Nuremberg Revisited
History and evolution of medical ethics
NUREMBERG REVISITED HISTORY AND EVOLUTION OF ETHICS IN MEDICINE A Historical Account of Human Research, Ethical Failures, Regulatory Reform, and the Continuing Demand for Accountability Col (Dr) Arvind Singh Kushwaha, MDx Lapsus Creations © Lapsus 2026 | All Rights Reserved
Nuremberg Revisited 2
Nuremberg Revisited 3 Copyright © 2026 Col (Dr) Arvind Singh Kushwaha, MD All rights reserved. No part of this book may be reproduced, distributed, stored, or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, scanning, or any information storage and retrieval system, without prior written permission from the author, except for brief quotations used in reviews or other permitted uses under applicable copyright law. NUREMBERG REVISITED First Edition, 2026 ISBN: Printed and published by: Lapsus Creations by Lapsus Written by Col (Dr) Arvind Singh Kushwaha All characters, stories, examples, reflections, and situations presented in this book are intended to communicate ideas, lessons, and perspectives. Where real-life experiences or observations have influenced the writing, they have been adapted as appropriate for the purposes of storytelling. This book is not intended to provide medical, psychological, financial, legal, or professional advice. Readers should seek appropriate professional guidance for decisions requiring specialist advice.
Nuremberg Revisited 4 ABOUT THE AUTHOR Dr. (Col) Arvind Singh Kushwaha, BSc, MBBS, MD, FRSPH Dr. (Col) Arvind Singh Kushwaha is a physician, public health specialist, teacher, researcher and military medical professional whose career spans more than three decades across clinical medicine, military medicine, community health, medical education and healthcare administration. He completed his MBBS and MD in Preventive and Social Medicine at the Armed Forces Medical College, Pune, and has more than 28 years of teaching experience in Community Medicine. He has served as a Professor since 2011, including at AFMC Pune and AIIMS Nagpur. He served for 28 years in the Armed Forces Medical Services in clinical, field, public-health, administrative and academic appointments. His service included postings in Punjab, Jammu and Kashmir, Rajasthan and Siachen, where medicine often had to be practised amid difficult terrain, limited resources, long evacuation distances and rapidly changing operational circumstances. His public-health work has included healthcare planning and outreach during the Vaishnodevi and Amarnath yatras, medical camps and civilian programmes in Jammu and Kashmir, disaster-response work after the Bhuj and Jammu and Kashmir earthquakes, and community services through rural and urban health-training centres. He has also contributed to medico-social outreach through the Prayas Club initiative at AFMC Pune. As an academician and researcher, Dr. Kushwaha has published more than 50 papers in national and international journals, including 31 indexed in Scopus, and contributed more than 30 chapters to books and academic texts. His work includes contributions to the Manual of Health for the Armed Forces, WHO public-health teaching material, ECAB – Pediatric Clinical Update and editions of the Textbook of Community Medicine. His interests include infectious diseases, military medicine, epidemiology, public health, outbreak investigation, preventive medicine and health systems. He has taught, examined and mentored medical students and postgraduate trainees, and has represented India at military-medicine and public-health forums in Bangladesh, Nigeria, Thailand, Nepal and Switzerland, including the
Nuremberg Revisited 5 World Congress on Military Medicine. Following military service, he became a founding faculty member and Additional Professor of Community Medicine at AIIMS Nagpur. He later served as Founding Principal and Dean of Autonomous State Medical College, Auraiya, and as Principal and Dean of Shri Gorakshnath Medical College Hospital and Research Centre, Gorakhpur. During the COVID- 19 pandemic, he served as Nodal Officer for COVID management at AIIMS Nagpur and worked with Ministry of Health and Family Welfare response teams in Maharashtra and Beed district. In this memoir, he returns to the patients, teachers and everyday encounters that shaped his early years as a doctor. The stories reflect a professional life spent between the consulting room and the field, the classroom and public service, and remain grounded in the belief that medicine begins with clinical knowledge but is fulfilled through responsibility, humility and care for the person behind the diagnosis.
Nuremberg Revisited 6
Nuremberg Revisited 7 Contents PART I — BEFORE NUREMBERG: ORIGINS OF HUMAN EXPERIMENTATION 1. The Human Being as a Subject of Science 2. The Ethics of Early Medical Research 3. Colonialism, Race and the Power to Experiment 4. Eugenics, Forced Sterilisation and Scientific Racism PART II — THE SECOND WORLD WAR AND THE BIRTH OF MODERN RESEARCH ETHICS 5. Medicine Under the Nazi Regime 6. Concentration-Camp Experiments: Science Without Consent 7. Unit 731: Biological Warfare and the Limits of Post-war Accountability 8. The Doctors’ Trial and the Nuremberg Code PART III — AFTER NUREMBERG: VIOLATIONS IN MODERN SCIENCE 9. Tuskegee: Four Decades of Deception 10. Guatemala, 1946–1948: Deliberate Exposure Without Valid Consent 11. Soldiers as Research Participants: Duty, Discipline and Consent 12. MKUltra and Behavioural Experimentation 13. Human Radiation Experiments: Secrecy, Consent and the Cold War 14. Prisoners, Children and Institutionalised Patients PART IV — PHARMACEUTICAL RESEARCH, LOWER-INCOME POPULATIONS AND GLOBAL INEQUALITY 15. When Research Crosses Borders: The Geography of Risk and Benefit 16. Kano, 1996: Trovan and the Ethics of an Epidemic 17. The Placebo Question: HIV Prevention and Unequal Standards of Care 18. India’s Clinical-Trial Reckoning: Consent, Compensation and Accountability 19. When the Sponsor Owns the Evidence: Conflicts, Publication and Transparency 20. Who Owns a Human Sample? HeLa, Havasupai and the Limits of Consent PART V — THE EVOLUTION OF RESEARCH ETHICS AND REGULATION 21. From Nuremberg to Helsinki: How Ethical Principles Became a Professional Standard 22. The Belmont Report: Respect, Beneficence and Justice 23. CIOMS and Research in Lower-Resource Settings: From Protection to Partnership
Nuremberg Revisited 8 24. Ethics Committees and Independent Review: The Gatekeepers Under Pressure 25. Good Clinical Practice: Quality, Transparency and Regulatory Accountability PART VI — COVID-19: ETHICS UNDER EMERGENCY CONDITIONS 26. When Emergency Becomes a Justification 27. COVID-19 Treatment Trials: Hope, Uncertainty and the Cost of Bad Evidence 28. COVID-19 Vaccine Trials: Acceleration, Authorisation and Continuing Obligations 29. Mandates, Bodily Autonomy and Public-Health Ethics 30. Global Inequality in the Pandemic Response PART VII — THE FUTURE: ETHICS, TECHNOLOGY AND ACCOUNTABILITY 31. Genomics, Artificial Intelligence, Biobanks and Digital Surveillance 32. Human Challenge Studies, Gene Editing and Emerging Technologies 33. From Codes to Compliance: A Global Agenda for Accountability APPENDIX A — CHRONOLOGY OF MAJOR RESEARCH-ETHICS MILESTONES APPENDIX B — GLOSSARY OF KEY TERMS APPENDIX C — MAJOR ETHICAL CODES, DECLARATIONS, GUIDELINES AND REGULATORY FRAMEWORKS
Nuremberg Revisited 9 PART I BEFORE NUREMBERG: ORIGINS OF HUMAN EXPERIMENTATION Chapters 1–4
Nuremberg Revisited 10 CHAPTER 1 The Human Being as a Subject of Science The question beneath the experiment Every medical discovery involving people begins with a moral question as well as a scientific one: what may be done to a person in order to learn something that could help another? The question is older than the modern clinical trial. Physicians have observed illness, compared remedies, examined bodies and tested explanations for centuries. What changed over time was not simply the sophistication of instruments or statistics. It was the recognition that a person enrolled in research is not merely a source of information, a body on which a technique can be tried, or a means to a future benefit. The participant remains a person with interests, relationships, rights and the capacity to refuse. [1] The history of research ethics should not be written as a simple march from ignorance to enlightenment. Historical actors worked within different medical knowledge, legal systems and social assumptions. Some took extraordinary personal risks to investigate disease; some worked with patients whose alternatives were few; others used power over enslaved people, prisoners, soldiers, children or institutionalised patients to obtain access that free people might have refused. Scientific purpose and moral legitimacy have never been identical. A study can be methodologically ingenious and ethically indefensible, while a risky experiment may be justifiable only if its necessity, safeguards and consent withstand careful examination. [1, 7]
Nuremberg Revisited 11 The vocabulary also matters. The term “human subject” became common in twentieth-century regulation, but it can make a person sound like an object. Contemporary frameworks increasingly refer to human participants. This book uses historical terms when discussing historical documents, while keeping the person—not the protocol, institution or discovery—at the centre of the analysis. Observation, treatment and experimentation Not every clinical encounter is research. A physician may observe a patient in order to diagnose and treat that individual. Research usually asks a question intended to produce knowledge that can extend beyond the person receiving care; it may assign interventions, compare groups, collect samples under a protocol or analyse identifiable information. These categories can overlap. A clinician may learn from a patient while caring for them, and a study may include interventions that also offer a possibility of direct benefit. The overlap creates a recurring ethical risk: patients may believe that a doctor’s recommendation is solely for their personal care when the physician is also pursuing a research objective. Clear explanation and valid consent help maintain that distinction. [1, 3] Ancient and pre-modern medical traditions contained rules about professional duty, patient welfare, secrecy and the conduct of healers. Those rules were not equivalent to a modern framework for research involving human participants. They generally did not provide an independent review system, a universal right to withdraw from a protocol or a consistent process for distributing the benefits of scientific discovery. Nor should one assume that every earlier society treated the body as freely available for investigation. Religious, legal and cultural traditions placed their own limits on dissection, invasive procedures and the authority of healers; these limits varied widely across time and place. [1] The growth of anatomy, physiology, bacteriology and pharmacology created new opportunities to test ideas directly. Hospitals offered concentrated access to people with particular
Nuremberg Revisited 12 illnesses. Military institutions and prisons placed populations under administrative control. Poor patients often depended on charitable institutions for care. These settings could support important scientific advances, but they also created conditions in which refusal might be difficult, consent might be assumed, and an institution’s goals might prevail over an individual’s interests. The ethical question was not merely whether a doctor meant well. It was whether the person could understand the experiment, decline participation without penalty, and receive protection from unnecessary harm. [3, 7] The changing moral status of the participant For much of medical history, the physician’s authority was defended through paternalism: the professional was expected to decide what was best for the patient. Research intensified this imbalance because the benefits might accrue to future patients, the profession or the state, while the immediate risks fell on the person being studied. Some researchers believed that a worthwhile goal justified intervention; some assumed that professional judgement could substitute for permission. In other cases, people were denied the status of full moral agents because of their race, disability, poverty, incarceration or political subordination. Such beliefs made ethical boundaries dependent on who the participant was. [6, 7, 8, 9] The emergence of research ethics can be understood as a series of attempts to prevent those with power from defining the interests of others without accountability. Consent became a central principle, but consent alone is not enough. A person cannot make a meaningful choice if material risks are concealed, the science is invalid, the investigator has no plan to respond to harm, or refusal threatens the person’s access to essential services. Ethical research therefore requires a combination of scientific validity, fair selection, proportionate risk, independent scrutiny, transparent information, respect for withdrawal and arrangements for care and accountability. These requirements were articulated in modern codes, but their roots lie in earlier
Nuremberg Revisited 13 controversies and in testimony from people harmed by research. [1, 3] Not all historical episodes fit neatly into the category of abuse. Some involved self-experimentation by researchers who accepted risks themselves; others involved volunteers who understood at least some of the danger and sought a contribution to science or a practical benefit. The fact of volunteering does not, however, settle every question. Was the information sufficient? Was payment so important that it constrained choice? Could the individual stop after the study began? Were dependants affected? Did the investigator hold authority over the volunteer? Historical evaluation must consider these details rather than attach a single moral label to every experiment. [1] The central theme of this book is therefore not that science is inherently cruel or that formal regulation can eliminate all wrongdoing. It is that the pursuit of knowledge generates obligations. Researchers must justify the question, protect the people who help answer it, disclose what is uncertain and remain accountable when expectations are not met. The history that follows examines the repeated gap between those obligations and the arrangements that actually governed research. How the history will be assessed Throughout the book, five evidence labels are used. “Established violation” indicates a strong formal finding by a court, official inquiry or comparably authoritative record. “Documented ethical failure” denotes conduct supported by primary documentation and credible historical scholarship, whether or not a court ruled on it. “Contested allegation” identifies material facts or responsibility that remain disputed. “Ethical controversy” describes a genuine dispute about whether a design or policy met applicable standards. “Systemic risk” describes a structure that could permit abuse but is not proof that a particular violation occurred. These categories are not excuses or verdicts; they help the reader distinguish evidence from inference.
Nuremberg Revisited 14 The standard used in each chapter combines historical context with moral scrutiny. It is unfair to pretend that every modern form or review board existed centuries ago. It is equally mistaken to conclude that conduct was acceptable simply because a powerful institution authorised it or some professionals considered it normal. Historical context can explain why a practice seemed defensible; it does not remove the obligation to examine coercion, avoidable suffering, unequal power and the choices that were available at the time. [1, 3] The chapters begin with early experimentation and then move through colonial rule, racial science, wartime atrocities, post-war scandals, the rise of ethical codes, pharmaceutical trials, COVID- 19 and emerging technologies. The order is chronological, but the deeper question is continuous: who decides what risks may be imposed, who benefits from the knowledge, and who has the power to demand an answer when the system fails?
Nuremberg Revisited 15 CHAPTER 2 The Ethics of Early Medical Research Knowledge before the modern clinical trial Before randomised controlled trials, investigators compared treatments through observation, analogy, case reports and small experiments. The resulting evidence was often fragile, but the medical need was real. Epidemics killed large numbers of people; effective treatments were scarce; and physicians had few reliable means of distinguishing a remedy from coincidence. Early experimentation must be assessed against that uncertainty without romanticising it. A desperate clinical situation can create a strong reason to investigate, but desperation cannot by itself convert a person into an involuntary participant. [1] Inoculation against smallpox illustrates the circulation of medical knowledge across cultures. Variolation—deliberate exposure to material from smallpox lesions or scabs to induce a usually milder infection—was practised in several societies before its adoption by European elites. Histories have described forms of inoculation in parts of Asia, Africa and the Ottoman world, though the precise origins and routes of transmission are not always straightforward to reconstruct. In early eighteenth-century Britain, Lady Mary Wortley Montagu helped publicise the Ottoman practice after observing it during her time in Constantinople. Variolation reduced the risk of severe smallpox compared with naturally acquired infection in many settings, but it could itself cause serious illness and spread the disease to others. The ethical
Nuremberg Revisited 16 question was therefore not simply whether it worked, but who consented, who accepted the risks and who bore the consequences of transmission. [1] This history cautions against telling the story of medical progress as if knowledge flowed in only one direction—from European discovery to passive populations elsewhere. Medical ideas were carried by travellers, healers, communities and patients. Their contributions were not always acknowledged on equal terms. The ethics of research must include recognition of the people and knowledge systems from which an intervention emerged, as well as the individuals exposed to its testing. Jenner and the Limits of Retrospective Judgement In 1796, Edward Jenner investigated the observation that exposure to cowpox might protect against smallpox. His work helped establish vaccination as a transformative preventive intervention. The celebrated account includes the inoculation of James Phipps, a young boy, with material from a cowpox lesion, followed later by exposure to smallpox material to test whether protection had been produced. Jenner published his account in 1798. The investigation was conducted before modern rules for paediatric research, independent ethics review, or informed consent had developed. It would therefore be simplistic either to treat the episode as if it had met current standards or to erase the historical danger of smallpox and the limited alternatives available at the time. [2] A careful retrospective assessment asks what risks were foreseeable, what information was given to the family, whether the child could have declined, how the later challenge was justified, and how the work compared with prevailing medical practices. Davies argues that Jenner’s experiment had a defensible public-health purpose in its historical setting, especially against the risks of smallpox and variolation, while acknowledging the value of examining it through modern ethical questions. The point is not to force a binary verdict on Jenner. It is to recognise that a
Nuremberg Revisited 17 breakthrough can be historically important while still raising questions that later ethical frameworks would treat as fundamental. [2] It is also important to distinguish an experimental exposure intended to answer a question from subsequent clinical use after an intervention has accumulated evidence. An early experiment does not have to be perfect to produce valuable knowledge; however, scientific success must not be used to retroactively erase the interests of the participant who made it possible. The person who bears risk deserves more than a place in the story of the discovery. Lind’s scurvy trial and the emergence of comparison In 1747, naval surgeon James Lind conducted a small comparative experiment among twelve sailors suffering from scurvy aboard HMS Salisbury. He arranged six pairs of patients, gave them a common basic diet, and compared different treatments then in use. The pair receiving oranges and lemons improved most strikingly. Lind described the work in his later Treatise of the Scurvy, contributing to the history of comparative clinical investigation. The trial was small, and it took decades for the British Navy to institutionalise the provision of citrus to prevent scurvy. [4] Lind’s study is often celebrated as an early controlled clinical trial, but a controlled design does not settle its ethics. The participants were ill sailors living within a strict naval hierarchy; records do not demonstrate that they were recruited under a modern process of informed consent. Their circumstances were different from those of contemporary research volunteers. At the same time, the study did not deliberately create scurvy and sought to compare remedies for an existing disease. A balanced account can recognise its methodological significance without assuming that modern protections were in place. [1, 4]
Nuremberg Revisited 18 The history also shows that the results of an experiment do not automatically change practice. Institutions need systems that recognise good evidence, implement it and monitor consequences. Research ethics and evidence implementation intersect: exposing people to risk without a plausible path from the findings to better care may fail the very social value that is supposed to justify participation. From professional discretion to explicit safeguards By the nineteenth century, laboratory medicine and clinical experimentation expanded rapidly. The risks of introducing new substances into patients became more visible, as did concerns about conducting experiments without adequate disclosure. Germany developed early formal guidance before the Nuremberg Code. Vollmann and Winau describe nineteenth-century recognition of consent concerns, a Prussian directive in 1900 addressing the use of new treatments, and German guidance on new therapy and human experimentation issued in 1931. These rules were limited and unevenly implemented; their existence nevertheless undermines the claim that no ethical ideas about consent existed before the Second World War. [3] The existence of guidelines is not proof of compliance. A directive may be ignored, interpreted narrowly or applied only to some types of participants. The key questions are whether participants were informed, whether the practice was independently scrutinised, what remedies were available for harm and whether officials enforced the stated rules. The twentieth century repeatedly demonstrated that written principles can coexist with grave violations when institutions lack the will or capacity to enforce them. [3] The yellow-fever experiments associated with Walter Reed’s US Army commission in Cuba in 1900 offer a more complicated example. Researchers investigated mosquito transmission and recruited volunteers, including soldiers and Spanish immigrants. Accounts describe a written consent document translated into
Nuremberg Revisited 19 Spanish, disclosure of danger, provision for medical care and compensation, and researchers who also accepted risk. The experiment helped establish the role of mosquitoes in transmission; a colleague, Jesse Lazear, died after becoming infected. The programme is often identified as an early milestone in documented consent, but it took place amid military hierarchy, large financial inducements, and serious risks. The surviving evidence should be checked against the commission’s original records before precise participant and mortality figures are repeated. [5] These examples demonstrate that research ethics did not appear fully formed after 1945. It developed unevenly through medical practice, public controversy, emerging regulations, and the testimony of people whose interests were poorly protected. The Nuremberg Code was a landmark, not an absolute beginning; the Declaration of Helsinki, the Belmont Report, CIOMS guidelines, and modern good clinical practice later expanded and operationalised the obligations examined here. [3]
Nuremberg Revisited 20 CHAPTER 3 Colonialism, Race, and the Power to Experiment Medicine within unequal political systems Clinical research does not occur outside society. Colonial rule structured who could access hospitals, who controlled medical institutions, which diseases were prioritised, who conducted research, and whose accounts were considered authoritative. Medical services sometimes offered real benefits and reduced suffering; the ethical record cannot be reduced to a claim that every colonial intervention was abusive. Yet colonial systems also created steep asymmetries of power. Populations could be described as a plentiful source of clinical material, while the institutions that controlled investigation and publication were located elsewhere. [6, 12] Historian Helen Tilley describes how colonial medical services and research programmes in Africa were entwined with imperial administration, uneven health infrastructure, and sometimes coercive campaigns. Research was often undertaken in settings where concepts of consent and subject protection had not been standardised and where the ability to refuse could be shaped by local authority, economic dependency, language barriers, or the prospect of receiving care. These conditions did not make every study unethical, but they created risks that could not be assessed simply by reading the protocol. [6]
