The Tuskegee Syphilis Study, perhaps the longest running study in the history of the U.S. Public Health Service, is a paradigmatic example of unethical human experimentation.1 From 1932 until 1972, researchers examined the course of untreated syphilis in 400 African-American males in rural Alabama.2 Study participants were denied access to chemotherapy for the disease, even when penicillin, a safe and effective treatment for syphilis, became available in the 1940s. Inadequate information was provided to participants and, at times, they were deceived as to the purpose of medical interventions. For example, subjects were told that lumbar punctures, done solely for research purposes, were “treatments.” It is estimated that 20% of the study participants died prematurely of untreated syphilis. The legacy of the Tuskegee Syphilis Study continues to this day as the African-American community remains distrustful of medical research.3 In an attempt to heal these wounds, President Clinton apologized to the study's survivors, and the African-American community, on May 16, 1997 saying, “[a]n apology is the first step, and we take it with a commitment to rebuild that broken trust. We can begin by making sure there is never again another episode like this one.”4
If trust is to be rebuilt and maintained, the research community must be prepared to ask hard questions about studies that raise ethical warning flags. In this issue, Collins and Jeffery report a retrospective analysis of 318 patients with syphilis who were intentionally infected with malaria at South Carolina and Georgia State institutions in association with National Institutes of Health installations between 1940 and 1963.5 We are told that of the patients treated at these institutions, about 60% were African-American and 40% were Caucasian. All of the patients in the report of Collins and Jeffery had neurosyphilis and, as a result, many if not most would have been incapable of giving consent. They were infected with Plasmodium falciparum, the most deadly form of malaria, and data on fever and the number of malaria parasites in the blood were fastidiously recorded. This information, which could not be obtained in this manner today, could be used as an historical control in research into a vaccine to prevent malaria, “the world's most important parasitic infection.”6 Prima facie, the study raises a number of ethical issues. The ensuing ethical analysis presumes that the data as presented by Jeffrey and Collins are accurate.
The retrospective analysis of Collins and Jeffery per se must be evaluated by current ethical norms as defined by the Common Rule—federal regulations governing the conduct of human subjects research. The authors report that “a data base has been extracted from parasitologic and fever records acquired between 1940 and 1963”—in short, the retrospective analysis is a chart review.5 The Common Rule states that “[r]esearch involving the collection or study of existing data, documents, records …” is exempt from federal policy provided “the information is recorded by the investigator in such a manner that subjects cannot be identified …”.7 Thus, so long as the authors recorded data from patient files without identifiers, they were neither required to have the study reviewed by an Institutional Review Board (IRE) nor obtain consent from subjects (or their next of kin) to retrieve data from their charts.
Even though the retrospective analysis per se was conducted ethically, we are left with considerable moral discomfort regarding the fate of the patients described therein. Was what happened to these people between 1940 and 1963 another Tuskegee? To answer this over-arching question, we will need to address several issues: Were patients exposed to inordinate research risks? Were they exploited? Was adequate consent obtained? And finally should the data about malaria be published at all? As above, if research is to be evaluated fairly, it must be appraised according to the prevailing standards when it was conducted. (This issue was encountered, and is explored at length, in the final report of the President's Advisory Committee on Human Radiation Experiments).8
The experience of these patients from 1940 to 1963, therefore, necessitates a more complex ethical analysis. Central to this analysis is the question: What, if anything, about the interventions to which these patients were exposed constitutes research? Was the intentional infection of the patients with malaria a research procedure? Were there other research procedures performed on these patients? Colleagues and I have argued elsewhere for a systematic approach to the ethical analysis of clinical research that begins with distinguishing interventions that fall under clinical practice from those properly understood as research.9,10 Clinical practice and research are governed by differing sets of rules. Clinical practice is governed by the norms of the physician-patient relationship and is subject to retrospective review by the institution, professional association, or the courts. Research, on the other hand, is regulated by federal policy and is subject to prospective review by IRB s. Implicit in this demarcation is the view that the IRE has no right to regulate clinical practice. Elements of care that fall within standard therapy, that is, “treatment accepted by the community of practitioners possessing the relevant expertise,” are not research and, therefore, fall outside of the proPer bounds of federal policy and IRE scrutiny.10
