginning the time-consuming review process again. Once a paper is accepted for publication, considerable time usually occurs before it is actually published, often 6 months to 1 year. After publication, months to years may elapse before the results of the research make their way into clinical practice or are widely accepted and taught by the scientific community. Codification of results into textbooks and removal of outdated or even disproved information from textbooks may take longer.
Although cumbersome, this process has on the whole been viewed as appropriate. Outside peer review—in theory and usually in practice—provides an objective assessment of the accuracy and relevance of research findings by experts in the same field. It has benefits for authors, who are often given advice by reviewers that strengthens or extends their work, and society benefits from the prevention of dissemination of information that is not sound. These advantages have been considered to outweigh the inherent
disadvantage of delay. However, many people have questioned whether in the case of clinical research that appears to show a significant therapeutic benefit for serious diseases or that demonstrates a previously unknown toxicity of accepted therapy the slow and orderly process of peer review and publication is in fact appropriate.
Studies pertaining to AIDS have not been the only ones in which disclosure before publication has been championed or actually occurred. Repeatedly in the past decade, authors, journalists, physicians, and public health officials have protested specific instances in which promising results were delayed from public announcement for months prior to journal publication. In response to these criticisms, some journals have adopted various "fast track" approaches to manuscript review and publication, in essence identifying some manuscripts as high priority and hastening them through the process. In addition, several government agencies have themselves asked for permission from journal editors to release information to the lay press and medical practicing communities before manuscript publication and in some cases, even before manuscript review. For example, in May 1988, the National Cancer Institute sent out a "Clinical Alert" informing the public that adding chemotherapy to initial surgery for women with node-negative breast cancer increased survival. Publication of the findings in the New England Journal of Medicine, which had agreed to the "Clinical Alert," did not occur until 9 months later. Similar announcements in the past 5 years have included information concerning a protective effect of corticosteroids for spinal cord injury, a news conference to announce an unexpected mortality from two widely used antiarrhythmic drugs in a large clinical trial, and announcements of termination of a clinical trial of AZT versus a placebo in delaying the progression to AIDS among HIV-positive asymptomatic subjects.
Such announcements have been criticized, however, as not providing sufficient information to permit physicians and patients to proceed with clinical care, for which they would want to have detailed information, such as drug dosage, drug side effects and interactions, and any caveats concerning the applicability of the results to types of patients other than those included in the trial. In addition, although most promising studies that have been announced in this way had some level of review by such groups as independent data-safety monitoring boards, early release of their results came with the risk that the information would be found to be inaccurate, which would lead to the need for rapid retraction and, perhaps, to morbidity and mortality.
The issue of whether a more pragmatic approach to traditional peer review is needed is just beginning to be fully considered. Some people have suggested that guidelines should define studies that are of sufficient importance and whose inherent design and internal review mechanism are
of sufficient rigor that they should be singled out for rapid, intensive review and early disclosure of relevant results. It is likely that continued refinements, large and small, will be proposed that will alter the process of research review and dissemination. Because of the large number of AIDS-related clinical studies, including clinical trials, and the urgency of AIDS research, it is likely that much of the debate and change will take place in the context of the epidemic.
The dissemination of AIDS research information through AIDS-specific, nontraditional means has been explosive. Among the most interesting of these are patient-oriented publications, which started out very modestly but have grown in sophistication and influence as well as circulation. AIDS Treatment News (mentioned above), which is produced in San Francisco by John S. James (a former computer programmer), has a circulation of about 5,000. Another influential newsletter is San Francisco Project Inform's PI Perspective. Supported by a staff of 11 and numerous volunteers, it has a circulation of 50,000. Project Inform also operates an AIDS treatment hotline (Bishop, 1991).
Many mainstream AIDS researchers and clinicians are subscribers to AIDS treatment newsletters, and a few even contribute to such publications. Often, these publications provide editorial comment on many aspects of the epidemic, such as regulation and access to drugs, HIV antibody testing among various groups at risk, and immigration and travel restrictions for HIV-positive individuals. Completion and updating of information regarding all known AIDS drugs has been done for several years by the American Foundation for AIDS Research through its AIDS/HIV Treatment Directory. The Public Health Service operates a toll-free hotline (1-800 TRIALS-A) that provides information on federally sponsored clinical trials and industry-sponsored efficacy trials.