Nuremberg Revisited 21 In India, colonial knowledge systems also classified populations through censuses, medical surveys, ethnography and administrative categories. Bernard Cohn’s study of colonial forms of knowledge examines how information gathering helped render Indian society legible to the state. Not every census or survey was biomedical experimentation, and such activities should not be mislabeled as clinical trials. Their relevance to research ethics lies in the wider relationship among classification, power, consent and the later use of information collected from people and communities. [12] An important distinction follows. Public-health interventions, clinical treatment, observational research, experimental procedures and administrative data collection do not carry identical ethical obligations. But all can raise questions about coercion, privacy, fairness and who benefits from the knowledge. Good historical analysis names the activity precisely rather than treating every state health campaign as the same kind of event. Race, slavery and the medical use of bodies The history of American medicine includes research and clinical innovation carried out in societies that denied full autonomy to enslaved people. Robert Baker’s history of American research ethics discusses the use of servants, enslaved people and patients in the development of medical knowledge. In such settings, a physician’s access to a person could be secured by ownership, employment dependency or institutional authority rather than by a choice comparable to the consent expected today. [7] J. Marion Sims’s nineteenth-century operations to treat vesicovaginal fistula are a central and disputed example. Between 1845 and 1849, Sims performed repeated operations on enslaved Black women, including Anarcha, Betsey and Lucy, and his work contributed to the development of a successful surgical repair. Historical and ethical scholarship disagrees over aspects of the circumstances, including the degree of willingness expressed by the women, the available alternatives, the use of anaesthesia at
Nuremberg Revisited 22 different stages and how the surviving record should be interpreted. Wall’s 2006 review argues that some common accusations simplify the historical evidence and that the women may have participated in attempts to treat a devastating condition for which alternatives were limited. Other ethical analyses emphasise the coercive structure of slavery and the impossibility of equating any apparent cooperation with freely given consent in a system of ownership. [7, 8] The appropriate conclusion is neither to pronounce every disputed fact settled nor to let uncertainty erase the power imbalance. Enslavement constrained legal freedom, bodily autonomy, and the practical ability to refuse. Medical benefit to an individual does not, by itself, establish that every procedure was ethically justified, and later success does not compensate for the absence of meaningful choice. At the same time, historical accuracy requires careful use of primary records and consideration of the medical conditions, available treatments, and standards of the period. The case is important partly because it shows how the same historical record can be used to ask different questions—and why those questions need to be made explicit. [7, 8] Racial hierarchy also influenced whose pain was believed and whose health was considered valuable. In settings where stereotypes portrayed certain groups as less sensitive to pain or less deserving of care, the ordinary duty to minimise suffering could be applied unequally. Such beliefs were not sound biology; they were social assumptions capable of entering clinical practice. The lesson is not that race is irrelevant to medicine, but that social categories can be misused as if they were fixed biological explanations, thereby legitimising unequal treatment and research burdens. [6, 7, 8, 9] Colonial research and the distribution of benefit Colonial-era campaigns against infectious disease provide a mixed record. Some generated knowledge that contributed to prevention
Nuremberg Revisited 23 and treatment; some relied on compulsory measures, limited disclosure or administrative force. When authorities were responsible for both protecting a population and maintaining labour, trade, or military systems, their priorities could conflict. A campaign might be called public health while also serving colonial administration. Tilley’s review stresses the need to examine both intended and unintended harms, the quality of local infrastructure, and the treatment of indigenous healing knowledge. [6] The ethical questions remain recognisable in modern international research. Are local investigators and institutions involved in setting the question? Does the study respond to a health priority identified by the host community, or mainly to the commercial or academic interests of an external sponsor? Is consent available in languages participants understand? Are community leaders consulted without being allowed to replace each individual’s consent? Will successful products be affordable and available locally? Will samples and data be returned, shared or stored under fair terms? These questions are especially pressing when the research site has limited bargaining power. [6, 12] The term “human laboratory” captures a fear that poorer communities may bear risks while better-resourced countries capture the benefits. It can be a useful warning but not a conclusion to be applied without evidence. Research conducted in a low-income country may be scientifically important, ethically justified and locally valuable. Conversely, a study based in a wealthy country can exploit prisoners, institutionalised patients or economically vulnerable people. Geography is an indicator of possible inequality, not a substitute for analysing consent, governance, scientific value and benefit-sharing. A history that does not flatten its subjects Colonialism and race should not become an all-purpose explanation that removes the agency of patients, healers,
Nuremberg Revisited 24 community leaders, or local researchers. People accepted interventions for many reasons: hope of treatment, trust, practical need, curiosity, payment, social obligation, or a desire to help future patients. Ethical analysis must take those reasons seriously. It must also recognise that choice can be narrowed by dependency, poverty, political coercion, or lack of alternatives. Respect for agency means neither assuming total passivity nor overlooking structural constraints. [6, 7] This chapter therefore treats inequality as a context to investigate, not a shortcut to judgement. Each case requires attention to the procedure performed, the knowledge available at the time, the mechanism of recruitment, the possibility of refusal, the risk imposed, independent oversight, and the allocation of benefit. Where the documentary record is disputed, the disagreement should be described rather than resolved by rhetoric. Those habits of evidence are essential to a book about accountability: a case made without care for facts can be easily dismissed, while a case that distinguishes certainty from uncertainty is harder to evade.
Nuremberg Revisited 25 CHAPTER 4 Eugenics, Forced Sterilisation, and Scientific Racism The Promise of Improving the Population Eugenics emerged in the late nineteenth century as a movement claiming that human populations could be improved by managing reproduction. Francis Galton coined the term and promoted the study of inherited human characteristics. In the early twentieth century, eugenic ideas gained support in parts of Britain, the United States, Germany, and other industrialised countries. They were promoted by scientists, physicians, reformers, and officials who believed that social problems such as poverty, disability, criminality, and mental illness could be addressed through selective reproduction. Many claims rested on weak heredity models, social prejudice, or an unjustified conversion of complex social conditions into supposedly fixed biological traits. [9] Eugenics was not a single, uniform programme. It included positive measures encouraging some people to reproduce and negative measures designed to restrict the reproduction of those labelled “unfit.” The latter included marriage restrictions, segregation, and compulsory sterilisation. The ethical danger lay in combining an apparently scientific classification with state authority over intimate bodily decisions. A medical label could become a legal instrument: once a person was categorised as genetically undesirable, officials and courts might claim the power to prevent that person from having children. [9, 10]
Nuremberg Revisited 26 Eugenic policies were often justified as efficient, preventive, and economically prudent. They treated the future population as an object of administration and presented coercion as social improvement. The use of technical language did not make the underlying values neutral. Judgements about whose lives were worth living, whose families were desirable, and whose reproduction should be prevented were shaped by class, disability stigma, racism, sexism, and ideas about social conformity. Research and statistics sometimes gave these judgements the appearance of objective fact, even when the methods were incapable of supporting them. [9] The United States and the authority of law Compulsory sterilisation programmes developed in several jurisdictions before and during the twentieth century. In the United States, the Supreme Court’s 1927 decision in Buck v. Bell upheld Virginia’s sterilisation statute. The case involved Carrie Buck, a young woman committed to an institution, whose reproductive rights were overridden by a legal system that accepted eugenic assumptions. The decision became a deeply troubling symbol of the way judicial authority can legitimate state control over the body. [11] The United States Holocaust Memorial Museum notes that eugenic ideas were widespread internationally and that US sterilisation laws were part of the context followed by Nazi racial- hygiene policies. This does not imply that American and Nazi programmes were identical in scale, purpose or consequences. It does mean that Nazi eugenics did not develop in an intellectual vacuum: foreign examples were studied, and some of the same assumptions about heredity, disability and social worth circulated across borders. [9] Compulsory sterilisation programmes disproportionately targeted people with disabilities, people living in institutions, poor people and racialised minorities. The categories and numbers varied by country and period, and estimates must be tied to a specific
Nuremberg Revisited 27 source and definition. For publication, totals should not be compared unless the periods, jurisdictions and forms of coercion being counted are consistent. The underlying ethical point does not depend on a single number: people were denied reproductive autonomy on the basis of social judgements presented as medical necessity. [9, 10, 11] Nazi racial hygiene and the medical profession After 1933, the Nazi regime rapidly translated eugenic and antisemitic ideas into state policy. The Law for the Prevention of Hereditarily Diseased Offspring established compulsory sterilisation for specified categories. The United States Holocaust Memorial Museum estimates that approximately 400,000 people were forcibly sterilised in Nazi Germany. Such procedures were enforced through a combination of medical diagnosis, special courts, bureaucratic procedures and police power. The appearance of due process did not make the policy just; professional and legal structures helped operationalise persecution. [9, 10] Eugenics also shaped Nazi policies that classified disabled people as burdens on the state. The clandestine killing programme commonly known as Aktion T4 murdered institutionalised patients. This programme was not itself a human research trial and should not be categorised as one. It belongs in this chapter because it demonstrates how medicine and public administration could be mobilised to decide which lives deserved protection. The institutional and professional culture that accepted coercive selection and killing provides essential context for the medical atrocities explored in the next part of this book. [9] At concentration camps, physicians also conducted forced experiments, including sterilisation experiments on prisoners. These practices crossed from coercive policy into the direct use of captive human beings as experimental material. The later Doctors’ Trial documented how the absence of voluntary consent, the imposition of severe suffering and the lack of a defensible risk- benefit relationship could not be excused by the state’s authority
Nuremberg Revisited 28 or by claims of scientific ambition. The Nuremberg Code would make voluntary consent its first principle. [3, 10] The enduring warning Eugenics is not merely a historical story about false ideas that were eventually abandoned. Its deeper warning concerns the power of institutions to translate contested biological claims into coercive social policy. A scientific discipline can be respected, a law properly enacted and a procedure technically performed, yet the overall system can still violate human dignity. Ethics must therefore examine not only whether a method works but also the ends to which it is put, the populations it targets and the safeguards available to challenge it. [9, 10, 11] The history also cautions against using present-day genetics as a licence to revive old social hierarchies in new language. Modern genetic science has genuine clinical and public-health value, but complex traits and social outcomes cannot be reduced casually to inherited “fitness.” Contemporary genomics demands careful interpretation, respect for privacy, non-discrimination and protection against coercive use. These issues return in Part VII, where the book examines genetic data, artificial intelligence, biobanks and emerging technologies. The chapters that follow move into the Second World War, when professional complicity, racial ideology and state violence produced some of the clearest documented violations in the history of human experimentation. The transition should not be read as implying that atrocities began in 1933 or ended in 1945. Rather, the war exposed in devastating form what can happen when consent and dignity are treated as dispensable, and it helped create the conditions for an international effort to define what research must never do to the people whose participation makes it possible.
Nuremberg Revisited 29 PART II THE SECOND WORLD WAR AND THE BIRTH OF MODERN RESEARCH ETHICS Chapters 5–8
Nuremberg Revisited 30 CHAPTER 5 Medicine Under the Nazi Regime When a profession accepts the state’s definition of whose life has value, its technical competence can become an instrument of persecution. The corruption of a healing profession The medical crimes committed under National Socialism cannot be understood as a sudden departure from an otherwise untouched profession. They arose through the convergence of racial ideology, state coercion, professional ambition, bureaucratic obedience and a scientific culture in which some people were increasingly regarded as biological material rather than as persons. The Nazi state did not abolish medicine. It recruited medical expertise, redirected it and gave physicians authority over policies that violated the people medicine was supposed to protect. The resulting history is disturbing precisely because it involved trained professionals working through institutions that retained the appearance of scientific and administrative legitimacy. [13, 14] From 1933, Nazi policies increasingly classified human beings according to purported racial and hereditary worth. Jewish people, Roma and Sinti, disabled people, and others marked as undesirable were subjected to exclusion, coercion and violence. The language of racial hygiene and eugenics gave political prejudice a scientific vocabulary. Medical research was not the only arena in which this happened, but biological claims helped
Nuremberg Revisited 31 make persecution appear as public health, social improvement or national protection. [13, 14] This chapter is not an argument that every physician in Germany participated in these crimes or that medicine as a discipline inevitably leads to abuse. It is an account of how professional authority can be captured by an authoritarian system, and how ordinary mechanisms—appointments, research grants, hospital administration, military requirements, professional networks and official orders—can normalise conduct that should have remained unthinkable. From exclusion to coercive medical policy The Nazi regime's programme of forced sterilisation preceded the concentration-camp experiments that later became emblematic of medical atrocity. The 1933 Law for the Prevention of Hereditarily Diseased Offspring created a legal mechanism for compulsory sterilisation of people considered to have hereditary conditions. Special courts decided cases, often on the basis of diagnoses and assumptions that reflected the racial and eugenic priorities of the state. The lesson is important: legal authorisation and medical participation do not by themselves establish ethical legitimacy. A law can institutionalise injustice, and a diagnosis can be turned from a tool of care into a rationale for coercion. [13, 14] During the war, the state expanded from coercive sterilisation to the systematic killing of people with physical and mental disabilities. The programme commonly known as Aktion T4 involved the organised murder of institutionalised patients. It was not simply a research project, and it must not be collapsed into the category of experimentation. It is nevertheless essential context for the later medical crimes: physicians, administrators and institutions participated in a system that decided some lives were unworthy of protection. Techniques, personnel and administrative practices developed in this setting also intersected with the wider machinery of persecution and mass murder. [13, 14]
Nuremberg Revisited 32 The boundary between medicine and state violence was not crossed in one single moment. It was weakened by a sequence of decisions: discriminatory rules were accepted; patients lost rights; consent became irrelevant for targeted groups; clinicians took part in selection and classification; and bureaucratic routine hid the consequences behind euphemism. A case file, a clinical examination or a transport list could be presented as ordinary administration while contributing to an apparatus of harm. The researcher, the military and the state Some Nazi medical experiments were explicitly designed to address military problems: survival at altitude, exposure to cold, seawater consumption, infectious disease and the treatment of battlefield wounds. Such stated aims did not make the experiments ethically acceptable. Military medicine routinely studies dangerous conditions, but legitimate military research still requires protections for participants, a defensible scientific design and limits on risk. The prisoners at the camps were not service members volunteering for a clearly explained study. They were captive people who could not freely refuse, whose welfare was not the controlling consideration and whose suffering was treated as expendable. [13, 15] The professional setting matters. Researchers could describe a study as urgent, patriotic or beneficial to future patients. Such words may carry genuine meaning in ethical research, but they can also function as moral camouflage. A claim that a project might save many lives in the future cannot, by itself, erase the rights of the people deliberately endangered in the present. Nor can research value be assumed merely because an experiment was conducted by a physician, funded by a government or recorded in technical language. Race as a research instrument Other experiments were intended to advance Nazi racial ideology. The most notorious example is Josef Mengele's work at Auschwitz, including experiments involving twins and people
Nuremberg Revisited 33 selected because of perceived racial or physical characteristics. The procedures were performed without meaningful consent and in circumstances in which death or permanent injury could result. In some instances, death was an intended or accepted consequence. These activities did not represent ethically flawed versions of normal clinical research; they were part of a system that treated prisoners as objects. [13, 16] The racial selection of subjects is a central ethical fact. The victims were not recruited because they had freely chosen to accept a scientifically reasonable risk. They were exposed because Nazi ideology had stripped them of equal moral status. This violated not only consent but also the principle that all human beings have an equal claim to bodily integrity and protection. It also compromised scientific judgement: when a research question is driven by pseudoscientific racial premises, the entire project may be methodologically and morally corrupted from its inception. Responsibility beyond the individual It is tempting to describe these events as the work of a few monstrous individuals. Individual responsibility was real, but the institutional question is larger. Research is rarely the product of one person alone. It requires access to participants, facilities, administrative authorisation, staff, records, supplies, and channels through which findings are communicated. Atrocities of this scale therefore raise questions about professional associations, universities, military and government agencies, camp administrations, and colleagues who enabled, tolerated, or failed to challenge the work. [13, 14] The subsequent prosecution of a number of physicians was consequential, but it did not mean that every participant in Nazi medical crimes was tried. The US Holocaust Memorial Museum notes that some perpetrators were prosecuted while others escaped justice, including Mengele, who fled to South America and died in Brazil in 1979. Survivors also endured lasting physical and psychological consequences, including fear and distrust of
Nuremberg Revisited 34 medical care. [13, 16] The aftermath therefore involved more than punishment: it involved damaged communities, a breached relationship between medicine and patient, and an enduring question about how a profession should respond when its own institutions have betrayed their purpose. Lessons for research ethics The history of Nazi medicine helps identify several failure modes that remain relevant. First, consent is meaningless where refusal is not possible. Second, a scientific objective cannot justify every means used to pursue it. Third, a government order does not cancel professional responsibility. Fourth, apparently routine administrative procedures can become mechanisms of abuse when the subjects of research are denied rights. Finally, the selection of people considered socially or biologically inferior is an early warning that research is serving ideology rather than human welfare. These lessons should not be used as rhetorical shortcuts. A contemporary study is not equivalent to Nazi experimentation simply because it involves risk, government sponsorship or a disputed policy. Historical comparison requires care. But the core question survives across contexts: are people regarded as rights- bearing participants whose welfare constrains the research, or as resources to be used for an objective defined by someone else? That question leads directly to the experiments themselves and then to the trial that attempted to place ethical limits on them.
Nuremberg Revisited 35 CHAPTER 6 Concentration-Camp Experiments: Science Without Consent The defining feature was not simply suffering. It was the deliberate removal of the victim’s right to decide whether the suffering would be imposed. A catalogue of experiments—and a necessary caution Medical experiments were conducted in several Nazi concentration camps. The US Holocaust Memorial Museum groups them broadly into three categories: experiments intended to improve the survival of military personnel; experiments to develop drugs and treatments; and experiments intended to advance Nazi racial and ideological objectives. The categories overlapped. In each, the fundamental ethical failure was the use of people imprisoned by the regime without voluntary consent or adequate protection. [14] The historical record does not justify treating all projects as identical. They differed in aim, procedure, duration, personnel, and consequences. Some were designed around military survival, others around infectious disease, drug testing, surgery, or sterilisation. Their shared ethical character lies in coercion, disregard for the individual, a willingness to impose severe risks, and the absence of safeguards that could protect a participant from the investigator’s decisions.