Was the administration of malaria for neurosyphilis from 1940 to 1963 a standard therapy or research? From the perspective of our post-penicillin era, intentionally infecting persons with malaria for neurosyphilis seems drastic, to say the least. But in the pre-penicillin era, syphilis was a major public health problem: in the 1920s, for example, perhaps 20% of patients in U.S. mental institutions had neurosyphilis.11 The disease was treated with the arsenicals arsphenamine and neoarsphenamine, but these agents were both toxic and relatively ineffective for patients with advanced neurosyphilis.12 Against this background, it is understandable that an effective treatment for neurosyphilis would be greeted with enthusiasm. Writing only 9 years after the first use of malaria in syphilis, O'Leary wrote in JAMA that “[e]xperience with the fever treatment of Wagner von Jauregg [i.e., malaria] during the last years in 213 cases … has definitely proved that it is the most valuable method of treatment suggested for paretic parenchymatous neurosyphilis.”13
Infecting patients who have neurosyphilis with malaria was a standard therapy, and is properly placed within the bounds of clinical practice. It is not, therefore, a research intervention. Even as a standard therapy, infection with malaria seems to have been one of the more effective treatments available. One modem review of the literature estimates that about equal proportions of patients with neurosyphilis treated with malaria went into a full remission, had a partial response, or no response.14 Even after the advent of penicillin, infection with malaria was commonly administered in combination with penicillin to treat neurosyphilis until the mid-1960s in the United States and the early 1970s in the United Kingdom.14
Whether other procedures administered count as research or not is a difficult question. We are told that blood was drawn to measure parasite levels daily at first and then two or three times weekly later in the illness. Both fever and parasitemia were used to guide treatment to control (rather than cure) the malaria in these patients. In so far as blood was taken to guide clinical care of the patient, these blood tests were a part of the clinical care of the patient and not research. It is unclear, however, whether blood was drawn more often than needed to support even optimal clinical goals. If this were the case, these additional tests would constitute research procedures. It is also not clear whether the data collection initially was a part of the patients' charts or was collected centrally with the intention of producing generalizable knowledge. While ambiguities remain, we can say with confidence that the research interventions to which patients were exposed constituted at most additional blood tests and the fastidious recording of data—by most accounts, procedures associate with minimal risk.
One of the central injustices of the Tuskegee Syphilis Study was the exploitation of African-American research subjects. They were wantonly denied access to necessary medical treatment. No evidence of exploitation, however, can be found for the patients with neurosyphilis. While penicillin was unavailable until the mid-1940s, there is no indication that patients with neurosyphilis were denied access to penicillin subsequent to that time. Indeed, our review of the role of malaria infection in the treatment of neurosyphilis indicates that, at the time, it was the state of the art treatment for the disease. It is, therefore, laudable that 60% of those infected with malaria for neurosyphilis were African-American—a community often excluded from access to high-quality medical care.
Collins and Jeffery tell us that “[g]eneral consent for the treatment of patients for mental disorders was granted when patients were admitted to the hospitals, either by the family or the courts.” While sufficient for treatment within the bounds of clinical practice, such consent may not be sufficient if any research procedures were added to the care of these patients. (While it is unclear that the infection of patients with neurosyphilis involved any research interventions, our analysis will proceed assuming a “worst-case scenario,” namely, that additional blood tests and data collection were done for research purposes.)