From the beginning of the epidemic, the problem of teaching busy practitioners about a new disease of rapidly growing prevalence and importance was appreciated. A variety of newsletters for primary care practitioners were established to meet this need, such as AIDS Alert, AIDS News, and AIDS Clinical Care (published by the Massachusetts Medical Society, publishers of the New England Journal of Medicine). These publications are uniquely suited to providing up-to-date information on AIDS through a mixed format of in-depth clinical reviews, annotated summaries of articles in leading peer-reviewed journals, and reporting of late-breaking stories of interest. They fill a gap between mass media reporting and traditional peer-reviewed journals and as such are uniquely suited to the exigencies of the epidemic.
Science and Health Reporting in the Mass Media
During the 1980s medical and scientific reporting in the mass media became increasingly well supported and sophisticated, changing from simple repetition of the abstracts and press releases that accompanied publication of articles in medical journals to careful and then highly critical commentary. Major newspapers often employ several full-time science and medical reporters, and local and network television shows regularly air health segments. As science and medicine have become big business, they have become big stories. Because of the dramatic nature of the HIV/AIDS epidemic, the changes it has generated in all aspects of science and medicine, and its extraordinary social ramifications, AIDS has become the biggest of those stories and as such has driven many of the media changes that have occurred (Kinsella, 1989).
Ironically, there was almost total avoidance of AIDS by the media early in the epidemic. The turning point came in the mid-1980s, and it is often tied to the coverage given to Rock Hudson when it was learned that he had AIDS. Now, virtually all AIDS stories of importance and most minor ones are covered by the daily print media. Indeed, many AIDS researchers rely first and foremost on several key newspapers to keep themselves abreast of AIDS research news. The number of public service announcements, local and national "AIDS specials," and AIDS segments on major television news shows has been remarkable. Indeed, the Public Broadcasting System had for some time a regular feature, "AIDS Quarterly," hosted by anchor Peter Jennings; it has continued but is now the "Health Quarterly."
It is impossible at this juncture in the HIV/AIDS epidemic to reach any definitive conclusions about its long-term impact on clinical research and the regulation of new drugs. It is apparent that patient activism, a political climate favoring deregulation, and the exigencies of the AIDS epidemic have generated the most significant reevaluation of the research and regulation processes to occur since World War II. As this chapter shows, the AIDS epidemic has had profound effects on the conduct and nature of clinical research and on the regulation of drugs.
First, not since randomized clinical trials became the orthodox mode of clinical investigation had the most basic approaches and assumptions regarding research methodologies been open to searching critique in the context of an epidemic disease. AIDS has led to a fundamental reconsideration of basic methodologies for establishing the efficacy and safety of pharmaceuticals. New approaches to experimentation—to shorten the time required, as well as to alter the criteria for assessment—are actively being
explored. Moreover, the very legitimacy of randomization has been called into question as the ethics of withholding treatments has loomed large in the context of the epidemic. At this time it seems clear that randomized clinical trials will continue to have a significant role in clinical investigation, but that they will be designed in new ways in order to answer a revised set of questions. Moreover, it seems likely that the new methodologic approaches will not be limited to AIDS investigations, but, rather, will be diffused to other research fields. Although randomized clinical trials continue to offer significant scientific advantages for the evaluation of new drugs, it seems likely that in the context of AIDS, alternative approaches to clinical investigation will be proposed and evaluated in the years ahead.
Second, there has been a dramatic shift in the face of the epidemic from a restrictive ethos for regulating new drugs to a new, less restrictive environment. Nevertheless, the balance between protection and access is likely to be the focal point of intense debate in the future. The evolution of the parallel track, expanded compassionate drug allocations, and the general trend to expedite trials—all are clear indicators of the impact that the HIV/AIDS epidemic has had on drug regulation. Although the epidemic has spurred these changes, it now seems likely that in certain instances attempts will be made to expedite the regulatory process for drugs for other diseases, especially those for life-threatening or previously untreatable conditions. Reports of toxicity, morbidity, or, especially, drug-related mortality, however, could lead to new restrictions in the future.
Third, changes in clinical trials and the regulation of new drugs are obviously having a related impact on the pharmaceutical industry. Although the issue of the relationship between federal research and private industry has been most vocally raised in the instance of AZT, fundamental questions of public-private cooperation in research and development have been posed by the epidemic. To date, many of these complex issues have been resolved essentially on an ad hoc basis, but that public-private relationship, which is outside the scope of this chapter, merits much more attention. Moreover, the incentives and disincentives for the development of new drugs in a new regulatory environment also needs much more scrutiny.