Nuremberg Revisited 36 High-altitude experiments at Dachau Beginning in 1942, physicians associated with the German air force and aviation research institutions conducted high-altitude experiments at Dachau. Using a low-pressure chamber, they simulated conditions at altitude in order to investigate what might happen to aircrew forced to parachute from damaged aircraft. Prisoners were exposed to physiological conditions that could lead to severe injury or death. [14, 15] The military objective was intelligible: air forces wanted to understand the limits of human survival. But scientific relevance does not settle the ethical question. The men in the chamber were not volunteers who had been given information and a meaningful choice. Their incarceration created an absolute power imbalance. They could not leave, negotiate the risk or demand that the study stop. Their bodies became a means to answer a question posed by the state and the military. The case also challenges a persistent misconception: that technically sophisticated equipment or a measurable physiological endpoint makes a study scientific in the ethical sense. A device can produce data while the experiment remains an abuse. Ethical research must ask not only whether the instrument measures something, but whether the knowledge sought is necessary, the design is defensible and the people exposed to the procedure have rights that constrain the researcher. Freezing and hypothermia experiments At Dachau, researchers also conducted experiments involving extreme cold and hypothermia, including exposure to cold water. They sought information relevant to German airmen who might survive a crash into icy water. Prisoners endured dangerous cold stress, and some died. The experiments became a later point of debate because questions arose about whether the resulting data might have any scientific value. [14, 17] That debate exposes a second ethical problem: what should be done with data obtained through an atrocity? There is no simple
Nuremberg Revisited 37 answer that removes the need for careful judgement. Some argue that using such data risks legitimising the methods, degrading the victims a second time, or giving an undeserved aura of validity to poor-quality work. Others argue that limited information might, in exceptional circumstances, help prevent future deaths. Robert L. Berger's historical analysis in the New England Journal of Medicine documented the controversy over the Dachau hypothermia studies and the disagreement among scientists about whether their results should be used. [17] The argument should not begin from the assumption that criminally obtained data are either automatically worthless or automatically useful. The quality of the underlying methods, the possibility of independent confirmation, the necessity of using the material and the risk of normalising the abuse all matter. In this particular history, the ethical wrong is not erased by later debate about data utility. Even if some observation were accurate, accuracy would not retrospectively create consent or justify the means by which it was obtained. Seawater, infection and drug experiments Prisoners at Dachau were also used in studies designed to test ways of making seawater potable. Across several camps, researchers conducted experiments involving infectious diseases and possible treatments. At Buchenwald, for example, physicians and scientists tested vaccines and treatments for diseases including typhus, typhoid, cholera and diphtheria; the US Holocaust Memorial Museum records that these experiments caused hundreds of deaths. [14, 18] At Ravensbrück, women prisoners were subjected to experiments involving wounds, bone grafting, and sulfonamide drugs. Victims known as the 'rabbits' were subjected to procedures that produced severe injury and infection. Photographs taken clandestinely by prisoners became evidence of the experiments and a means by which the victims attempted to document what had been done to them. [14]
Nuremberg Revisited 38 The use of drugs in a camp did not make the activity a therapeutic trial in the modern sense. A clinical trial seeks a valid answer while protecting participants and respecting their rights. Here, prisoners were deliberately exposed to injury or disease, and the research process did not operate within a system of voluntary participation, independent oversight, or effective safety monitoring. The distinction between treatment and experimentation was itself compromised: a person who cannot refuse and has no control over the procedure is not made a voluntary research participant because a drug is administered afterward. Sterilisation experiments The regime pursued methods of mass sterilisation for racial and ideological reasons. At Auschwitz, gynaecologist Carl Clauberg conducted experiments intended to develop non-surgical methods of sterilisation. The US Holocaust Memorial Museum records that he used toxins injected into the fallopian tubes of several hundred women, many of them Jewish; the procedures could cause severe pain, sepsis, organ failure, and death. Horst Schumann used X- rays in forced sterilisation experiments on Jewish prisoners, causing burns, disfigurement, and other injuries. [10] These were not merely studies with a disproportionate risk. The objective itself was connected to a policy of coercive population control. The subjects were chosen because the regime regarded their reproduction as something it had the right to prevent. Their reproductive autonomy was not simply ignored during research; its destruction was the point of the programme. This distinction matters for research ethics today. A study can be unethical because its procedures are unsafe or its consent inadequate. It can also be unethical at a more fundamental level because the purpose is to violate the rights of a targeted group. Evaluating only the technical procedure and ignoring the political purpose would give an incomplete account.
Nuremberg Revisited 39 Mengele and experiments driven by racial ideology At Auschwitz, Josef Mengele selected twins and other prisoners for experiments informed by Nazi racial theory. The US Holocaust Memorial Museum describes experiments involving disease, invasive procedures, sterilisation, unnecessary surgery and the killing and dissection of prisoners. In some cases, death was intended or knowingly accepted as part of the procedure. No meaningful consent was sought. [16] Mengele's notoriety can obscure the institutional setting in which his work occurred. He was not operating in a world where prisoners possessed recognised rights but one rogue doctor ignored them. The camp system had already stripped the victims of their freedom and placed them under a regime of racial persecution and violence. Medical experimentation was one element in a much broader system that denied them personhood. The racial premise also mattered scientifically. A study organised to substantiate a political doctrine is vulnerable to confirmation bias and false inference. Even where measurements are recorded, the coercion and ideological purpose raise doubts about the questions asked, the interpretation of results and the scientific claims made. Ethical failure and scientific unreliability need not always be identical, but under Nazi racial medicine they frequently reinforced one another. How should these experiments be assessed? Four questions should be kept distinct. What happened to the participants? What did investigators intend to learn? What evidence supports the reported procedure and outcome? And what, if anything, can responsibly be inferred from the resulting data? These are separate questions. A historical account should not exaggerate details for dramatic effect, but neither should it soften
Nuremberg Revisited 40 a documented violation through euphemism or technical language. The evidence is drawn from trial material, survivor testimony, camp records, photographs, institutional histories, and later scholarship. Each source has limits. Some records were destroyed; testimony may be incomplete; casualty estimates may vary; and not every experiment has been reconstructed in equal detail. The responsible response is to make uncertainty visible where it exists while recognising the overwhelming documentary record of systematic non-consensual experimentation. [13, 14] The most important lesson is that no study can be ethically defended solely by pointing to the importance of the question. Consent, the possibility of refusal, proportionality of risk, protection against harm, and the investigator's duty to stop are not decorative requirements added after a protocol is designed. They are part of what makes research morally legitimate. The Nuremberg Code would later express these obligations in unusually direct language.
Nuremberg Revisited 41 CHAPTER 7 Unit 731: Biological Warfare and the Limits of Post-war Accountability The history of Unit 731 is also a history of evidence: crimes committed in secrecy, records scattered across countries, and post-war decisions that shaped what could be known and who would be held responsible. A secret programme within the Japanese military During Japan's war in China and the wider Second World War, military research units investigated biological warfare. The best- known was Unit 731, associated with Japanese Army physician Shirō Ishii and based in the region of Harbin in Japanese- occupied Manchuria. Human beings were used in experiments connected with infectious disease, physiological response and biological warfare. Victims included Chinese, Korean and Russian people. The US National Archives' historical collection on Japanese war crimes describes the archival record and the difficult history of documenting these activities. [19, 23] The secrecy surrounding the programme, the collapse of the Japanese empire, the destruction or dispersal of records and the politics of the early Cold War all affected the historical record. It is therefore necessary to approach Unit 731 with both moral clarity and evidentiary discipline. The existence of human experimentation is well documented; the exact scale of particular activities, responsibility in particular incidents and some claims about specific groups of prisoners have varied levels of certainty.
Nuremberg Revisited 42 Human beings as material for biological warfare research The research programme was not a conventional clinical investigation designed primarily to improve the health of participants. Its operational context was military biological warfare. Historical research has documented deliberate infection and experiments on live human beings, together with studies concerning cold exposure and other physiological effects. The National Archives review notes that wartime publications by Ishii and other members, combined with post-war records and scholarship, provided evidence of human experimentation. [19] The ethical violations were severe. Captives could not freely refuse. Their vulnerability was not incidental to recruitment: confinement and military power made experimentation possible. The distinction between a participant and a target disappeared. In a legitimate clinical investigation, the person enrolled remains a rights-bearing individual whose welfare constrains the project. In the Unit 731 setting, the human subject was treated as material from which information could be extracted for military purposes. The experiments also raise the issue of dual-use research: scientific knowledge that may be applied to both protective and harmful ends. Such work demands special scrutiny even when conducted ethically, because findings about pathogens, susceptibility or transmission may support public health or facilitate biological attack. Unit 731 went far beyond the ordinary ethical challenges of dual-use research; its human experimentation was tied to coercion and military programmes that violated basic human rights. What the records establish—and what they do not Post-war understanding of Unit 731 has developed through Japanese, Chinese, Russian and US documents; testimony; scholarship; and declassified intelligence records. The US National Archives' volume on Japanese war crimes describes the recovery of important records and the uneven path by which the
Nuremberg Revisited 43 history became better known. It also stresses that some questions remain unresolved. [19] One disputed issue concerns whether Allied prisoners of war held at Mukden, also called Hoten, were used as experimental subjects. The National Archives' historical essay notes that some writers and former prisoners have argued that Unit 731 medical personnel conducted experiments there, while US agencies reported finding no definitive evidence. The essay also records scholarly views that the visit may have been intended to address hygiene and illness rather than conduct experiments. This question should therefore be presented as unresolved rather than asserted as settled fact. [19] Similarly, estimates of the total number of people killed through experimentation, biological attacks and related activities vary across publications. A manuscript should state which estimate it uses, what it counts, which sources support it and why other estimates differ. Inflated or untraceable numbers can undermine a history that is otherwise well founded. Precision is not a concession to perpetrators; it is a form of respect for the historical record and for the victims. The post-war bargain: research data and immunity One of the most consequential questions concerns the treatment of Unit 731 personnel after Japan's surrender. The National Archives' historical study describes US investigations of Ishii and his associates and records evidence that some US officials sought information about the biological warfare programme while considering protection from prosecution. The essay discusses correspondence and records concerning efforts to discourage prosecution of prominent members, and it notes that some documents about the transfer of data have not been located. [19] This record requires careful wording. It supports the conclusion that US authorities considered access to the data valuable and that efforts were made to protect members of the programme from prosecution. It does not mean that every detail of a supposed
Nuremberg Revisited 44 formal bargain is established by a single, complete contract. The archival picture is incomplete, and different aspects of the exchange have been reconstructed from multiple records. The book should explain both the available evidence and the gaps. The moral conflict is stark. Intelligence and military officials could regard the results as strategically valuable, particularly amid rising Cold War tensions. Yet the data had been produced through severe abuse. A decision to acquire such information while shielding those responsible would prioritise strategic utility over accountability and could deny victims the recognition and justice that prosecution might have provided. It would also create an incentive for future perpetrators to believe that information of sufficient military value might protect them from punishment. Unit 731 therefore belongs in a history of research ethics for two reasons. First, for the original crimes committed against human subjects. Second, for the post-war decisions that shaped accountability and the potential use of the resulting data. The history did not end when the laboratory doors closed or when the war ended. It continued through decisions about archives, prosecution, intelligence and public memory. Who was responsible? Responsibility cannot be reduced to the individual who carried out an experiment. A military programme on this scale requires command structures, facilities, logistics, human recruitment or capture, scientific staff and institutional support. The National Archives' review records debate about the extent of responsibility among military leaders and the broader medical establishment. It also notes criticism that some people associated with the programme later occupied influential professional positions in post-war Japan. [19] The challenge is to identify responsibility at the level supported by evidence. A document showing that an official knew a programme existed does not automatically establish that the official knew every detail of a specific experiment. Conversely, the absence of a
Nuremberg Revisited 45 signed order does not prove that no institutional responsibility existed. Historical investigation must consider patterns of authorisation, resources, reporting, direct testimony, contemporaneous records and the reasonable knowledge of those in command. Lessons: secrecy is not a substitute for ethics Unit 731 illustrates how secrecy can weaken every ordinary safeguard. Participants cannot assess risks when a programme is concealed from them. Independent reviewers cannot scrutinise protocols they cannot access. Journalists, courts and civil society cannot examine claims that remain classified. After the event, destroyed or withheld records can prevent families and communities from learning what happened. Secrecy can also make it easier for institutions to define the programme only through the information they find useful. The appropriate lesson is not that every classified research programme is unethical. Military medicine can legitimately protect sensitive operational information. The lesson is that secrecy must not remove ethical limits, independent authority, records retention or accountability. Where research involves people, a national-security claim cannot itself prove that their rights were protected. The story of Unit 731 is a warning against a utilitarian argument without boundaries: that information might save more lives than it cost. Such arguments become especially dangerous when the people bearing the costs are prisoners, civilians or members of a population deemed expendable. The next chapter returns to the tribunal that attempted to define a different relationship between science and the human subject—and to the limits of that effort.
Nuremberg Revisited 46 CHAPTER 8 The Doctors’ Trial and the Nuremberg Code The law could punish some perpetrators. The Code attempted something wider: to state what a researcher must never forget about the person in front of them. From war crimes to medical responsibility After the Second World War, the Allies established proceedings to prosecute people responsible for war crimes and crimes against humanity. The Doctors' Trial was one of the subsequent Nuremberg proceedings conducted by an American military tribunal. It opened on 9 December 1946 as United States v. Karl Brandt et al. and involved 23 leading German physicians and administrators. The case covered concentration-camp experimentation and the killing of people with physical and mental disabilities under the Nazi regime. [20, 22] This proceeding must be distinguished from the principal International Military Tribunal, which tried prominent Nazi leaders in 1945–1946. The Doctors' Trial was a subsequent trial focused on medical crimes and related conduct. The US Holocaust Memorial Museum records that the proceedings lasted almost 140 days, included testimony from 85 witnesses and approximately 1,500 documents, and culminated in a verdict on 20 August 1947. Sixteen defendants were found guilty; seven were sentenced to death and executed on 2 June 1948. [20]
Nuremberg Revisited 47 The case forced the tribunal to confront a question that remains central to research ethics: what rules should govern experimentation on human beings, particularly where the claimed objective is scientific knowledge or public benefit? Some defendants argued that no clear international rules distinguished permissible from impermissible human experimentation and compared their work with research conducted in other countries before the war. [21] That argument did not establish that the experiments were ethical. Instead, it exposed a gap: medicine had long relied on professional norms and local rules, but there was no single internationally influential statement laying out a concise set of principles for experimentation on human subjects. The tribunal's response became one of the most enduring consequences of the case. How the ten principles emerged During the trial, two American physicians, Andrew Ivy and Leo Alexander, contributed to the debate on permissible research. On 17 April 1947, Alexander submitted a memorandum outlining principles for legitimate medical research. The tribunal's judgment of 19 August 1947 included a section entitled 'Permissible Medical Experiments'. It set out ten points that became known as the Nuremberg Code. The verdict followed on 20 August. The US Holocaust Memorial Museum and Yale Law School's Avalon Project reproduce the Code and place it in the context of the Doctors' Trial. [20, 21] The Code was not merely a demand for better technique or more careful record-keeping. It began with the proposition that voluntary consent is absolutely essential. It also addressed scientific necessity, preparation, the avoidance of needless suffering, proportionality of risk, qualified investigators, the subject's freedom to withdraw and the investigator's duty to stop an experiment when continuation could cause injury, disability or death. [21]
Nuremberg Revisited 48 The wording is unusually direct because it places responsibility on the investigator. The duty to establish the quality of consent belongs personally to each person who initiates, directs or engages in the experiment. It cannot be discharged by saying that a form was signed, a subordinate obtained permission or an institution approved the protocol. Consent must be meaningful, and those conducting the work remain responsible for ensuring that it is so. The ten principles in practical language First, consent must be voluntary and informed. Participants need the capacity and freedom to decide, without force, fraud, deceit, duress or other coercion. They must have adequate information about purpose, duration, methods, hazards and possible effects. This requirement is not satisfied by a signature if the person does not understand the study or cannot genuinely refuse. [21] Second, the research should seek socially valuable results that cannot be obtained by other means and should not be random or unnecessary. Third, it should be grounded in prior knowledge and appropriate scientific preparation. These points resist research undertaken merely because a group of people is available for study. Scientific value is a condition of proceeding, not an excuse that overrides every other obligation. [21] Fourth, the experiment must avoid unnecessary physical and mental suffering. Fifth, the Code says research should not be conducted where there is prior reason to expect death or disabling injury, subject to a narrowly worded exception in the original text. Sixth, risk must not exceed what is justified by the humanitarian importance of the problem. Seventh, adequate preparations and facilities must protect the participant from injury, disability or death. [21] Eighth, only scientifically qualified people should conduct the work, with the highest degree of skill and care.
Nuremberg Revisited 49 Ninth, the human subject must be free to end participation when continuation seems physically or mentally impossible. Tenth, the investigator must be ready to terminate the experiment if there is probable cause to believe that continuation could cause injury, disability or death. The last two principles are particularly important because they make protection dynamic. Ethical responsibility does not end when a participant enters a study; it continues throughout the experiment. [21] What the Code did—and did not—settle The Code became a landmark because it stated basic requirements in a form that could be remembered and invoked. But it was not a complete modern research-regulation system. It was not a standalone international treaty, and its legal force and direct application across jurisdictions have been debated. Its moral and historical influence is nevertheless extensive. The US Holocaust Memorial Museum describes it as one of the most lasting products of the Doctors' Trial and notes that it informed later statements on medical ethics. [20] The Code focused strongly on individual consent and the responsibilities of the investigator. Later frameworks developed broader systems for research oversight, including independent review, risk-benefit assessment, additional protections for vulnerable groups, continuing scrutiny and principles for international research. The Declaration of Helsinki, adopted by the World Medical Association in 1964, and later documents such as the Belmont Report and CIOMS guidelines expanded the ethical architecture. The later frameworks should not be presented as if they were fully contained in the 1947 Code; they responded to subsequent experience and changing research environments. Nor did the Code end unethical human research. Investigations into post-war and Cold War programmes, including US government experiments, demonstrate that ethical declarations
Nuremberg Revisited 50 can coexist with institutional failures. A code cannot protect participants unless researchers follow it, sponsors do not undermine it, regulators enforce applicable requirements and records permit independent examination. Nuremberg provided an essential statement of principle, not a guarantee of compliance. The use—and misuse—of Nuremberg in contemporary debate The Code is sometimes invoked in modern arguments about public health, vaccination, emergency powers and government policy. Its relevance depends on precise classification. A clinical trial is research involving human subjects; a public-health mandate is ordinarily a policy intervention; a marketing authorisation is a regulatory decision; and routine medical care is a different category again. These categories may raise overlapping ethical questions, but they are not interchangeable. The Nuremberg Code's first principle is directly relevant when a person is enrolled in an experiment: consent must be voluntary and informed. Its broader values—proportionality, harm reduction and the investigator's duty to stop unsafe research—also influence the ethical analysis of research programmes. But the Code alone does not decide every legal or ethical question concerning public-health policy. Those questions require examination of the applicable law, evidence available at the time, the nature of the intervention, proportionality, available alternatives and the rights affected. Similarly, an adverse event does not by itself prove unethical research, just as a beneficial outcome does not prove that a study was ethically sound. Research ethics is concerned with design, consent, risk, safeguards, data integrity, fairness and accountability as well as outcomes. A careful history must resist using the Code either as a slogan for every objection or as a ceremonial document invoked only after the facts have been settled.