Particularly problematic is the Nuremberg Code requirement that “[t]he voluntary consent of the human subject is absolutely essential. This means that the person involved should have legal capacity to give consent …”—a statement that effectively prohibits research involving children or the mentally infirm.15 The Nuremberg Code, however, had little impact in North American researchers: “From their perspective, the Code had nothing to do with science and everything to do with Nazis.”16 The first National Institutes of Health policy was developed in 1953 (coinciding with the opening of the Clinical Center) and required written consent for research participation involving “normal” (i.e., healthy) subjects. Research involving patients, however, was held to a lower standard and only required “voluntary agreement based on informed understanding” when research presented risks that the physician believed to be “unusually hazardous.”17 This curious double standard was a product of the paternalism of medical practice at the time and was not corrected until reforms in federal policy were initiated in the mid-1960s. Nonetheless, we must conclude the lack of written—or perhaps even specific—consent for research procedures (if any) performed on the neurosyphilis patients is consistent with practice at the time.
The infection of neurosyphilis patients with malaria forced us to ask a disturbing question: Is this another Tuskegee Syphilis Study? Our ethical analysis revealed that infection with malaria was a standard treatment for those with neurosyphilis in the years 1940 to 1963 and not, therefore, research at all. Research interventions at most involved a few additional blood tests and centralized recording of data—minimally risk interventions, approvable under today's stringent ethical standards for research. There is no evidence that these patients were exploited, indeed, many were lucky to have access to state-of-the-art care. Finally, even if patients were exposed to some minimally-risky research interventions, the standards for informed consent of the day were not violated. We must conclude, therefore, that this is not another Tuskegee.
When a study raises ethical problems, Levine rightly recommends that the “decision should be made to publish the article along with an editorial in which the ethical questions are raised.”18 The approach is wise, in part, because it recognizes that there is a value in airing ethical issues in research. At times, such ethical analysis leads to pleasant surprises, as is the case here. The papers of Collins and Jeffery present data that will be invaluable to future malaria research, including vaccine development. It is gratifying to conclude that this important end was achieved in an ethically supportable manner.
Acknowledgement: I am grateful to Francoise Balyis, Robert Crouch, Nuala Kenny, and Robert J. Levine for helpful suggestions on an earlier draft of this paper.
Freedman B. 1995. Unethical research Reich WT, ed. Encyclopedia of Bioethics, Volume IV. New York: Simon & Schuster MacMillan, 2258–2261.
Harris Y, Gorelick PB, Samuels P, Bempong I, 1996. Why African-Americans may not be participating in clinical trials. 1 Natl Med Assoc 88: 630–634.
Collins WE, Jeffery GM, 1999. A retrospective examination of sporozoite- and trophozoite-induced infections with Plasmodium falciparum. Am J Trop Med Hyg 61 (suppl):4–48.
45 CFR 46.101(a)(4).
Advisory Committee on Human Radiation Experiments. 1996. The Human Radiation Experiments: Final Report of the President's Advisory Committee. New York: Oxford University Press, 113–131.
Freedman B, Fuks A, Weijer C, 1992. Demarcating research and treatment: a systematic approach for the analysis of the ethics of research. Clin Res 40: 653–660.
Freedman B, Weijer C, 1992. Demarcating research and treatment interventions: a case illustration. IRB Rev Hum Subjects Res 14: 5–8.
Sartin JS, Perry HO, 1995. From mercury to malaria to penicillin: the history of the treatment of syphilis at the Mayo clinic— 1916–1955. J Am Acad Dermatol 32: 255–261.
O'Leary PA, 1927. Treatment of neurosyphilis by malaria: report on the three years' observation of the first one hundred patients treated.. JAMA 89: 95–100.
Rothman RJ, 1995. Human research: historical aspects. Reich WT, ed. Encyclopedia of Bioethics. Volume IV. New York: Simon & Schuster MacMillan, 2248–2258.
Advisory Committee on Human Radiation Experiments. 1996. The Human Radiation Experiments: Final Report of the President's Advisory Committee. New York: Oxford University Press, 65.