Fourth, to a degree that few could have anticipated, AIDS has influenced the nature and meaning of the ethics of human investigation. AIDS has forced clinical medicine to consider the variable of time in a new way: added to the normative care and precision of clinical investigation has been the new demand of urgency. The epidemic will not pause for the traditional modes of science; AIDS has forced the acceleration of the procedures and processes of clinical investigation, as well as the mechanisms of regulation. The scientific community has accepted the acceleration of clinical trial phases in the context of AIDS, recognizing the urgency of early access and allocation
of new drugs for a disease that is fatal and principally strikes young people. Similar pressures were building prior to AIDS for early access to new cancer drugs. However, there is also serious concern in the research community that an effective research tool may be compromised and that the implications for both safety and efficacy of new therapeutic agents have not been fully assessed. Thus, the long-term impact of these changes is not known. Although, in the past, the very term ''therapeutic research" was held in disrepute, the lines between experimentation and therapy have now been blurred. Activists have chanted, "a drug trial is health care, too." The basic concept of human experimentation has been radically altered—from protecting individuals from research to attempting to ensure individuals access to research. These changes can be fully understood only in the context of the powerful and effective social and political activism that the HIV/AIDS epidemic has generated.
This section is based in part on a paper prepared for the panel by Harry M. Marks of the Institute for the History of Medicine at the Johns Hopkins University, "Historical Perspectives on Clinical Trials" (August 1990).
Abrams, D.I. (1990) Alternative therapies in HIV infection. AIDS 4:1179-1187.
AIDS Coalition to Unleash Power (ACT-UP) (1989) A National AIDS Treatment Research Agenda. Paper distributed at the Fifth Annual International Conference on AIDS, Montreal, June 4-9.
Annas, G.J. (1988) AIDS, judges and the right to medical care. Hastings Center Report 18(4):20-22.
Arras, J.D. (1990) Noncompliance in AIDS research. Hastings Center Report 20:24-32.
Barber, B. (1976) The ethics of experimentation with human subjects. Scientific American 234:25-31.
Bayer, R. (1990) Beyond the burdens of protection: AIDS and the ethics of research. Evaluation Review 14:443-446.
Beecher, H.E. (1966) Ethics and clinical research. New England Journal of Medicine 274:1354-1360.
Bishop, K. (1991) Underground press leads way on AIDS advice. New York Times December 16:A-10.
Bozette, S.A., F.R. Sattler, J. Chiu, A.W. Wu, D. Gluckstein, et al. (1990) A controlled trial of early adjunctive treatment with corticosteroids for Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. New England Journal of Medicine 321:1451-1457.
Brandt, A.M. (1978a) Polio, politics, publicity, and duplicity: ethical aspects in the development of the Salk vaccine. International Journal of Health Services 8:257-270.
Brandt, A.M. (1978b) Racism and research: the case of the Tuskegee syphilis study. Hastings Center Report 8:26-27.
Carter, R. (1966) Breakthrough: The Saga of Jonas Salk. New York: Trident.
Chalmers, T., J. Block, and S. Lee (1972) Controlled studies in clinical cancer research. New England Journal of Medicine 287:75-78.
Cotton, P. (1990a) Examples abound of gaps in medical knowledge because of groups excluded from scientific study. Journal of the American Medical Association 263:1051, 1055.
Cotton, P. (1990b) Is there still too much extrapolation from data on middle-aged white men? Journal of the American Medical Association 263:1049-1050.
Delaney, M. (1989a) The case for patient access to experimental therapy. Journal of Infectious Diseases 159:416-419.
Delaney, M. (1989b) Patient access to experimental therapy. Journal of the American Medical Association 261:2444-2447.
D'Emilio, J. (1983) Sexual Politics, Sexual Communities: The Making of a Homosexual Minority in the United States, 1940-1970. Chicago: University of Chicago Press.
Dresser, R. (1992) Wanted: single, white male for medical research. Hastings Center Report 22(1):24-29.
Eckholm, E. (1986) Should the rules be bent in an epidemic? New York Times July 13:E30.