Nuremberg Revisited 51 A code must be carried by institutions and people The most enduring contribution of Nuremberg is the insistence that the participant is not merely a means to an end. The Code makes the investigator responsible for consent, the researcher responsible for scientific necessity and the person in charge responsible for stopping dangerous work. These duties apply at the moment a study is planned, when participants are recruited, while the procedure is underway and when unexpected harm emerges. The Doctors' Trial was a criminal proceeding about past conduct; the Nuremberg Code was an attempt to define principles for future conduct. The distinction matters. Prosecution can establish responsibility and impose punishment, but it cannot by itself create an everyday culture of ethical research. That requires professional training, independent ethics committees, transparent protocols, adverse-event reporting, protection from retaliation, whistle-blower pathways, public access to results and remedies for people harmed by research. The story of the Code is therefore neither a story of complete moral progress nor one of futile idealism. It records an important moment when the world articulated a clear limit: a human being cannot be forced into an experiment merely because the investigator believes the result could be valuable. The subsequent history of research ethics can be read as an ongoing attempt to make that limit effective in institutions that possess money, authority, scientific prestige and political power. Questions for the reader 1. Does a socially valuable aim ever justify imposing a serious experimental risk on a person who has not consented? 2. Can consent be voluntary when the participant depends on the researcher, institution, prison authority, employer or military chain of command?
Nuremberg Revisited 52 3. What independent evidence should be required before ethically compromised data are used in clinical guidance or later research? 4. What should happen when a state or institution decides that scientific or strategic value outweighs the rights of participants? 5. What mechanisms, beyond a code of principles, are needed to ensure that ethical rules are obeyed?
Nuremberg Revisited 53 PART III AFTER NUREMBERG: VIOLATIONS IN MODERN SCIENCE Chapters 9–14
Nuremberg Revisited 54 CHAPTER 9 Tuskegee: Four Decades of Deception The gravest betrayal was not that the disease was studied. It was that the men were denied the information and treatment they needed to make decisions about their own lives. A study of untreated disease In 1932, the United States Public Health Service, working with the Tuskegee Institute in Alabama, began a study intended to document the natural history of untreated syphilis in Black men. The study enrolled 600 men: 399 who had syphilis and 201 who did not. Participants were not given adequate information about the study. Researchers described their medical problem using the local expression 'bad blood', a phrase that could refer to several conditions, and offered examinations, meals and burial insurance. The men were not purposefully infected for the study; those in the syphilis group already had the disease when recruited. This distinction is important because the actual ethical violations are grave enough without adding an inaccurate claim. [24, 25] The study was framed as an observational investigation of the natural course of infection. But the participants were not simply people who happened to be observed while receiving ordinary care. Their understanding of what was being studied was deficient, and they were not asked for voluntary informed consent in the sense now recognised as a central requirement of ethical research. The arrangement created an asymmetry: federal researchers possessed the scientific purpose and relevant
Nuremberg Revisited 55 information, while the men were encouraged to rely on the institution without being told the full nature of their involvement. [24] The racial context was inseparable from the study's ethical significance. The participants were poor Black men in the segregated American South. Their economic and social circumstances limited their choices, while institutional authority shaped how the offer of medical examinations and support was experienced. The wrong was not merely a failure to use a modern consent form. The study took place in a society where racial inequality could influence whose health mattered, who received information and who was expected to accept professional authority without question. When effective treatment became available The ethical position changed sharply as penicillin became accepted as the treatment of choice for syphilis in the 1940s and was widely available by the 1950s. The study continued. The men were not offered the available treatment through the study, and steps were taken that obstructed some opportunities for treatment. A research question about the natural history of an untreated disease was allowed to take precedence over the interests of the people living with it. The CDC's historical account specifically states that participants were not offered treatment even after it became readily available. [24, 25] This is the core of Tuskegee. The men were not deliberately infected by investigators, but they were misled about the nature of the project and were not offered penicillin as treatment while the study continued. The difference between deliberate infection and withholding effective treatment matters historically, scientifically and ethically. Precision strengthens the case rather than weakening it: the documented conduct already constitutes a profound breach of trust and participant protection. [24] The harm extended beyond the direct physical effects on the men. Untreated syphilis could have consequences for health,
Nuremberg Revisited 56 relationships and families. The study also created the possibility of transmission to partners. The CDC records that some women contracted syphilis from men in the study's infected group, and the ethical consequences therefore reached beyond the people initially enrolled. A protocol that treats participants only as sources of longitudinal data can fail to account for the people connected to their lives. [25] Exposure, whistleblowing and the end of the study Questions about the ethics of the study were raised before its eventual termination. Public Health Service employee Peter Buxtun challenged the continuation of the project in the late 1960s. The study only came to a decisive public crisis in 1972, when Associated Press reporting exposed the arrangement. An advisory panel reviewed the study and concluded that it was ethically unjustified because the results were disproportionately meagre compared with the known risks to participants. The study ended in 1972. [24] That sequence raises a critical point about ethical oversight. A system cannot be considered effective merely because it has a formal hierarchy or because individual staff members raise concerns internally. What matters is whether concerns receive a prompt, independent response and whether participant welfare takes precedence over institutional investment in a project. The persistence of Tuskegee for four decades was not only a failure of the initial design; it was a failure to correct course when the scientific and ethical circumstances changed. [24, 26] After termination, legal and public accountability followed. A class-action lawsuit resulted in a 1974 out-of-court settlement of $10 million. A health-benefit programme was established for survivors and later extended to certain family members. In 1997, President Bill Clinton issued a formal presidential apology. These measures acknowledged harm and provided assistance, but they could not restore the years in which participants were denied information and treatment. [24]
Nuremberg Revisited 57 Scientific value and ethical legitimacy A defence sometimes made of observational research is that knowledge of natural history can help medicine understand a disease. That proposition is generally true: natural-history studies, registries and long-term cohorts can be valuable. But scientific value does not make every way of obtaining knowledge permissible. A legitimate observational study does not require deception about the central purpose, and ethical research cannot proceed by deliberately leaving participants without treatment that is clinically indicated merely to preserve an untreated comparison. Once effective therapy exists, the burden of justification is not satisfied by pointing to the scientific interest in observing the consequences of non-treatment. [24, 26] Tuskegee illustrates the danger of allowing the research question to become more important than the participants. A project can maintain a scientific vocabulary, collect follow-up data and produce publications while nevertheless failing the basic obligations of medicine. In research, clinical expertise does not displace the need for consent. Nor does the existence of a control group remove the obligation to consider what is owed to each person in that group. Methodological convenience cannot be allowed to define the limits of ethical responsibility. Consequences for research ethics The study's exposure contributed to major changes in US human- subjects protection. The National Research Act was enacted in 1974, and later reforms strengthened independent ethical review. The Belmont Report, published in 1979, set out respect for persons, beneficence, and justice as core principles. The CDC's own historical account links Tuskegee to subsequent changes in research practices and the emergence of formal protections. [26] Yet regulatory reform does not erase the need to examine institutional culture. Forms can be signed without genuine understanding; ethics committees can be treated as procedural hurdles; and participants may continue to be vulnerable because
Nuremberg Revisited 58 of poverty, discrimination, or dependence on a provider. Tuskegee should therefore be read not as a closed historical episode but as a test for present systems: when the interests of a study conflict with a participant's welfare, which interest actually prevails? The case also requires care when discussing present-day mistrust. It is understandable that people who learn of Tuskegee may question institutions that once deceived participants. But using this history to dismiss all modern research would make a different mistake. The appropriate response to documented wrongdoing is stronger transparency, reliable independent review, meaningful remedies and honest acknowledgement—not the rejection of evidence as such. Tuskegee is a warning about what happens when scientific authority is insulated from accountability. It is also a reason to make accountability visible.
Nuremberg Revisited 59 CHAPTER 10 Guatemala, 1946–1948: Deliberate Infection Without Consent The question was not whether sexually transmitted infections deserved study. It was whether vulnerable people could be exposed to them without knowing, choosing or consenting. A programme beyond the borders of the United States Between approximately July 1946 and December 1948, researchers associated with the United States Public Health Service and the Pan American Sanitary Bureau conducted sexually transmitted disease research in Guatemala; follow-up work continued through 1953. The programme included intentional exposure to syphilis, gonorrhoea and chancroid among prisoners, soldiers and psychiatric patients. The Presidential Commission’s subject database identified 1,308 people intentionally exposed; it documented some form of treatment for 678 of them. Separate diagnostic research involved 5,128 people, including children, leprosy patients and other institutionalised or dependent populations, and included blood draws and lumbar or cisternal punctures. These figures describe different categories of activity and should not be added together as though they were unique participants. The Commission cautioned that the archival record was incomplete and that some retrospective accounts contained inaccurate or incomplete data. [27, 124]
Nuremberg Revisited 60 The Presidential Commission for the Study of Bioethical Issues published its report, 'Ethically Impossible', in 2011. The Commission said it had reviewed more than 125,000 pages of original records and more than 550 published sources, and concluded that the Guatemala experiments involved gross violations of ethics even when assessed against the researchers’ own understanding of the practices of their time. The report also explicitly describes limits in the evidence: the documents were scattered and incomplete, and some accounts were retrospective. Thus, the ethical conclusion is strong, while some individual-level counts, sequences and responsibility questions must still be tied to the records and caveats stated in the report. [27, 124] Guatemala is especially important in a history of global research because the project took place outside the sponsoring country's own population. Research across national borders is not inherently unethical, and international cooperation can be essential for public health. But moving research to a different jurisdiction can create additional risks: participants may have less access to information, weaker mechanisms for complaint, limited independent representation, and less power to influence what is studied or how results are used. Guatemala shows what happens when the geographic distance between funder and participant is accompanied by an ethical distance. Scientific purpose did not excuse the methods The studies were connected to the search for ways to prevent and treat sexually transmitted infections, including evaluation of prophylactic measures. Their stated scientific purpose does not resolve the moral problem. The researchers deliberately exposed people to pathogens or sought infection in settings where valid informed consent was not obtained. Some procedures relied on vulnerable individuals whose circumstances made refusal difficult or impossible. [27] The Commission's conclusion is especially consequential because it rejected the suggestion that the work could be excused by the
Nuremberg Revisited 61 standards or practices of the period. It found gross violations even judged against the researchers' own contemporary understanding of ethical requirements. This is an important distinction from arguments that rely on applying today's full regulatory framework retrospectively. Ethical criticism can take historical context seriously while recognising that deliberate infection without consent was incompatible with basic principles of medical responsibility. [27] The Guatemala case also exposes how a study's label can obscure its true ethical character. Research may be described as prevention research, inoculation research, public-health investigation or the development of a prophylactic method. The label is not decisive. The decisive questions are what was done to the participant, what was disclosed, what was understood, whether the participant could freely refuse, and what protections were in place. In Guatemala, the decisive failures concerned the basic relationship between investigator and subject. Vulnerability and layered responsibility The participants were not a single homogeneous group. The programme involved multiple settings and populations, and each requires careful reconstruction from the records. Prisoners live under custodial authority; people in psychiatric institutions may depend on the institution for care and daily life; and people with limited social or economic power may find it difficult to challenge researchers connected with government or medicine. Vulnerability does not mean that every person in a category lacks decision-making capacity. It means that conditions can make coercion, misunderstanding or exploitation more likely and demand stronger safeguards. [27] The Commission attributed moral culpability to individuals who approved, conducted, facilitated and funded the experiments to varying degrees. This offers a useful model for the book's broader analysis: responsibility is distributed, but it is not thereby eliminated. The investigator who performs a procedure, the
Nuremberg Revisited 62 supervisor who approves it, the institution that provides infrastructure and the sponsor that funds it may bear different kinds and degrees of responsibility. A rigorous account should identify the evidence for each role rather than attributing identical knowledge or intent to everyone involved. [27] A structural analysis is necessary as well. The research could proceed because institutional systems permitted access to vulnerable people, because meaningful consent and independent review were lacking, and because the participants did not have effective ways to stop the study. Focusing only on one or two named researchers risks leaving intact the arrangements that made the experiments possible. At the same time, structural explanation must not become an excuse that dilutes the accountability of identifiable decision-makers. Why the discovery mattered The Guatemala experiments became known publicly in the early twenty-first century after historical research uncovered records that had not been widely recognised or examined. President Barack Obama apologised to the people of Guatemala in 2010, and the Presidential Commission was charged with conducting a detailed fact-finding inquiry. The resulting report established a public account of the programme and its ethical significance. [27] The long interval between the experiments and their public examination raises questions about archives, institutional memory and the rights of people who cannot obtain a record of what happened to them. Research ethics requires more than a consent process at the beginning. Accountability also depends on the preservation of protocols, correspondence, case records and adverse-event information, so that later generations can determine whether people were protected and whether institutions fulfilled their obligations. The case also illustrates that official investigation can be a corrective act, but it is not equivalent to complete justice. A report can acknowledge what happened and condemn the conduct; it
Nuremberg Revisited 63 may not restore health, repair family consequences or provide a direct remedy to every person affected. The ethical aftermath of a study must include attention to recognition, apology, compensation where appropriate and safeguards against recurrence. Lessons for research across countries Guatemala should inform present-day debates about global clinical research without turning them into a simplistic claim that all research in lower-income countries is exploitative. Such countries also need and conduct important research into locally significant illnesses. Excluding populations from research can leave them with less evidence and fewer accessible interventions. The ethical requirement is fair and relevant inclusion—not exclusion, and not exploitation. [27, 28] At minimum, cross-border studies require a credible scientific rationale for the selected population, independent review in appropriate jurisdictions, consent in a language and form participants understand, protections against coercion, realistic arrangements for treatment and care when harm occurs, and a fair account of who may benefit from the findings. Local ethics review should be meaningful rather than ceremonial, and the institutions and communities involved should have a voice in priorities and governance. The Guatemala case belongs in this history because it reveals the gap between universal claims about science and the unequal distribution of power in research. The ethical standard must follow the person, not the passport. A participant's dignity cannot depend on the country in which the protocol is implemented or on the income level of the population from which subjects are recruited.
Nuremberg Revisited 64 CHAPTER 11 Soldiers as Research Participants: Duty, Discipline and Consent A signed consent form cannot answer every question about freedom when the person signing it is subject to military authority. The military as a research environment Military organisations have legitimate reasons to study protection against chemical agents, infectious diseases, environmental extremes, fatigue, and other threats to service members. Military medicine has to prepare people for risks encountered in operational settings, and research can make protective equipment, prophylaxis, and treatment more effective. The ethical problem begins when the chain of command, secrecy, or institutional priorities weaken the service member's ability to understand the study, refuse participation, or seek redress. [29, 30] The soldier is not automatically incapable of voluntary consent. Service members may participate in legitimate trials with appropriate information and safeguards. But rank, discipline, dependence on superiors, career concerns and the pressure to demonstrate loyalty can affect voluntariness. A researcher must not assume that a soldier's obedience to lawful orders extends to consent for experimentation. Military hierarchy makes independent safeguards particularly important because a
Nuremberg Revisited 65 participant may worry that refusal will be viewed as disloyal, career-limiting or inconsistent with unit identity. The US Government Accountability Office's 1994 review of Cold War-era programmes reported that precise totals could not be established, but hundreds of thousands of people had been used as subjects in federal national-security experimentation between 1940 and 1974. It found that in some cases basic safeguards were absent or not followed, that some subjects suffered immediate injury or later adverse effects, and that lack of records made claims difficult. The GAO also noted that no mechanism then ensured that the formal policies on informed consent and volunteer protection were consistently implemented. [29] The Nuremberg Code and American military policy The post-war history contains a striking contradiction. The Nuremberg Code articulated the need for voluntary consent and protection against undue harm in 1947. The GAO's 1994 review records that a 1953 memorandum from the US Secretary of Defense directed the military services, in substance, to use the Code as a guide for human experimentation. Yet later government reviews still found cases in which consent or protective rules were not followed. [29] The contradiction matters. A principle can be formally adopted while operational practices diverge from it. In a classified programme, the participant may be told only part of the purpose; risk information may be presented in technical language; records may be held in separate agencies; and medical follow-up may be limited. The existence of a written policy provides an important baseline, but it does not demonstrate that each participant understood the study or that every researcher complied. The GAO also reviewed the Army's use of volunteers and whether participants were fully informed about the nature of experiments. Its 1984 report reflects congressional concern that military volunteers should understand what they were being asked to do. These reviews are valuable because they show that the issue was
Nuremberg Revisited 66 not simply a distant historical allegation: the government itself recognised weaknesses in the documentation, implementation and oversight of programmes conducted in the name of national security. [29, 31] Edgewood Arsenal and chemical-agent testing One of the best-documented examples is the Edgewood Arsenal programme in Maryland. The US Department of Veterans Affairs describes Army chemical research conducted between 1955 and 1975 to evaluate the effects of low-dose chemical warfare agents on military personnel and to test protective clothing and pharmaceuticals. According to the VA's summary of Department of Defense information, about 7,000 service members participated in experiments involving more than 250 different chemicals. Some volunteers experienced symptoms during exposure, and long-term follow-up was not planned as part of the original studies. [30] These facts need to be read carefully. The existence of an exposure study does not establish that every subject was deceived or that every exposure caused long-term harm. Some participants volunteered and were informed to varying degrees, and the VA describes subsequent assessments that did not find significant health effects in the studied cohort at the group level. At the same time, incomplete long-term follow-up and historical uncertainty can make it difficult for an individual veteran to determine what exposure occurred or whether a later condition is related. The ethical concern includes both the quality of consent and the adequacy of continued medical responsibility. [30] The Edgewood experience shows why research records are part of participant protection. If the study sponsor does not preserve detailed exposure records, a participant may have difficulty obtaining an accurate medical history decades later. A research programme can end administratively while its ethical obligations continue. The ability to identify the agent, dose range, protocol,
Nuremberg Revisited 67 adverse event and follow-up plan is not a clerical extra. It is a condition of later care, review and accountability. Volunteers, coercion and the meaning of choice The term 'volunteer' can conceal important differences. One service member may have received a clear explanation, time to consider the risks and reassurance that refusal would not affect career standing. Another may have received a brief description that omitted classified details or may have felt unable to decline an offer made in a hierarchical setting. The ethical status of a programme should therefore not be decided by the presence of a volunteer form alone. The process of recruitment, the information shared and the practical consequences of refusing all matter. [29, 31] Military research also raises a special question about public service. Service members accept certain occupational hazards in order to perform duties. That does not mean they have consented to every research exposure that a military institution might consider useful. The occupational expectation to face a known operational danger differs from being enrolled in an experiment designed to generate generalisable knowledge. Research that directly protects participants may have a strong ethical rationale, but it still requires transparent explanation and suitable oversight. A second problem is the distinction between classified objectives and participant information. National-security needs may justify withholding particular operational details from the public. They do not automatically justify concealing material health risks from the person exposed. If full disclosure is truly impossible, the ethical case for enrolling a participant becomes more difficult, not easier. Independent review must assess whether the experiment can proceed without violating the participant's ability to make an informed decision.
Nuremberg Revisited 68 Harm, remedy and the limits of retrospective proof The GAO reported that some participants in Cold War programmes experienced immediate injury, death or later health effects, but it also noted that the effects of many experiments were difficult to determine and that records were often insufficient. These qualifications must appear together in a credible account. It would be inaccurate to state that every subject suffered permanent injury; it would be equally inaccurate to infer that the absence of a traceable record proves no harm occurred. [29] Retrospective assessment is intrinsically difficult. People may present decades later with symptoms that have multiple possible causes. Exposure records may be missing, agencies may have changed names or responsibilities, and programmes may have been classified. A fair system should not promise causal certainty where science cannot provide it, but it should assist former participants in obtaining records, reconstructing likely exposures, receiving appropriate clinical assessment and understanding whatever remedy is available. The VA's current information on Edgewood acknowledges the long interval since testing and explains that no contemporary test can confirm all exposures that took place decades ago. It recommends history-taking and clinical evaluation where veterans are concerned. This modern information is not a substitute for accountability; it is a practical acknowledgement that exposure reconstruction and care can remain relevant long after research has concluded. [30] The wider lesson: service does not remove rights Military medicine is often described through duty, sacrifice and unit cohesion. Those values can support ethical research when they encourage responsible protection of personnel. They become dangerous when loyalty is used to silence questions or when a participant is treated as a resource belonging to the organisation. The ethical duties owed to service members do not disappear
Nuremberg Revisited 69 because the research is classified, the sponsor is the state or the stated purpose is national defence. A robust system should provide an independent chain for consent, written and understandable disclosure of material risks, documentation of each protocol and exposure, access to an advocate where appropriate, monitoring of adverse events, a clear right to decline without career penalty, and long-term access to relevant records. Research involving service members should also receive scrutiny from reviewers who understand the realities of command and service culture, but who are not subordinated to the operational interests promoting the study. The American experience offers a case study of both abuse and attempted correction. Federal reviews, service-member testimony, disclosure programmes and revised rules show that institutions can investigate their own past, though often only after public or congressional pressure. The central lesson is not that all military research is unethical; it is that military necessity must operate within ethical limits, and those limits are meaningful only when participants can rely on them in practice.
Nuremberg Revisited 70 CHAPTER 12 MKUltra: Intelligence, Drugs and Unwitting Subjects The destruction of records was not a minor administrative failure. It narrowed the possibility of knowing what had been done to people and of holding institutions to account. A programme born in secrecy Project MKUltra was a Central Intelligence Agency programme of research into behavioural modification and substances that might be useful in clandestine operations. It began in the early 1950s and expanded into numerous subprojects. The CIA's programme included research involving drugs such as LSD, hypnosis, electroshock, sensory deprivation and other approaches to influencing behaviour. The 1977 Senate hearing documented the programme and the discovery of records that had survived an earlier destruction order. [32] The programme emerged from Cold War fears that adversaries might be developing drugs or techniques for interrogation, coercion or behaviour control. Security concerns can explain why officials sought such research; they do not excuse using people without their knowledge. Some subprojects involved legitimate scientific research with no clear evidence that every participant was unwitting. Others involved subjects who were not told that they were receiving a substance or were part of an experiment. This variation must be retained rather than presenting every MKUltra activity as identical. [32]
Nuremberg Revisited 71 The Senate hearing described over 150 subprojects identified in surviving financial records. Research was funded through universities, hospitals, prisons, mental institutions and other settings, sometimes without institutions or researchers being told that the funding originated with the CIA. Some subjects did not know they were subjects at all. The combination of secret sponsorship and undisclosed participation undermined accountability at two levels: the individuals could not make informed choices, and the institutions around them could not reliably assess the true purpose of the work. [32] Unwitting drug tests Some of the most troubling activities involved the administration of psychoactive substances to people who did not know they were being tested. The Senate record describes arrangements involving agents of the Federal Bureau of Narcotics and testing in social settings, including covert operations in safehouses in New York and San Francisco. The programme sought to observe effects on people whose behaviour could be monitored by agents, but those agents were not necessarily trained scientific observers. The surviving record notes that subjects could become ill for hours or days and that at least one instance required hospitalisation. [32] The ethical failure was fundamental. A person cannot weigh the benefits and risks of participation if they do not know a substance is being administered or that their reactions are being observed for a research purpose. Even where a compound was given in a small dose, the subject had no opportunity to disclose relevant medical history, assess personal risks, request an alternative or stop the procedure. Deception was not a side effect of the design: it was part of how the study could be conducted covertly. This also raises a scientific problem. Covert administration and uncontrolled settings make the observations difficult to interpret. The Senate record described limitations in the quality of observation and follow-up. The programme's secrecy therefore harmed both ethics and knowledge: it stripped people of choice
Nuremberg Revisited 72 while often producing data of questionable scientific value. The pursuit of a clandestine advantage did not guarantee reliable evidence. [32] The Human Cost and the Death of Frank Olson The history of MKUltra is associated with the death of US Army scientist Frank Olson after he was secretly given LSD in 1953 during a gathering involving intelligence and military personnel. Olson died after falling from a hotel window in New York. The circumstances, his state of mind, the actions of colleagues, and the adequacy of later investigations have been the subject of public controversy. The record surrounding his death should be handled with care: it is established that he received LSD without informed consent and died soon afterward, but not every claim about the circumstances has been conclusively proved. [32, 33] The Olson episode matters because it shows that an experimental exposure can have consequences extending beyond the immediate observation period. Psychoactive drugs may affect perception, judgement, anxiety, and behaviour; a subject's response is unpredictable, and a person kept unaware may be especially vulnerable because they cannot seek help on the basis of an accurate account of what occurred. The ethical obligation to anticipate and minimise harm applies even when the researcher hopes to observe a short-lived effect. The programme also demonstrates how secrecy can frustrate families. When a death occurs and relevant details are hidden, relatives may be unable to understand the sequence of events or pursue informed questions. An institution's obligation does not end with the death or disappearance of the immediate subject; it includes preservation of records, honest disclosure, cooperation with investigation and appropriate recognition of harm. The Destruction of Records In 1973, CIA Director Richard Helms ordered the destruction of MKUltra records, and Sidney Gottlieb, who had directed the
Nuremberg Revisited 73 programme, carried out the order. Many operational records were destroyed. In 1977, a search prompted by a Freedom of Information Act request uncovered seven boxes of financial records at a records centre. These surviving documents identified around 150 subprojects and documented some institutional links, but they did not provide a complete account of what had been done, to whom, or with what results. The Senate hearing records this extraordinary limitation. [32] The destruction meant that later investigators could not reconstruct many protocols, identify all subjects or determine the full range of exposures and consequences. This does not mean every missing detail can be assumed to represent a crime, nor does it permit exact totals to be invented. It means the institution's own choices made the historical record incomplete and reduced the possibility of individual remedy. Where a programme involves human participants, documentation is an ethical safeguard, not merely an administrative expectation. Record destruction also created a problem of institutional knowledge. The Senate was forced to reconstruct the programme from incomplete financial documents and other surviving materials. Some institutions later learned that they had received CIA-related funding without being aware of the ultimate sponsor. The inability to identify participants and protocols was therefore not accidental historical bad luck alone; it was shaped by decisions to keep the programme compartmentalised and later destroy files. [32] Oversight and the limits of classified research The 1977 Senate hearing asked why MKUltra had occurred, what safeguards existed and whether abuse could recur. It documented activities that were contrary to the values and legal protections expected in a democratic society. The history shows that intelligence research can be especially vulnerable to ethical failure when sponsorship is concealed, normal review is circumvented,
Nuremberg Revisited 74 and the population under study is treated as a target rather than a participant. [32] Not all intelligence research is necessarily human-subjects research, and not all classified research is automatically unethical. But secrecy cannot be allowed to make the ethical review process disappear. Sensitive research requires secure but independent review, clearly defined authority, records that survive beyond an individual programme, a lawful basis for every exposure, participant protections and post-hoc accountability where disclosure was initially restricted for legitimate reasons. A secret programme should not be able to define its own moral boundaries and then erase the evidence by which those boundaries could be assessed. MKUltra is also a reminder that the language of national security can distort the balance between ends and means. A government may argue that adversaries use coercive techniques and that its own research is necessary to understand or counter them. But a programme that secretly exposes its own or another country's civilians to substances does not become ethical merely because it claims defensive intent. The rights of the subject are not suspended by a threat assessment. What the case teaches The first lesson is that informed consent requires the disclosure of material facts, including the fact that a person is participating in research and is receiving an investigational substance. The second is that risk review must reflect the real procedure, not a sanitised description. The third is that institutions need records detailed enough to support monitoring, adverse-event review and later accountability. The fourth is that subjects need access to care and remedies even when a programme ends or changes its name. The final lesson is about the relationship between science and democratic oversight. A research programme that cannot tolerate independent scrutiny is likely to have a problem beyond classification. Ethical protection must function precisely when a
Nuremberg Revisited 75 sponsor possesses unusual authority and participants possess little information. MKUltra shows what happens when a secret research agenda is allowed to override consent, evidence preservation, and accountability.
Nuremberg Revisited 76 CHAPTER 13 Human Radiation Experiments: Secrecy, Risk and Delayed Accountability A low dose, a controlled exposure or an experiment described as routine still requires an ethical justification—and an honest account of what the participant was told. The atomic age and human subjects The development of nuclear weapons and radiation science during the Second World War and the Cold War created new questions for medicine and research ethics. Federal agencies supported studies involving radioactive tracers, radiation exposure and medical procedures associated with the investigation of radiation's effects. The Advisory Committee on Human Radiation Experiments, appointed by President Bill Clinton in 1994, reviewed government-sponsored or government-conducted experiments from 1944 to 1974 and reported its findings in 1995. The US Department of Energy's archival roadmap and the Committee's final report opened a substantial body of records for public scrutiny. [34, 35] The committee's investigation identified nearly 4,000 human radiation experiments sponsored by the federal government during the period; many involved low doses of radioactive tracers. This figure should not be mistaken for a count of 4,000 unethical experiments. The studies varied considerably in purpose, risk, disclosure and consent. The ethical task is to determine which individuals were exposed, what they were told, whether a valid
Nuremberg Revisited 77 research purpose and appropriate safeguards existed, and what later records show. [34] The radiation history is often reduced in public discussion to a handful of the most alarming examples. These examples deserve attention, but a serious historical account must also consider the broader research system: how studies were authorised, who reviewed them, how agency and institutional responsibilities were divided, how information was communicated to participants and the public, and what happened when later scrutiny revealed gaps. The committee's work was valuable because it did not rely solely on anecdote; it systematically examined a large documentary record and made findings and recommendations. [34] Disclosure and consent The fundamental ethical concern in some radiation experiments was that people were exposed without adequate disclosure or without understanding the research purpose and the risks. In other cases, participants were informed to varying degrees or the intervention was part of a recognised clinical context. The existence of radioactive material in a study does not automatically prove that the study was abusive; nor does the fact that a dose was low automatically make non-disclosure acceptable. Risk and consent are separate questions. A small risk may still require consent, and a beneficial objective does not eliminate the duty to tell a participant what is being done to them. [34, 35] The committee's review also highlighted the problems created by secrecy. Records were spread across agencies and archival collections, definitions of 'human radiation experiment' varied, and it was difficult to identify some of the people who had taken part. The GAO reported in 1994 that the extent of Cold War radiation experiments might never be completely known because of problems locating and analysing documents and determining who had participated. [35] The long interval between exposure and later review creates a difficult ethical situation. A participant may have no memory or
Nuremberg Revisited 78 documentation of a procedure, while an institution may no longer possess the detailed records needed to clarify it. A later diagnosis may or may not be related to the exposure. Ethical accountability requires acknowledging both the scientific uncertainty of individual causation and the institutional duty to preserve records, explain known facts and assist participants in obtaining care or remedies where justified. Examples and differences in context Public discussions of radiation research have included studies involving radioactive tracers, total-body irradiation, radioisotopes and other procedures. The appropriate approach is to evaluate each study rather than to treat every experiment as having the same ethical status. Some research concerned clinical questions in the context of serious disease; other studies raised concerns about whether individuals understood the purpose or the nature of exposure. The committee's final report examined multiple categories and distinguished findings about consent, risk, scientific merit and the circumstances in which research took place. [34] One important issue is the difference between treatment and research. A patient may receive radiation as part of an attempt to diagnose or treat an illness. Research may also use radiation to answer a generalisable question. The boundaries can overlap, but the ethical requirements do not vanish when the patient is seriously ill or the investigator believes the procedure may offer benefit. People should understand whether an intervention is experimental, what alternatives exist, what is known and unknown about risks, and whether the purpose is primarily clinical care or the production of knowledge. Another issue is the role of children, pregnant women, hospital patients and other people whose options may be constrained by dependency or circumstance. The mere inclusion of a vulnerable population is not proof of misconduct; some studies must involve particular groups to answer a relevant health question. But such
Nuremberg Revisited 79 inclusion increases the obligation to justify the design, minimise risk, assess consent capacity or proxy authority where applicable, and ensure an independent reviewer can evaluate whether the research is ethically defensible. The committee and public accountability The Advisory Committee on Human Radiation Experiments conducted a public investigation and issued a final report in 1995. Its work included a review of documents, oral histories and a broad examination of the historical context. The Department of Energy separately published a roadmap and a list describing more than 400 experiments associated with DOE and its predecessors, while other agencies conducted additional searches and reviews. The multiplicity of reports reflects the breadth of the federal research enterprise and the difficulty of reconstructing it from records held by different organisations. [34, 35] The GAO's 1994 report expressed concern that agencies used inconsistent definitions and had difficulty locating participants. It also reported that the advisory committee was working under a short time frame while the scale of the records made complete assessment challenging. This illustrates the difference between making information public and fully reconstructing the human experience. A list of experiments is necessary, but it does not by itself establish what an individual participant knew, what was done to them or what harm followed. [35] Public apology, disclosure and archival access are meaningful parts of accountability. They can restore some truth to a record previously shaped by secrecy. But accountability also requires an accessible path for former participants and families to ask questions, receive medical assessment and pursue remedies. A historical programme cannot be ethically evaluated only from the standpoint of the institution that designed it. It must also be considered from the perspective of the people whose bodies were used to generate its evidence.
Nuremberg Revisited 80 What the radiation history teaches First, research records should preserve the basis for consent, the identity of the agent or material, the expected risk, the protocol, deviations, adverse events and follow-up arrangements. Second, independent review must be able to see the real procedure rather than a summary designed to protect the sponsor. Third, research involving national security or sensitive technology must remain subject to a lawful system of ethical oversight. Fourth, researchers and sponsors should retain the capacity to identify and inform participants if later knowledge changes the assessment of risk. [34, 35] The radiation history also reminds us that evidence changes over time. Ethical evaluation should consider what was reasonably known at the time, but it must not use uncertainty as a blanket excuse. If a study involved a material risk unknown to participants, the ethical status depends on what investigators knew or should have known, what safeguards were feasible and whether information was withheld. Conversely, a later scientific concern should not automatically be used to accuse every earlier investigator of wrongdoing without examining the evidence. The ultimate lesson is that technology can change faster than institutions' ability to govern it. A scientific field associated with enormous public benefit may also create pressure to accelerate research, maintain secrecy or treat disclosure as an obstacle. Ethics is the mechanism by which scientific power is constrained by the claims of the human beings who make research possible. When records are missing and consent is uncertain, the obligation to investigate is not a distraction from science. It is part of public accountability for science.
Nuremberg Revisited 81 CHAPTER 14 Prisoners, Children and Institutionalised Patients A person can be technically asked for permission and still have very little freedom to refuse. Ethical review must look at the reality of choice, not merely the appearance of consent. Why some populations are especially vulnerable Research involving prisoners, children and people living in institutions raises particular ethical questions because the investigator may hold authority over people whose freedom is already restricted. Prisoners can face limited privacy, inadequate health care, restricted income and dependence on custodial staff. Children may lack the legal capacity to consent and may not understand the risks of a procedure. People in long-term institutions may depend on administrators for shelter, care and daily life. These circumstances do not mean that such populations must never be included in research. They mean the research needs a clear justification and additional protection. [36, 37] The history of medical research includes examples in which vulnerable populations were recruited because they were easy to access, available in large numbers or less able to challenge investigators. Such recruitment can lead to unfair distribution of risk. It can also create a misleading claim of voluntariness: a person may sign a form because participation appears to be the only route to better living conditions, medical attention, money or institutional admission. A valid process must examine whether
Nuremberg Revisited 82 those advantages distort the participant's judgement or become a form of pressure. This chapter considers four different cases: the Willowbrook hepatitis studies; the Jewish Chronic Disease Hospital cancer-cell experiment; research in Holmesburg Prison; and the malaria studies at Stateville Penitentiary. Their methods and ethical profiles differed, and they should not be presented as identical. Together, however, they illuminate the recurring problem of consent under dependency and the reasons why later regulations imposed additional safeguards on research involving prisoners and other vulnerable groups. Willowbrook: hepatitis research in a state institution Studies of the natural history and prevention of viral hepatitis began at Willowbrook State School in New York in 1955. The institution housed children with intellectual disabilities, and hepatitis was common in the crowded conditions. Saul Krugman and colleagues conducted research that included intentional exposure of children to hepatitis. The studies remained controversial for decades, including among physicians and bioethicists. Krugman later published a retrospective account of the work, while other ethicists criticised the consent procedures and use of children in challenge experiments. [38, 39] Those who defended the work pointed to the high background incidence of hepatitis at Willowbrook, the effort to understand the disease and the potential development of preventive approaches, including the use of immune serum. They also argued that some parents were informed and consented. These claims should be presented because they formed part of the historical debate. But parental consent does not automatically settle the ethics of enrolling a child in research that intentionally exposes the child to infection. Consent must be informed and voluntary, and proxy permission must be evaluated against the child's interests, level of risk and availability of alternatives. [38, 39]
Nuremberg Revisited 83 A major concern was the environment in which families made decisions. Admission to Willowbrook was difficult, and some accounts describe pressure or the perception that participation could help secure admission. Where access to essential care is linked, explicitly or implicitly, to research participation, consent may become compromised. The issue is not simply whether a parent said yes, but whether refusal was a realistic option without a penalty to the child or family. Research can be ethically problematic even where parents believe that they are acting in their child's best interests. The question of deliberate infection in children Willowbrook raises a wider ethical issue: under what circumstances, if any, can deliberate infection be used in research? Controlled human challenge studies continue to be debated in modern medicine. Their ethical acceptability depends on factors such as the seriousness of the disease, the availability of prevention and treatment, the vulnerability and competence of participants, the quality of consent, risk minimisation, scientific necessity, and independent review. The contemporary debate generally places special restrictions on challenging children or people unable to consent, especially where there is no prospect of direct clinical benefit. [39] The Willowbrook experiments cannot be assessed only by observing that infection was common in the institution or that the research contributed to scientific knowledge. High background risk does not remove the obligation to protect a child from an additional experimental risk. The relevant question is whether the study was scientifically necessary and ethically justified, whether risk was proportionate, and whether children were recruited because their circumstances made them convenient. The child's own welfare must remain central even when a parent or institution agrees to research. The studies contributed to a public debate that helped sharpen attention to informed consent, vulnerable participants and the
Nuremberg Revisited 84 relationship between research and institutional power. A balanced chapter should acknowledge the scientific arguments made in favour of the work while clearly examining why many ethics scholars regard the procedures and recruitment circumstances as ethically unacceptable. Historical context can explain the controversy, but it cannot replace ethical evaluation. Jewish Chronic Disease Hospital: cancer cells without adequate disclosure In 1963, physicians at Brooklyn's Jewish Chronic Disease Hospital injected live cancer cells into 22 chronically ill patients as part of research intended to study the body's response to foreign cells. The controversy centred on the fact that patients were not adequately informed that the injections contained live cancer cells and that the procedure was experimental rather than part of their ordinary treatment. Barron H. Lerner's historical analysis in the New England Journal of Medicine describes the case as a major controversy about research without informed consent. [40] Investigators argued that the cells were unlikely to produce cancer in the recipients and that the procedure was intended to study immune rejection. Even if a particular scientific assessment of risk was held by the investigators, that did not resolve the disclosure problem. Patients were entitled to know the nature of what was being injected and why. When a procedure is presented in a way that prevents a patient from understanding its essential character, the consent process fails at the point where it matters most. This case demonstrates that a procedure need not cause a documented injury to be ethically unacceptable. The central violation was the use of patients in an experiment without adequate disclosure and valid consent. Ethical research protects a person's right to make decisions, not merely their chance of avoiding a bad outcome. A participant who is deceived but experiences no lasting harm has still been denied the right to decide whether to participate.