Edgar, H., and D.J. Rothman (1991) New rules for new drugs: the challenge of AIDS to the regulatory process. In D. Nelkin, D.P. Willis, and S.V. Parris, eds., A Disease of Society: Cultural and Institutional Responses to AIDS. New York: Cambridge University Press.
Faden, R.R., and T.L. Beauchamp (1986) A History and Theory of Informed Consent. New York: Oxford University Press.
Feinstein, A. (1967) Clinical Judgment. Baltimore, Md.: Williams and Wilkins.
Fried, C. (1974) Medical Experimentation: Personal Integrity and Social Policy. New York: American Elsevier.
Gray, B. (1975) Human Subjects in Medical Research: A Sociological Study of the Conduct and Regulation of Clinical Research. New York: John Wiley and Sons.
Green, S.B., S.S. Ellenberg, D. Finkelstein, A.B. Forsythe, L.S. Freedman, et al. (1990) Issues in the design of drug trials for AIDS. Controlled Clinical Trials 11:80-87.
Gross, J. (1991) Turning disease into a cause: breast cancer follows AIDS. New York Times January 7:A1.
Hand, R. (1989) Alternative therapies used by patients with AIDS (letter). New England Journal of Medicine 320:672-673.
Hardy, L.M., ed. (1991) HIV Screening of Pregnant Women and Newborns . Committee on Prenatal and Newborn Screening for HIV Infection, Institute of Medicine. Washington, D.C.: National Academy Press.
Harris, R. (1964) A Real Voice. New York: Macmillan.
Henninger, D. (1990) Will the FDA revert to type? Wall Street Journal December 12:A16.
Hill, A.B. (1951) The clinical trial. British Medical Bulletin 7:278-282.
Hill, A.B. (1963) Medical ethics and controlled trials. British Medical Journal 1:1043-1049.
Insight Team (of the Sunday Times of London) (1979) Suffer the Children: The Story of Thalidomide. New York: Viking Press.
James, J.S. (1989) AIDS Treatment News. Berkeley, Calif: Celestial Arts.
Johnson, N., R.J. Lilford, and W. Brazier (1991) At what level of collective equipoise does a clinical trial become ethical? Journal of Medical Ethics 17:30-34.
Jones, J.H. (1981) Bad Blood: The Tuskegee Syphilis Experiment. New York: Free Press.
Jonsen, A. (1984) Public policy and human research. Pp. 3-19 in J. Humber and R. Almeder, eds., Biomedical Ethics Reviews. Clifton, N.J.: Humana Press.
Jonsen, A.R. (1991) Old Ethics, New Medicine. Cambridge, Mass.: Harvard University Press.
Katz, J. (1984) The Silent World of Doctor and Patient. New York: Free Press.
Katz, J., and A.M. Capron (1975) Catastrophic Diseases: Who Decides What?: A Psychological and Legal Analysis of the Problems Posed by Hemodialysis and Organ Transplantation . New York: Russell Sage Foundation.
Kinsella, J. (1989) Covering the Plague: AIDS and the American Media. New Brunswick, N.J.: Rutgers University Press.
Kolata, G. (1991) U.S. weighs release of an AIDS drug before tests are completed. New York Times June 19:A-11.
Levine, C. (1988) Has AIDS changed the ethics of human subjects research? Law, Medicine & Health Care 16:167-173.
Levine, C. (1990) Women and HIV/AIDS research: the barriers to equity. Evaluation Review 14:447-463.
Marks, H.M. (1990) Historical Perspectives in Clinical Trials. Working paper prepared for the National Research Council Committee on AIDS Research and the Behavioral, Social and Statistical Sciences, Washington, D.C.
Martin, J.M., and H.S. Sacks (1990) Do HIV-infected children in foster care have access to clinical trials of new treatments? AIDS & Public Policy Journal 5:3-8.
Merton, V. (1990) Community-based AIDS research. Evaluation Review 14:502-537.
Mintz, M. (1967) By Prescription Only, rev. ed. Boston: Beacon Press.
Mitchell, J.L. (1988) Women, AIDS, and public policy. AIDS and Public Policy Journal 3(2):50-52.
National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research (1978) The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research. Washington, D.C.: U.S. Department of Health and Human Services.
National Commission on AIDS (1991) America Living with AIDS: Transforming Anger, Fear and Indifference into Action. Washington, D.C.: National Commission on AIDS.