Nuremberg Revisited 85 Holmesburg Prison: the market for research subjects Holmesburg Prison in Philadelphia became a prominent site for medical, cosmetic, and chemical testing during the mid-twentieth century. Research included tests of topical products, drugs and chemical agents, and sponsors included commercial companies and government-related interests. The history was brought to wider public attention through investigations and Alan Hornblum's book, “Acres of Skin”. A National Academies review of prisoner research summarises accounts that between 1962 and 1966, 33 pharmaceutical companies tested 153 experimental drugs at Holmesburg. Those figures should be attributed to the historical account rather than treated as an official census of all studies conducted at the facility. [36] Some inmates signed up willingly and were paid. But payment in prison occurs in a setting where income and choices are constrained, and an inmate may have strong reasons to accept opportunities not readily available outside. The ethical question is not whether money was offered; it is whether the amount and surrounding circumstances compromised the participant's ability to weigh the risks freely. It is also necessary to examine whether participants received full information about immediate and long- term effects, whether they could stop participation, and whether they had access to independent medical care for adverse events. [36, 37] The use of prisoners as a readily available population created a commercial as well as an ethical concern. When researchers and sponsors can recruit a large, controlled group at comparatively low cost, a prison can begin to resemble a research market. Such a setting creates the risk that the convenience of recruitment becomes a hidden driver of study design. The problem is not that prisoners are incapable of making any decision about research. It is that liberty is restricted and ordinary alternatives are reduced, so the conditions for truly voluntary choice require special scrutiny.
Nuremberg Revisited 86 Stateville: malaria research, wartime urgency and consent During the Second World War, researchers conducted malaria studies at Stateville Penitentiary in Illinois, where prisoners were deliberately infected with malaria to test the safety and efficacy of antimalarial drugs. The work responded to an urgent military problem: malaria threatened personnel in theatres of war, and effective treatments were needed. Some prisoners volunteered and received inducements; historical accounts describe the programme as having been viewed by some contemporaries as a model for ethical research. Later scholarship has re-examined the study through the Nuremberg Code and current principles, focusing on voluntariness, risks and exploitation. [41, 42] Stateville is a useful counterpoint to cases in which consent was entirely absent. The ethical issue is more nuanced. Researchers may have explained the study, obtained signed permission, and selected the participants because they could be followed closely. Yet prison confinement limits autonomy, and the value of a payment or other advantage may be magnified when a person's ordinary choices are restricted. The correct assessment must examine recruitment, the voluntariness of consent, the risks imposed, the availability of alternatives, the quality of care and whether the study's benefits justified those risks. The Stateville research also shows why ethical judgement develops over time. An experiment once praised for its organisation or contribution to drug development may later be viewed more critically when the social conditions of recruitment and the meaning of voluntariness are examined. This does not require pretending that researchers had no legitimate scientific objective. It requires recognising that scientific merit and ethical legitimacy are separate questions. What the cases have in common—and what differs The four cases should not be collapsed into one category. Willowbrook involved children with intellectual disabilities and
Nuremberg Revisited 87 deliberate infection; Jewish Chronic Disease Hospital involved inadequate disclosure to chronically ill patients; Holmesburg involved a large-scale prison research market with varied procedures and sponsors; Stateville involved prisoner volunteers in malaria research that has received differing ethical assessments over time. The methods, risk levels, intentions, and consent circumstances were different. [36, 38, 39, 40, 41, 42] The common questions, however, are striking. l Could participants refuse without losing something important? l Did they understand the real nature of the procedure? l Were risks minimised and proportionate? l Were the people selected because the scientific question required their inclusion, or because their circumstances made them easy to recruit? l Did independent reviewers consider the local power relationships? l Was there a plan for clinical care, adverse events and follow- up? These are the questions that turn a catalogue of incidents into a framework for ethical assessment. Modern US regulations impose additional protections for research involving prisoners. The Department of Health and Human Services requires special review, restricts the categories of research that can be approved under its prisoner-subject provisions, and requires ethics committees to examine whether any advantages offered to prisoners are so large that they impair judgement in the limited-choice environment. The regulations also require fair selection and comprehensible information. These protections emerged from an understanding that formal consent alone may not be sufficient where freedom is constrained. [37]
Nuremberg Revisited 88 The enduring lesson: protect choice in reality Ethical research does not promise to remove all risk. It requires risks to be scientifically justified, disclosed and minimised; participants to be selected fairly; and consent to be valid. When research involves children, prisoners or institutionalised patients, the process must be attentive to dependency and unequal power. The question is not simply whether the paperwork appears complete. It is whether the person had a meaningful opportunity to understand, deliberate, refuse and withdraw. The histories of Willowbrook, the Jewish Chronic Disease Hospital, Holmesburg and Stateville also show why public accountability matters. Ethical norms change, but each historical case leaves records from which future investigators can learn. A balanced account should avoid both extremes: describing every study involving a vulnerable population as abusive, or excusing unethical conduct because it was once accepted by some professionals. The evidence and circumstances of each case matter. The essential test is whether the person remains a subject with rights or becomes a resource controlled by the researcher, sponsor, or institution. When dependency is strong, the ethical burden on investigators and review bodies increases. It is their responsibility to demonstrate that participation is justified and genuinely voluntary—not the responsibility of a child, prisoner, or institutionalised patient to overcome the power imbalance alone.
Nuremberg Revisited 89 PART IV PHARMACEUTICAL RESEARCH, LOWER- INCOME POPULATIONS AND GLOBAL INEQUALITY Chapters 15–20
Nuremberg Revisited 90 Chapter 15 When Research Crosses Borders: The Geography of Risk and Benefit Modern clinical research crosses borders because disease does not respect them, because therapies must be tested in populations who may use them, and because international collaboration can bring expertise and investment to places where research infrastructure is limited. Multinational studies can answer questions that a single-country trial cannot: whether a treatment works across different genetic backgrounds, diets, disease burdens, health systems and patterns of care. They can also build laboratories, train investigators and give participants access to close monitoring that would otherwise be unavailable. There is no inherent ethical defect in a trial conducted in a lower-income country, just as there is no automatic guarantee of ethical conduct in a wealthy one. The ethical problem arises when the reasons a population is selected also make its members easier to recruit, harder to protect, or less likely to share in the benefits. Poverty can make even modest reimbursement feel indispensable. A public hospital with scarce medicines may become an unusually attractive recruitment site. A patient may confuse a research visit with ordinary treatment, or believe that the investigator can offer care only if the patient agrees to participate. A parent may sign a form without having a realistic opportunity to understand it. A community may bear the inconvenience, uncertainty and risk of a
Nuremberg Revisited 91 study while the resulting product is priced beyond its reach or is not relevant to its most urgent health needs. These risks are structural rather than proof of bad faith by every individual researcher. They arise from the distribution of bargaining power. Sponsors often control the protocol, funding, data systems and product-development strategy. Local investigators may depend on the grant for staff and institutional prestige. Ethics committees may lack independent resources, specialist expertise or freedom from institutional pressure. Regulators may be overstretched. Participants, meanwhile, may have little access to the protocol, no independent adviser and limited ability to challenge an institution on which they depend for care. The question is not simply where the trial happens. A defensible ethical review asks why this population is being studied, whether the question addresses a health priority relevant to it, what care participants will receive, and what obligations survive the last study visit. The Council for International Organizations of Medical Sciences (CIOMS), in guidelines prepared with the World Health Organization, states that research in low-resource settings should respond to the health needs or priorities of the communities or populations in which it is conducted. That principle rejects a model in which disadvantaged people are recruited merely because they are available, while the scientific question and resulting product serve only someone else. [28, 44] “Responsiveness” should be assessed before recruitment, not added as a public-relations statement after a product has been developed. Does the research concern a disease that burdens the host population? Is the intervention likely to be affordable and deliverable in its health system? If it succeeds, who will be able to obtain it? If it fails, will participants still receive clinically appropriate care? Will local investigators share leadership, analytic access and authorship? Will the host system retain useful
Nuremberg Revisited 92 capacity, or will the operation leave with the sponsor and the data? These questions are not all resolved by individual consent forms. Consent under conditions of dependency Informed consent is a process of communication, not a signature collected at the end of recruitment. A valid process gives people information in a language and form they can understand; explains that the intervention is experimental where relevant; describes material risks, uncertainty, alternatives and compensation arrangements; and makes clear that refusal will not jeopardise ordinary care. The person must have enough time and practical freedom to decide. For children, authorisation by a parent or legally authorised guardian and age-appropriate assent are distinct obligations; the fact that a school, hostel or state agency has custody of a child does not automatically make that institution an appropriate substitute for independent parental permission. Researchers should not assume that low literacy means a person lacks decision-making capacity. Nor should they mistake comprehension checks for a test that participants must pass. The ethical task is to improve the explanation, allow questions and recognise the limits of what a person can reasonably decide in a pressured situation. Community engagement can add context, but community leaders cannot consent on behalf of every competent adult. Individual voluntariness and appropriate community consultation are complementary, not interchangeable. After the final visit The distribution of benefits and burdens is especially visible after a trial ends. Participants may still need an intervention that appeared beneficial; a host country may lack the resources to purchase the eventual medicine; and communities may be left without an accessible summary of the study’s results. The 2013 revision of the World Medical Association’s Declaration of Helsinki strengthened the prominence of post-trial provisions, while the current Declaration continues to require that protocols
Nuremberg Revisited 93 address post-trial arrangements and that participants who still need a beneficial intervention have appropriate provisions made for them. It also requires public registration and dissemination of results, including negative or inconclusive findings. [44] Post-trial access is not a promise that every participant must receive a drug indefinitely regardless of clinical circumstances. It requires advance planning, a clear explanation to participants, and a credible arrangement among sponsors, researchers, health systems and governments. The plan may involve continued treatment, referral, access through a public programme, or another ethically justified route. What is indefensible is to leave the question until the sponsor has already obtained the data it needs. A test for ethical partnership A useful standard is reciprocity without coercion. The host population should not be required to purchase access to its own contribution at a price that makes the research benefit theoretical. It should not be exposed to avoidable risk merely because its regulatory system is weaker. At the same time, local needs do not mean every trial must guarantee a successful product or that communities must receive a share of all commercial revenue. The obligations depend on the study, the risks, the intervention and the agreements made before recruitment. But meaningful local relevance, independent oversight, protection against research- related injury, responsible data governance and transparent results are baseline duties, not optional favours. International standards have become more elaborate. Good Clinical Practice guidance assigns responsibilities to sponsors, investigators, and other stakeholders, and the 2025 ICH E6(R3) principles emphasise quality by design, proportionate risk management, participant protection and reliable data. [45] These tools matter only if ethics review is independent, investigators can act on safety concerns, participants can complain without retaliation, and violations have consequences. A carefully drafted
Nuremberg Revisited 94 protocol cannot protect a person if every institution responsible for enforcing it prefers the study to continue. The history of cross-border research is therefore neither a story of inevitable exploitation nor a story in which regulation has solved the problem. It is a continuing contest over who sets the scientific question, who carries uncertainty, who controls the evidence, and who benefits when the experiment succeeds. The following chapters show what happens when those questions are answered under the pressure of an epidemic, a search for affordable treatment, a weak oversight system, or a market that rewards favourable results.
Nuremberg Revisited 95 Chapter 16. Kano, 1996: Trovan and the Ethics of an Epidemic In 1996, an outbreak of meningococcal meningitis struck Kano State in northern Nigeria. Children were among those affected, and the epidemic created the worst setting for deliberation: fear, urgent need, scarce resources and families desperate to save a sick child. Médecins Sans Frontières was among the organisations providing emergency care. Pfizer sent a team to conduct a study of trovafloxacin, marketed as Trovan, an investigational antibiotic, in children with meningitis. The trial enrolled approximately 200 children, comparing Trovan with ceftriaxone, an established antibiotic. A BMJ report in 2011 stated that five children in the Trovan arm and six in the ceftriaxone arm died, and that Pfizer had not reported the outcomes of ten children at that time. These counts should be attributed to that report rather than treated as a complete, independently re-analysed clinical dataset. The deaths alone cannot establish that either treatment caused them; meningitis itself can be fatal. The controversy concerned the design, consent, oversight and care decisions surrounding the study. [46, 47, 120, 121] The study later became one of the best-known examples in the debate about international clinical research. Nigerian authorities investigated whether the trial had the necessary approvals and complied with applicable rules. A BMJ report published in 2001 quoted the supervising Nigerian physician as saying that his office had backdated an approval letter; the same report stated that
Nuremberg Revisited 96 Pfizer acknowledged that local ethics approval might not have been properly documented. These were reported statements and allegations surrounding an official inquiry, not a final judicial determination of every disputed fact. In United States appellate litigation brought by Nigerian children and their guardians, the Second Circuit described the allegations in the complaints under the procedural standard applicable to a motion to dismiss and reversed and remanded the lower-court judgments. The appellate decision did not establish every underlying allegation after a trial. In July 2009, Pfizer announced a settlement with Kano State under which the state would dismiss its civil and criminal Trovan- related cases; Pfizer expressly denied wrongdoing or liability. A careful history must preserve the distinctions among reported allegations, an appellate procedural ruling, an announced settlement and proof on the merits. [46, 47, 48, 120, 122, 123] The moral force of the setting An epidemic does not suspend research ethics; it makes safeguards more urgent. Families may be unable to distinguish a research offer from emergency treatment. A clinician in a crowded ward may be both the person offering potentially lifesaving care and the person recruiting the child into a study. Parents may fear that refusal will reduce access to treatment. Even a verbally stated explanation can fail if the family is not told clearly that the medicine is experimental, that alternatives exist, or that participation is voluntary. Consent involving children requires particular care. The adult’s permission is not a formality, and it cannot be inferred from presence in a hospital, cooperation with staff or acceptance of treatment. The parent or guardian must understand the distinction between clinical care and research, including the comparative aims and uncertainties. In an emergency, the prospect of benefit may be real, but so is the possibility that hope will overpower a family’s ability to assess risk. Independent consent staff, translation, documentation and a clear route to
Nuremberg Revisited 97 standard treatment are not bureaucratic luxuries; they help preserve the possibility of a real choice. The question of ethical review became central to the dispute. An external sponsor’s approval process cannot substitute for competent review in the country where research occurs. Local review is not a nationalist obstacle to science; it provides contextual knowledge and legal accountability. At the same time, a local ethics committee must itself be independent, competent and adequately resourced. A stamp on an application cannot legitimise an inadequate protocol, and a committee that does not exist or cannot review a trial leaves participants exposed. What settlement can—and cannot—resolve The dispute ultimately led to prolonged litigation and financial settlements involving Nigerian authorities, families and Pfizer. In its July 2009 announcement, Pfizer described the agreement with Kano State and expressly denied wrongdoing or liability. A BMJ report in 2011 separately reported payments of $175,000 to each of four families whose children had died during the study. These sources describe different settlements and should not be conflated. Neither a company announcement nor a settlement report is a judicial finding that every contested allegation was proven. Settlements can end specific claims or provide compensation without conclusively resolving the full historical record. The significance of Kano lies not only in the litigation outcome but in what it exposed: the risk that the urgency of an epidemic could be used to justify a study whose oversight and consent arrangements were contested; the difficulty of obtaining reliable documentation years after events; and the damage to public trust when people believe their children were treated as means to a commercial end. [121, 122] The case also illustrates a wider social consequence. When communities see outside institutions arrive during a health crisis, collect data, and depart while questions remain unanswered, suspicion can spread beyond the individual study. Research
Nuremberg Revisited 98 misconduct or perceived misconduct may weaken confidence in later public-health interventions, even when those later interventions are unrelated. The history of vaccination resistance in northern Nigeria has been discussed in relation to broader political, religious and social distrust, with the Trovan controversy among several factors in the wider environment. It would be too simple to attribute a later vaccination boycott to one drug trial alone, but it would be equally mistaken to ignore how perceived research injustice can corrode trust. [48] Lessons for emergency research First, the protocol must explain why the trial is necessary during an emergency and why its design offers a credible prospect of benefit without withholding effective treatment unnecessarily. Second, recruitment and consent should be insulated, as far as possible, from the immediate authority of the treating team. Third, an independent safety process must be able to change or stop the study quickly. Fourth, participants and families need transparent follow-up and a path to compensation for research- related harm. Fifth, documentation must be preserved so that the public can assess what happened. Finally, there must be a realistic plan for communication and accountability after the emergency has passed. None of these safeguards means that clinical trials should never take place during epidemics. In emergencies, research can be indispensable: it can identify effective therapies and prevent repeated use of unproven treatments. But the urgency that makes research valuable also creates incentives to lower scrutiny and recruit people whose choices are constrained. The correct response to a crisis is a faster, better-organised ethical review— not the abandonment of review. Kano remains a caution because the ethical quality of research is judged not by the sponsor’s good intentions alone, nor by whether a drug may ultimately prove useful, but by the fairness of the process through which knowledge is produced. If parents are not
Nuremberg Revisited 99 properly informed, if authorisation is uncertain, if safety decisions cannot be independently scrutinised, or if subsequent accountability becomes a matter of competing statements, then even research undertaken amid a genuine health emergency may leave a legacy of distrust.
Nuremberg Revisited 100 Chapter 17. The Placebo Question: HIV Prevention and Unequal Standards of Care In the 1990s, a controversy over HIV prevention trials forced medical researchers to confront one of the hardest questions in research ethics: what is the proper control group when a proven intervention exists in one part of the world but cannot readily be delivered in another? The immediate issue was prevention of mother-to-child transmission of HIV. In 1994, the AIDS Clinical Trials Group Study 076 showed that a relatively intensive zidovudine regimen substantially reduced transmission in the United States and France. The regimen involved antenatal treatment, intravenous administration during labour, medication for the infant and clinical monitoring. It was effective, but costly and difficult to deliver in health systems where women often entered care late, deliveries took place outside facilities able to administer intravenous treatment, and health services had limited resources. [49, 50] Researchers and public-health agencies sought shorter, simpler and less expensive regimens that could realistically be implemented where HIV transmission was concentrated. Some studies used placebo controls. Critics argued that these designs were unethical because an effective intervention had already been shown to work; participants assigned to placebo would not receive that protection. In a widely discussed 1997 article, Peter Lurie and Sidney Wolfe argued that these trials exposed pregnant women and infants to avoidable risk in order to make research faster or
Nuremberg Revisited 101 cheaper. The controversy was about more than the tablet count: it asked whether economic inequality could be used to redefine what counted as an acceptable standard of care. [49] The argument in favour of local feasibility Defenders of the trials argued that the regimen proved in high- income settings could not simply be transplanted into every health system. Its complexity, expense and dependence on repeated antenatal attendance limited feasibility. They argued that a simpler regimen needed to be tested in the actual circumstances where it would be used, and that a placebo design could produce a clear answer efficiently. Harold Varmus and David Satcher, writing in the New England Journal of Medicine in 1997, acknowledged the ethical challenge but stressed that research aimed at interventions countries could implement was different from testing a product solely for the benefit of wealthier nations. [50] The defenders’ strongest point was practical: an intervention that exists on paper but cannot be delivered reliably may not be the best public-health strategy for a particular setting. The critics’ strongest point was moral: the reason people could not access an intervention should not automatically become the reason researchers are permitted to withhold it from them. A community’s poverty cannot, by itself, turn a proven protective treatment into an ethically irrelevant option. Both concerns are real, and they cannot be resolved by saying either that local standards always govern or that the most expensive intervention available anywhere must be used in every research arm. How to assess a placebo design The correct analysis begins with the scientific question and the evidence available at the time, not with the convenience of the sponsor. What was the proven intervention? Was it effective for the same indication and population? Was it feasible in the host setting, and if not, why? Could the question have been answered through an active comparator, an add-on design, or a shorter
Nuremberg Revisited 102 regimen that preserved some established protection? Would a placebo arm expose participants to an increased risk of serious or irreversible harm? Were participants told honestly what care was available elsewhere and what they might be denied by randomisation? Did the host community help identify the intervention it actually needed? The Declaration of Helsinki has been revised repeatedly in response to such debates. Its current guidance prefers comparison with the best proven intervention, permitting placebo or no intervention where no proven intervention exists, or where compelling and scientifically sound methodological reasons justify the design and participants will not face additional risk of serious or irreversible harm from not receiving the best proven intervention. It cautions against abuse of the exception. [44] The principle is not an absolute ban on placebo; it is a demanding justification test. The concept of “standard of care” must therefore be handled carefully. It can refer to current local clinical practice, a legally accepted standard, the best evidence-supported treatment, or the care that a health system can reliably deliver. Those are not always the same. A research protocol that treats the lowest level of care in a resource-poor setting as the only relevant comparator risks embedding injustice in its design. Conversely, insisting on a regimen that a health system cannot sustain may yield results that do not answer the population’s most urgent implementation question. Scientific validity and ethical acceptability need to be examined together. Beyond the placebo debate The controversy changed the way researchers and institutions discuss global health research. It helped establish that a trial can be ethically troubling even when participants sign consent forms and the protocol is statistically sound. It also showed why post- trial access, affordability, community relevance and independent review must be negotiated before recruitment. A short regimen
Nuremberg Revisited 103 might be useful, but who would manufacture it? Would the country be able to afford it? Would participants be told if the sponsor had no plan to supply it? The value of the knowledge depends partly on whether it can improve health outside the trial report. No single historical controversy settles every present-day placebo decision. The ethical answer depends on the intervention, the clinical risks, the feasibility of alternatives, the population’s health priorities, and the evidence available at the time. What the HIV trials teach is that “this is usual practice here” is not, by itself, a sufficient ethical argument. Researchers must justify why the comparison is necessary, minimise avoidable risk, engage local communities, and plan honestly for the consequences of success. The deeper lesson concerns inequality in the production of evidence. When wealthy societies can demand the best available care while poorer societies are asked to accept placebo because the best care is unaffordable, the research system risks converting an injustice in access into an advantage in experimentation. Ethical research should work in the opposite direction: its design should help identify interventions that can be used in real settings without treating deprivation as a licence to lower protection.
Nuremberg Revisited 104 Chapter 18. India’s Clinical-Trial Reckoning: Consent, Compensation and Accountability India became an important setting for clinical research because it has a large and diverse population, a substantial burden of communicable and non-communicable disease, experienced clinicians, hospitals capable of conducting complex studies, and a large pharmaceutical sector. International trial activity also raised questions about the strength of regulatory oversight, the independence of ethics committees, the adequacy of consent, and the remedies available when a participant was injured. These questions were not merely theoretical. During the early 2010s, public controversy, parliamentary scrutiny and Supreme Court proceedings brought the protection of research participants into sharper focus. Clinical trials and compensation In 2012, the Supreme Court of India was asked to consider public- interest litigation concerning clinical trials and safeguards for participants. Subsequent orders addressed regulatory supervision, informed consent, monitoring and compensation. In its October 2013 proceedings in the Swasthya Adhikar Manch litigation, the Court recorded undertakings concerning audio-visual recording of the informed-consent process for specified trials and preservation of records while maintaining confidentiality. Such directions reflected concern that consent needed to be demonstrable and
Nuremberg Revisited 105 auditable rather than merely asserted by a sponsor or investigator after a dispute arose. [51] Audio-visual recording is not a substitute for understanding. A participant may speak on camera and still be misinformed, intimidated or unable to comprehend the information. Recording also raises legitimate concerns about privacy, data security and whether participants may feel pressured to agree because they are being filmed. Its ethical value depends on an adequate explanation, secure storage, restricted access and a process that allows people to ask questions privately. The underlying principle is that the consent process must be real, documented and open to independent scrutiny. Compensation for trial-related injury is another central safeguard. It distinguishes research from an arrangement in which participants carry all the downside while sponsors obtain the data. A credible system needs rules for recording adverse events, assessing causality, providing medical management and deciding compensation without requiring a participant to fight a powerful institution alone. Causality can be difficult: illness may arise from the underlying disease, a study procedure, an investigational product or an unrelated event. Difficulty is not a reason to abandon participants; it is a reason to specify independent assessment, timelines, documentation and appeal routes before the first person is enrolled. India’s New Drugs and Clinical Trials Rules, 2019, consolidated provisions for new drugs and clinical trials, including requirements concerning permissions, ethics committees, informed consent, safety reporting and compensation for injury or death related to a trial. [52] The existence of rules, however, does not establish that every trial site follows them. Effective protection depends on enforcement, adequate staffing, independent audits, transparent reporting and meaningful penalties for serious non- compliance.
Nuremberg Revisited 106 The HPV demonstration project A particularly consequential Indian controversy concerned a human papillomavirus (HPV) vaccine demonstration project implemented from 2009 in collaboration with the Program for Appropriate Technology in Health (PATH), the Indian Council of Medical Research (ICMR), and the governments of Andhra Pradesh and Gujarat. The vaccines used were Gardasil and Cervarix. The Government of India inquiry report records that 14,091 girls in Andhra Pradesh and 10,686 in Gujarat received the first dose (24,777 first-dose recipients in total); the corresponding third-dose counts were 13,791 and 9,637. These are dose-specific counts, not interchangeable measures of completion. The project was funded by the Bill & Melinda Gates Foundation, and the vaccine doses were donated by their manufacturers. The Ministry of Health and Family Welfare suspended the projects in 2010 after reports of deaths among participants and rising concern about the conduct of the programme. The 72nd Parliamentary Standing Committee Report is the principal primary document for the committee’s later findings and should be read alongside the inquiry materials and the position taken by PATH. [53, 54, 119] The central ethical controversy cannot be resolved by the fact that deaths were reported after vaccination alone. The 72nd Report records an expert assessment of seven reported deaths—five among recipients of Gardasil in Andhra Pradesh and two among recipients of Cervarix in Gujarat—which found no common pattern suggesting vaccine causation. The same section criticises the timeliness and uniformity of adverse-event reporting, the investigation process and the adequacy of safety monitoring. This is not equivalent to proving that every death was unrelated, nor does the report establish that the vaccines caused the deaths. Temporal association is not proof of causation; an absence of a common pattern is also not a complete substitute for individual clinical adjudication. The ethical questions about consent, oversight, reporting and participant protection remain distinct from the causal question. [119]
Nuremberg Revisited 107 The Government of India appointed an inquiry committee, and the Parliamentary Standing Committee on Health and Family Welfare examined the matter in its 72nd Report, presented in August 2013. The report’s consent section records that, among 9,543 Andhra Pradesh forms, 2,763 were signed by hostel wardens; it also cites a review of 100 Andhra Pradesh forms in which 69 lacked a witness signature. It lists further irregularities, including undated forms, signatures that did not match the named guardian, vaccination dates preceding consent signatures, and forms completed by school or hostel staff without written authorisation from parents. The report records that officials from the Department of Health Research and ICMR maintained that approvals were in order, while acknowledging implementation discrepancies. The committee’s critical findings should be attributed as findings of a parliamentary inquiry, not as a court judgment. The original report’s paragraph on Gujarat consent forms contains an apparent numerical inconsistency in the reproduced figures (545 versus 5,454); this manuscript therefore does not repeat the disputed Gujarat figure without a further primary-record check. [53, 119] PATH, for its part, stated that the project had been implemented with approvals from ethics committees and relevant Indian authorities, and argued that the vaccines had a favourable safety record and that the deaths were not causally linked to vaccination. [54] The parliamentary committee sharply criticised the conduct and regulatory handling of the project and recorded concerns about consent, monitoring and conflicts of interest. A balanced historical account should preserve the disagreement and the status of each source: PATH’s statement is the sponsor’s position; the 72nd Report is a parliamentary inquiry document; and neither alone substitutes for a transparent case-by-case causality assessment. The criticisms of consent and oversight do not, on their own, establish that HPV vaccines as a class are unsafe; the claim that a vaccine has a favourable safety profile does not excuse failures in consent, monitoring or accountability. [119]
Nuremberg Revisited 108 Why classification matters Calling a programme a “demonstration project” rather than a “clinical trial” is not a neutral choice if it alters the legal or ethical obligations applied to it. When an intervention is already licensed somewhere, studies may still be designed to generate evidence about implementation, safety, effectiveness or acceptability. The label alone cannot determine the protections participants need. The 72nd Report argued that the PATH project should have been treated as a clinical trial under the applicable framework. The committee’s conclusion should be read with the actual protocol, approvals and legal definitions, but its central governance lesson is clear: sponsors and public institutions should not be able to lower safeguards through nomenclature. The use of children from vulnerable or remote communities made the quality of consent particularly important. Parents and guardians need understandable information about the purpose of vaccination, the research question, possible harms, the status of the product and what follow-up will occur. A guardian’s signature is not meaningful if the person has not received the necessary information. A school, hostel or local official can help communicate; it should not simply replace the parent’s informed decision unless a lawful and ethically justified process specifically authorises that arrangement. Researchers must also take care that the offer is not confused with a compulsory school programme or the only way to receive ordinary care. Reform must survive public scrutiny The Indian experience points to a broad lesson for research governance. Regulations, court orders and parliamentary reports are not evidence of failure in themselves; they are mechanisms through which failures can be exposed and corrected. Yet a system loses credibility if serious recommendations remain without clear responsibility, if compensation decisions are opaque, or if the institutions under examination help to shape the inquiry without
Nuremberg Revisited 109 transparent declarations of interest. Independent review needs independence in practice, not only in its letterhead. For every clinical trial, the public should be able to establish who approved the protocol, how participants were recruited, how consent was obtained, what monitoring was available, how adverse events were adjudicated, whether compensation was paid, and how results were disclosed. This does not mean releasing identifiable medical records or breaching participant privacy. It means creating accountable systems whose decisions can be examined without exposing individuals to further harm. A further legal distinction is necessary. In Kalpana Mehta v. Union of India (2018), the Supreme Court considered how a Parliamentary Standing Committee report may be relied upon in judicial proceedings. The report remains an important public and institutional record, but its statements should not automatically be described as a judicial finding that a particular party committed a legal wrong. In March 2026, the Government of India told the Lok Sabha that actions following the 72nd Report had been taken into cognisance by the Committee in its 81st Report and listed subsequent regulatory reforms, including the New Drugs and Clinical Trials Rules, 2019, ethics-committee registration and prospective trial registration. This later official answer describes government-reported reforms; it does not independently resolve every historical allegation or establish that all earlier concerns were remedied. [124, 125] India’s trial history should not be reduced to a blanket claim that research conducted in the country is unsafe or unethical. Such a claim would be inaccurate and would obscure the contributions of Indian investigators and institutions to useful medical knowledge. The defensible conclusion is more precise: large-scale research creates obligations proportionate to its risks, and those obligations are not met by approval documents alone. Consent, safety, compensation and independent accountability must work together, particularly when participants are children, poor,
Nuremberg Revisited 110 dependent on institutions, or unfamiliar with the distinction between care and research.
Nuremberg Revisited 111 Chapter 19. When the Sponsor Owns the Evidence: Conflicts, Publication and Transparency Pharmaceutical companies have contributed to major advances in treatment, diagnostics, and prevention. Clinical development is expensive, technically demanding, and often impossible without partnerships among industry, universities, hospitals, regulators, and public research bodies. The ethical question is not whether industry should participate in research. It is how to ensure that a sponsor’s legitimate commercial interests do not distort the design, conduct, interpretation, or publication of evidence on which patients and clinicians rely. A conflict of interest exists when a secondary interest—such as financial gain, career advancement, institutional funding, or loyalty to a sponsor—creates a risk of improperly influencing a primary professional judgment. A conflict is not proof that a person has acted dishonestly. It is a condition requiring disclosure, management, and, where necessary, independent control. The US Institute of Medicine’s 2009 report on conflicts of interest in medical research described the value of collaboration with industry while warning that extensive financial ties can undermine the integrity of research, education, patient care and guideline development if they are not appropriately governed. [55]
Nuremberg Revisited 112 The evidence can be distorted without falsifying a number Misconduct is often imagined as invented data or a fabricated experiment. Those are serious possibilities, but the evidence base can also be distorted by choices that look ordinary in isolation: selecting a comparator that favours the sponsor’s product, choosing an outcome that is easier to move than a clinically important endpoint, changing analytic plans after results are known, emphasising subgroup findings, failing to report adverse outcomes clearly, delaying publication of negative studies, or presenting a secondary outcome more prominently than the pre- specified primary one. Any of these choices can mislead when readers cannot compare the published paper with the original protocol and full results. A prominent example comes from a 2008 analysis by Erick Turner and colleagues of clinical trials submitted to the US Food and Drug Administration for 12 antidepressant drugs approved between 1987 and 2004. Among 74 studies they identified, 31 per cent were not published. Publication and reporting were strongly associated with whether the FDA had judged a study’s results positive. The published literature made 94 per cent of trials appear positive, whereas the FDA data classified about half as positive. The authors estimated that selective publication and reporting inflated the apparent effect size. The study did not show that every investigator or company deliberately falsified data; it demonstrated that the visible literature can present a systematically rosier picture than the complete regulatory record. [56] For patients, this is not a minor academic problem. Clinical guidelines, prescribing decisions, regulatory assessments, procurement and public-health policy depend on a complete picture of benefits and harms. If a failed study remains invisible, another group may repeat work unnecessarily, a clinician may overestimate the benefit of a drug, and a participant’s contribution may be wasted because the knowledge was never
Nuremberg Revisited 113 shared. Negative results are not failed science; if the study was well designed, they are part of the answer. Registration is necessary but not sufficient The International Committee of Medical Journal Editors announced in 2004–2005 that journals would generally consider trials for publication only if they had been registered before enrolment began. Prospective registration was intended to reduce selective publication and selective reporting, prevent needless duplication, and allow participants and ethics committees to identify studies that were planned or ongoing. [57] Registration creates a public record of the question researchers said they intended to answer, making later changes easier to identify. Registration, however, does not guarantee that outcomes will be reported accurately, that a protocol will be followed, or that the public can see the data needed to interpret results. WHO’s 2017 joint statement on clinical-trial transparency called for prospective registration and timely public disclosure of results from all clinical trials, describing the incomplete reporting of studies as an ethical and scientific problem that can lead to poor decisions and wasted resources. It set a global norm of making summary results available within 12 months of primary study completion, alongside journal publication as appropriate. [58] Transparency also requires trial protocols and statistical analysis plans to be available, changes to be explained, and adverse events to be presented in a way that allows meaningful evaluation. Individual participant data can support independent verification and further research, but sharing must be designed around privacy, consent, data security, and the risk of re-identification. Public access cannot be an excuse to disclose sensitive records; confidentiality cannot be an excuse to keep all scrutiny within the sponsor’s walls.
Nuremberg Revisited 114 The independence of the researcher A sponsor-funded investigator may have excellent expertise and act with complete integrity. The risk arises when the contract gives the sponsor undue control over access to data, analysis, interpretation, manuscript preparation or the decision to publish. Academic institutions should examine whether contracts allow investigators to access the full dataset, verify the analysis and publish results regardless of their direction. Authorship should reflect actual contribution and accountability. A sponsor’s medical writing support may be legitimate if transparently acknowledged and the named authors retain control; undisclosed ghostwriting and honorary authorship obscure responsibility. Ethics committees should examine not only the risk-benefit calculation but also the practical conditions under which independent interpretation will occur. A trial that can be stopped by the sponsor but whose results cannot be published by the investigators creates a concerning asymmetry. Similarly, a researcher who receives substantial consultancy fees, equity or institutional funding from a company may reasonably have a conflict that requires independent review. The right response depends on its nature and degree: disclosure may suffice in some circumstances, while recusal, independent analysis or prohibition of particular roles may be necessary in others. Accountability without cynicism The purpose of transparency is not to presume that corporate research is false. Industry has the capacity to run large studies, manufacture interventions at scale and fund work that universities could not otherwise support. The point is to make evidence trustworthy even when financial incentives are strong. A system that discloses conflicts, registers protocols, reports all outcomes and permits independent scrutiny protects participants, responsible sponsors and the public together. A useful rule is that no organisation should have an unreviewable monopoly over the evidence it produces about its own product.
Nuremberg Revisited 115 Regulators need complete submissions; ethics committees need safety updates; investigators need access sufficient to fulfil scientific and publication responsibilities; and the public needs timely results that include harms, null findings and uncertainties. Commercial confidentiality may justify protecting manufacturing secrets or genuinely sensitive information, but it should not be used to conceal clinically important evidence from the people whose health decisions depend on it. Participants accept inconvenience and risk so that society can learn. Their contribution is not honoured when only the favourable parts of the answer are made visible. The ethical contract of research therefore extends beyond consent and bedside conduct: it includes the honest stewardship of data, fair interpretation, and public reporting of what was learned—even when the result is commercially disappointing.
Nuremberg Revisited 116 Chapter 20. Who Owns a Human Sample? HeLa, Havasupai and the Limits of Consent A blood sample can appear small and anonymous: a tube with a label, a specimen stored in a freezer, a fragment of tissue left after a biopsy. Yet the material may contain a person’s genetic information, family relationships, ancestry and clues to disease. Once collected, it can be copied, sequenced, transferred, commercialised and combined with other datasets. The original clinical encounter may be long over before a new use is contemplated. This creates ethical questions that are not exhausted by asking whether someone signed a form on the day the sample was taken. Henrietta Lacks and the cells that would not die In 1951, biopsy specimens taken from Henrietta Lacks, a 31-year- old patient being treated for cervical cancer at Johns Hopkins Hospital in Baltimore, gave rise to the HeLa cell line. The cells grew reliably in culture and became a foundational resource in biomedical research. They contributed to work across many areas of biology and medicine, and their use became international. The National Institutes of Health records that the specimens were obtained without Lacks’s knowledge or consent. [59] HeLa is a difficult historical case because it sits at the intersection of medical care, research practice, racial inequality, privacy and the changing legal status of tissue. In 1951, the legal and
Nuremberg Revisited 117 institutional framework surrounding leftover clinical specimens was not what it is today. That context matters when reconstructing individual responsibility; it does not erase the moral significance of a patient being unable to make an informed decision about the use of her biological material. Nor should the story be told as if the cells were valuable only because they became scientifically useful. Henrietta Lacks was a person receiving care, not a cell line waiting to be discovered. The consequences extended to her family. Genetic information from HeLa cells can reveal information relevant to living relatives who never supplied a sample. In 2013, NIH announced an agreement with members of the Lacks family for controlled access to HeLa whole-genome sequence data, balancing research use with privacy and the family’s preferences. NIH reaffirmed the agreement in 2023 and continues to set conditions for access to such data in NIH-controlled repositories. [59] The agreement did not erase the historical wrong or create a universal ownership rule for all human tissues. It represented a governance response in a distinctive case. The Havasupai: consent for one purpose, use for another The Havasupai Tribe case made the issue of secondary use particularly visible. Members of the Arizona tribe provided blood samples to researchers who were studying type 2 diabetes, a serious community health concern. The tribe later learned that samples had been used for research on other topics, including schizophrenia, population migration, and inbreeding, which had not been included in the purpose understood by participants. Some of these uses conflicted with Havasupai beliefs about ancestry and origin. In 2010, Arizona State University agreed to a settlement of $700,000 with 41 tribal members; the settlement also addressed the return of samples. The case is discussed in the New England Journal of Medicine as a warning about consent for stored biological materials and later research uses. [60, 61]
Nuremberg Revisited 118 The Havasupai case differs from HeLa in important ways. Havasupai members had intentionally contributed samples for research, but the controversy concerned purposes beyond what they understood and approved. Lacks’s specimens were obtained without her knowledge or consent during clinical care in a period with different rules. One case cannot simply be used as a legal precedent for the other. Together, however, they demonstrate that a specimen’s physical transfer to a laboratory does not dissolve the ethical relationship between the research enterprise and the people or communities from whom it came. Broad consent, specific consent, and community governance Modern research cannot always predict every future use of a specimen. A narrow form that names only one hypothesis may block valuable work that could not reasonably have been anticipated. But a vague statement that a sample may be used for “any future research” may leave participants unable to foresee uses that carry cultural, social or familial implications. There is no single consent formula that resolves all cases. The right approach depends on identifiability, sensitivity, the nature of the new use, the context in which material was obtained, the degree of risk and whether the proposed work aligns with the participant’s reasonable expectations. Broad consent can be ethically defensible when it describes the range of future research, explains governance and access procedures, identifies meaningful limits, and gives participants a clear account of whether withdrawal remains possible and what it can mean after data or cells have been distributed. It should not be used to conceal commercial development, ancestry research or other materially different uses. Some changes may require renewed consent or independent ethics review even when material has been de-identified. De-identification reduces certain risks but does not make all genetic or community-level harms impossible.