National Institutes of Health (NIH) (1990) Instruction and information memorandum OER 90-5, December 11, NIH, Bethesda, Md.
Nolan, K. (1990) AIDS and pediatric research. Evaluation Review 14:464-481.
Nussbaum, B. (1990) Good Intentions: How Big Business and the Medical Establishment are Corrupting the Fight Against AIDS. New York: Atlantic Monthly Press.
Office of Technology Assessment (1983) The Impact of Randomized Clinical Trials on Health Policy and Medical Practice: Background Paper. Washington, D.C.: U.S. Government Printing Office.
Okun, M. (1986) Fair Play in the Marketplace: The First Battle for Pure Food and Drugs. DeKalb, Ill.: Northern Illinois University Press.
Palca, J. (1990) Conflict over release of clinical research data. Science 247:374-375.
Quirk, P. (1980) Food and Drug Administration. In J.Q. Wilson, ed., The Politics of Regulation. New York: Basic Books.
Rothman, D.J. (1987a) Ethical and social issues in the development of new drugs and vaccines. Bulletin of the New York Academy of Medicine 63:557-568.
Rothman, D.J. (1987b) Ethics and human experimentation: Henry Beecher revisited. New England Journal of Medicine 317:1195-1199.
Rothman, D.J. (1991) Strangers at the Bedside: A History of How Law and Bioethics Transformed Medical Decision Making. New York: Basic Books.
Ruskin, J.N. (1989) The cardiac arrhythmia suppression trial (CAST). New England Journal of Medicine 321:386-388.
Rutstein, D.D. (1970) The ethical design of human experiments. In P.A. Freund, ed., Experimentation with Human Subjects. New York: George Braziller.
Schaefer, A. (1982) The ethics of the randomized clinical trial. New England Journal of Medicine 307:719-724.
Silverman, W.A. (1985) Human Experimentation: A Guided Step into the Unknown. New York: Oxford University Press.
Silverman, M., and P. Lee (1974) Pills, Profits, and Politics. Berkeley, Calif.: University of California Press.
Sinclair, U. (1906) The Jungle. New York: Doubleday Page & Company.
Steinbrook, R., and B. Lo (1990) Informing physicians about promising new treatments for severe illnesses. Journal of the American Medical Association 263:2078-2082.
Thomas, S.B., and S.C. Quinn (1991) The Tuskegee syphilis study, 1932-1972: implications for HIV education and AIDS risk education programs in the black community. American Journal of Public Health 81:1498-1505.
U.S. Congress, House of Representatives (1991) Drugs for Opportunistic Infections in Persons with HIV Disease. Hearing before the Human Resources and Intergovernmental Relations Subcommittee of the Committee on Government Operations, House of Representatives, 2nd session, August 1, 1990. Washington, D.C.: U.S. Government Printing Office.
U.S. Department of Health and Human Services, Advisory Committee on the Food and Drug Administration (1991) Final Report. Washington, D.C. : U.S. Government Printing Office.
U.S. General Accounting Office (1990) National Institutes of Health: Problems in Implementing Policy on Women in Study Populations. Testimony, June 18, Mark V. Nadel, Human Resources Division. Subcommittee on Health and the Environment, House Committee on Energy and Commerce. GAO/T-HRD-90-38. Washington, D.C.: General Accounting Office.
Valdiserri, R.O., G.M. Tama, and M. Ho (1988) The role of community advisory committees in clinical trials of anti-HIV agents. IRB: A Review of Human Subjects Research 10(4):5-7.
Wastila, L.J., and L. Lasagna (1990) The history of zidovudine (AZT). Journal of Clinical Research and Pharmacoepidemiology 4:25-37.
Weiss, R., and L. Mazade (1990) Surrogate Endpoints in Evaluating the Effectiveness of Drugs Against HIV Infection and AIDS. Roundtable for the Development of Drugs and Vaccines Against AIDS, Institute of Medicine. Washington, D.C.: National Academy Press.
Wolfe, S., and C. Coley (1980) Pills That Don't Work. Washington, D.C.: Public Citizen's Health Research Group.
Young, J.H. (1990) Pure Food. Princeton, N.J.: Princeton University Press.
Zelen, M. (1990) Randomized consent designs for clinical trials: an update. Statistics in Medicine 9:645-656.