Nuremberg Revisited 119 For Indigenous peoples and other communities with strong collective interests in genetic information, individual consent may not be the whole answer. A study can affect a community’s reputation, cultural narrative or relationship with other groups, even if no individual can be identified in a publication. Community engagement and governance may therefore be appropriate, alongside individual consent. But community approval cannot override the rights of competent individuals, and community leaders cannot consent on everyone’s behalf. Ethical governance must hold both personal autonomy and legitimate collective concerns in view. Commercial Value and Fair Benefit Human samples can create substantial commercial value, while donors may receive no direct financial payment. That fact alone does not establish exploitation: many discoveries arise from shared scientific infrastructure, and a simple entitlement to royalties for each tissue donation would be difficult to administer fairly. The important questions are whether the original collection was ethically justified, whether use stays within the scope of valid consent or a legitimate review process, whether privacy is protected, and whether the terms governing access and benefit are transparent. Benefit-sharing can take many forms: return of useful findings where appropriate, community health investment, access to interventions, capacity-building, acknowledgment, data governance and participation in setting research priorities. These are not interchangeable, and no one remedy fits every case. A commercial product should not be promised if its development is speculative. But a research organisation should not present samples as freely available raw material while ignoring commitments made to the people who supplied them. The ethical value of consent also changes over time. Participants may agree to a particular study, but not to every imaginable future use of their genome. Families may be affected by the information
Nuremberg Revisited 120 revealed. Community members may have no personal sample in the project and still experience harm from group-level claims. Research ethics therefore needs procedures for continuing governance: access committees, restrictions on use, transparency about collaborations, breach reporting, community consultation when justified, and secure handling of genomic data. The body is not simply a resource There are real tensions between scientific openness and personal control. Restrictions that are too broad can impede useful research; unrestricted access can ignore privacy, context and trust. The aim is not to treat each sample as untouchable property or to give one individual complete veto over all future knowledge. It is to recognise that science depends on relationships and that the human material collected in one setting carries obligations into future settings. HeLa and Havasupai remind researchers that technological possibility does not automatically confer ethical permission. The ability to sequence a genome, identify ancestry, develop a cell line or commercialise a biological resource tells us what can be done. It does not answer what should be done, who should decide, or how the risks and benefits should be distributed. Those questions require consent that is meaningful, governance that continues after collection, and humility about the difference between possessing a sample and having earned the trust of the person or community from whom it came. Synthesis: a practical accountability test Across these chapters, six questions provide a practical review of research that crosses institutional, national, or social boundaries:
Nuremberg Revisited 121 • Relevance: Does the research address a real health need of the population being asked to participate, and has that population had a meaningful role in shaping the question? • Choice: Can participants understand the study, distinguish research from ordinary care, refuse without losing care or benefits, and withdraw without retaliation? • Independent oversight: Are approvals valid in the relevant jurisdiction, are conflicts disclosed and managed, and can the ethics committee and safety monitors act independently? • Fair risk and remedy: Are avoidable risks minimised, clinical care available, research-related injury assessed promptly, and compensation arrangements clear before recruitment? • Benefit and stewardship: Is there a credible plan for affordability, post-trial provisions, community benefit where appropriate, appropriate handling of samples and data, and protection of privacy? • Public accountability: Is the study registered prospectively, are protocol changes and adverse events handled transparently, and will all results—including negative and inconclusive findings—be made available? No checklist guarantees ethical research. These questions are a starting point for independent judgement. They must be answered with evidence, not assurances alone, and revisited when the circumstances change. The central lesson is that an ethically acceptable study is not merely one that produces useful knowledge. It is one that produces knowledge through a process that respects those whose bodies, time, information, and trust made the knowledge possible.
Nuremberg Revisited 122 PART V THE EVOLUTION OF RESEARCH ETHICS AND REGULATION Chapters 21–25
Nuremberg Revisited 123 Chapter 21. From Nuremberg to Helsinki: How Ethical Principles Became a Professional Standard The Nuremberg Code did not end unethical research. It supplied an unmistakable statement of principle after the Second World War, but the existence of a code was not the same as an effective system for enforcing it. In the decades that followed, medical researchers, professional organisations, governments and the public had to confront a harder question: how could ethical requirements be made part of ordinary research practice rather than left as declarations of good intention? [44, 62, 63] From the courtroom to the clinic The Nuremberg Code emerged from the judgment in the Doctors’ Trial, part of the subsequent Nuremberg proceedings against physicians and medical administrators involved in crimes under the Nazi regime. Its ten principles included voluntary consent, a scientifically justified purpose, avoidance of unnecessary suffering, proportionate risk, qualified investigators, adequate facilities, and the participant’s right to withdraw. The code also placed responsibility on the investigator to end an experiment when continuation was likely to cause injury, disability or death. These principles confronted a central moral failure of the preceding era: the treatment of a person as a means to scientific, military or ideological ends. [62]
Nuremberg Revisited 124 Yet the Code was not adopted as a stand-alone universal treaty, and national research systems did not immediately create uniform mechanisms to make its standards routine. Its origin in a criminal trial also meant that some physicians treated it as a response to exceptional atrocities rather than as a framework relevant to ordinary clinical investigation. That interpretation was ethically shortsighted. Exploitation need not involve torture or a concentration camp. A study may be coercive when a patient believes refusal will jeopardise medical care, when an employee cannot safely refuse a superior, when a participant is misled about the purpose of a procedure, or when risk is concealed behind technical language. [63, 73] The subsequent history is therefore not a simple relay race in which each new document replaced the last. The Nuremberg Code, professional declarations, national regulations and institutional review systems have overlapping authority and purposes. The Code remains a foundational historical statement; the Declaration of Helsinki developed a professional framework for medical research; the Belmont Report articulated principles and their applications in the United States; CIOMS translated wider ethical principles into detailed guidance; and Good Clinical Practice addresses operational responsibilities in clinical trials. None can be understood in isolation from the political and institutional settings in which it was written. [43, 44, 62, 63, 64, 69] The Declaration of Helsinki, 1964 The World Medical Association adopted the Declaration of Helsinki in June 1964. Unlike the Nuremberg Code, which arose from criminal proceedings, Helsinki was framed as a statement of ethical principles for physicians engaged in medical research. That professional authorship was both a strength and a source of debate. It could speak directly to clinical practice, but it also required the medical profession to scrutinise its own conduct, including situations in which professional judgement and commercial or institutional interests might conflict. [44, 63]
Nuremberg Revisited 125 The Declaration evolved through successive revisions because research practice, technology and public expectations changed. The 1975 revision in Tokyo is historically important for strengthening the requirement for review by an independent committee before human research begins. Later revisions addressed the relationship between research and clinical care, placebo-controlled designs, publication and registration, compensation, post-trial responsibilities, vulnerable groups and the use of identifiable biological material or data. The text has never been a static monument: its changes record unresolved arguments about the boundary between scientifically necessary exposure to risk and unacceptable exploitation. [44, 63] A recurring controversy concerned the standard of care in trials carried out in settings where the best proven treatment was not readily available. Researchers argued that comparisons must sometimes reflect local realities if a study is to be feasible and produce relevant evidence. Critics warned that the absence of services produced by poverty should not be used to justify giving trial participants less protection than people in wealthier countries. Debates over placebo use and post-trial access became emblematic of the larger question of whether scientific efficiency could be allowed to override the claims of justice. The issue is not resolved simply by naming a study “pragmatic” or “context- sensitive”; the ethical justification must be made explicit and subjected to independent review. [44, 64] Why the revisions matter Helsinki has broadened its understanding of the relationship between research and social context. The 2024 Declaration explicitly highlights structural inequities, engagement with participants and communities before, during and after research, fair distribution of benefits, and the need to consider the harms of excluding underrepresented populations as well as the harms of including them. It states that the rights and interests of participants cannot be subordinated to the purpose of generating knowledge. It also insists that research be scientifically sound,
Nuremberg Revisited 126 that risks be continually assessed, and that suitable compensation and treatment be ensured for research-related harm. [44] The 2024 revision makes the ethics committee’s role particularly concrete. Research protocols must be reviewed before the study begins; the committee must have independence, authority, resources, competence and local understanding; and international collaborative research must receive ethics approval in both the sponsoring and host countries. The committee must be able to monitor research and recommend changes, withdraw approval or suspend a study. The researcher remains responsible for reporting findings at the end. A committee that can only approve a protocol but cannot challenge conduct after recruitment is incomplete as a protective mechanism. [44] Informed consent is also described as a process, not a signature. Information should be given in plain language and adapted to participants’ communication needs. People must be told about aims, methods, risks, anticipated benefits, funding, conflicts of interest, confidentiality, compensation and the right to refuse or withdraw without reprisal. Where a dependent relationship or risk of duress exists, consent should be sought by someone independent of that relationship. These provisions matter in ordinary hospitals, universities, military facilities, workplaces and communities—not just in research sites associated with historical scandal. [44] What Helsinki is—and what it is not The Declaration is an influential professional ethical statement, not a global statute that automatically creates identical legal duties in every country. Its practical authority comes through professional commitments, institutional policies, regulatory requirements, journal expectations and its use in interpreting ethical standards. A government may enact more specific protections, and sponsors may impose requirements that go beyond a declaration. Conversely, the existence of a national law
Nuremberg Revisited 127 does not guarantee that every provision will be observed. [44, 52, 66] The Declaration should be read as a whole. Extracting a single paragraph about placebo, consent or vulnerable groups while ignoring the surrounding duties can distort its meaning. The correct citation also depends on the historical question. A study conducted in 1976 should not be judged as if the exact 2024 wording had been in force at the time. Its conduct should be assessed against the standards and laws applicable then, while later standards may help explain how the international community’s ethical understanding developed. That method avoids both anachronism and the opposite error of treating historical practice as ethically acceptable merely because the language of today had not yet been written. [44, 63] A declaration needs institutions The central lesson is that a declaration can identify what should be protected but cannot, by itself, ensure that protection. Implementation requires independent review, trained investigators, reliable records, transparent reporting, resources for participant care, the ability to pause research, and remedies when harm occurs. The history from Nuremberg to Helsinki is therefore a history of an increasingly detailed ethical vocabulary— and of the persistent gap between stating principles and making institutions answerable to them. [44, 66, 67]
Nuremberg Revisited 128 Chapter 22. The Belmont Report: Respect, Beneficence and Justice The Belmont Report is one of the clearest attempts to explain why research involving human beings requires moral justification and how broad ethical principles should guide practical decisions. Prepared by the United States National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research and published on 18 April 1979, it followed the National Research Act of 1974. The Act created a national commission and strengthened the formal framework for review of research involving human subjects. The history is inseparable from public concern over Tuskegee and other episodes in which the interests of science or institutions had been placed ahead of the rights of participants. [43] The Report did not claim to provide an algorithm that could settle every ethical disagreement. It identified three foundational principles—respect for persons, beneficence and justice— and examined their practical application to informed consent, assessment of risks and benefits, and selection of research subjects. Its enduring importance lies in connecting moral language to decisions that a researcher, ethics committee, or institution must make. [43] Respect for persons: autonomy and protection Respect for persons has two dimensions. First, people with the capacity to make decisions should be treated as autonomous
Nuremberg Revisited 129 agents whose considered choices matter. They require information sufficient to make a meaningful decision and freedom from coercion or undue influence. Second, people whose autonomy is diminished or compromised may require additional protection. The principle does not mean that some categories of person have no agency; nor does it mean that protection should automatically exclude them from research. It calls for safeguards proportionate to the circumstances. [43] In practice, respect requires more than a form bearing a signature. The participant must understand that the activity is research, what participation involves, which procedures are experimental, what alternatives exist, which risks and benefits are reasonably foreseeable, how confidentiality will be handled, and how withdrawal works. A person should not be led to believe that joining a trial guarantees a better treatment or that refusal will bring punishment, loss of routine care or displeasure from a treating physician. The quality of consent must be judged by the process and the participant’s understanding, not by the number of pages in the form. [43, 44, 64] The Belmont framework also helps explain why an apparently voluntary choice may be ethically compromised. A patient dependent on a clinician, an employee recruited by a supervisor, a soldier recruited through the chain of command, or a person dependent on a research institution for food, shelter or essential care may face pressure that is not expressed as an explicit threat. Payment can reimburse costs and recognise time; it can also become unduly influential if its size or timing distorts a person’s evaluation of risk. These situations require careful assessment rather than a blanket assumption that all payment is coercive or that a signed form proves freedom. [43, 64] Beneficence: Minimise harm, make research worthwhile Beneficence in the Belmont Report carries obligations. Investigators and institutions should not merely avoid deliberate
Nuremberg Revisited 130 injury; they should make reasonable efforts to secure participants’ well-being by minimising foreseeable harm and maximising the value of the research. A risk-benefit assessment must consider the probability and magnitude of harms, the likelihood and importance of anticipated benefits, the relevance of the research question, and whether risks can be reduced through a different design or by using non-human methods, existing data or less invasive procedures. [43] The principle exposes a basic truth: a scientifically weak study can be unethical even when its procedures appear harmless. Participants may give time, undergo blood sampling, accept inconvenience or reveal sensitive information, yet the study may be incapable of answering its question. If the sample size is inadequate, the outcome is poorly chosen, the comparator is inappropriate or the analysis is not planned in advance, the burden imposed on participants may have no reasonable prospect of producing valuable knowledge. Sound science is therefore an ethical requirement, not a technical detail to be addressed after approval. [44, 64, 70] Beneficence is not a licence to expose people to any risk that might produce a useful result. The value of knowledge must be weighed against the burden placed on the person who bears that risk. The balance is especially demanding where participants cannot directly benefit from the intervention, where risks are uncertain, or where the anticipated benefit accrues mainly to future patients, sponsors or society. Independent review can test the investigator’s assumptions, but the committee must have adequate scientific competence and must not substitute a vague promise of future benefit for a convincing justification. [43, 44] Justice: who bears the burdens and receives the benefits? Justice asks how research burdens and benefits are distributed. It is not fair for convenient, dependent or socially marginalised people to carry a disproportionate share of risk while the resulting
Nuremberg Revisited 131 benefits are directed to groups better able to pay. The historical record shows how recruitment has sometimes followed institutional availability rather than a defensible scientific reason: prisoners, institutionalised patients, poor hospital patients, racial minorities and people under administrative control became accessible to researchers because their refusal was difficult or their claims carried little institutional weight. [43, 73, 74] Justice does not require that every group participate equally in every trial. Scientific objectives sometimes justify studying a defined population, and some groups must be included if an intervention is intended for them. The question is whether selection is scientifically appropriate and ethically justified. A trial of a medicine intended primarily for older adults, for example, cannot always be adequately informed by research in young, healthy volunteers alone. At the same time, no group should be selected merely because its members are cheap to recruit or less likely to complain. [43, 44, 70] Justice extends beyond recruitment. It includes fair access to the knowledge and interventions produced, reasonable arrangements for care and compensation when study-related injury occurs, and honest communication of results. Research that extracts data or biological samples from a community while ignoring its priorities may produce scientific value yet still raise concerns about fairness. Belmont’s language is concise, but its implications include the global geography of trials, the affordability of new products and the duty not to conceal inconvenient findings. [43, 58, 64] From principle to procedure The Report organises its three principles around three applications. Informed consent operationalises respect for persons; assessment of risks and benefits gives practical form to beneficence; and the selection of subjects is central to justice. The three applications support one another. Informed consent cannot make an unjust study fair; a favourable risk-benefit ratio cannot excuse deception; and fair recruitment cannot rescue research
Nuremberg Revisited 132 whose design is incapable of producing reliable knowledge. Ethical review should therefore examine the protocol as a connected whole rather than using any single safeguard as proof of acceptability. [43] The distinction between research and medical practice is also important. Treatment is generally directed towards the welfare of a particular patient, whereas research is designed to develop or contribute to generalisable knowledge. The two can overlap: a patient may receive clinical care while participating in a study. That overlap makes disclosure especially important, because patients may assume that every additional test is necessary for their own care. The Belmont Report notes that an innovative or experimental procedure is not automatically a research project, but radical innovations should be evaluated systematically rather than allowed to remain unexamined indefinitely. [43] Strengths and limitations Belmont’s three principles are influential because they are memorable and flexible. Their flexibility, however, also creates work for reviewers. “Respect”, “benefit” and “justice” do not answer every question without interpretation. The Report emerged from a specific United States legislative and social context. It did not fully anticipate contemporary problems involving genomic databases, artificial intelligence, digital surveillance, large-scale data sharing, or international networks in which control of a study’s data may be far from the community that supplied it. Later guidance, including CIOMS and the Declaration of Helsinki, addresses many of these issues more explicitly. [44, 64, 71] The Belmont Report should thus be treated as a foundational compass, not a complete operating manual. Its questions remain essential: • Is the participant respected as a decision-maker? • Is the balance of risks and benefits defensible? • Are the burdens and benefits fairly distributed?
Nuremberg Revisited 133 The answers must be tested through protocol review, institutional accountability, and observation of what actually happens after recruitment. The principles matter most when the easy answer is inconvenient for a sponsor, investigator, or institution. [43, 66, 67]
Nuremberg Revisited 134 Chapter 23. CIOMS and Research in Lower-Resource Settings: From Protection to Partnership International ethical principles do not implement themselves in every country in the same way. A consent process developed for a well-resourced hospital may be poorly suited to a rural community with multiple languages, different expectations of authority, limited access to independent medical advice or a health system unable to provide the intervention after a trial. The Council for International Organizations of Medical Sciences (CIOMS), working with the World Health Organization, developed guidance to translate broad principles into practice in varied social, economic and regulatory circumstances. Its guidelines are especially important to the question running through this book: how can research cross borders without exporting risks to populations with less power? [64] The CIOMS–WHO relationship CIOMS began work on international research ethics guidance in the late 1970s. Its guidelines have been revised over time, in part to respond to developments in research practice and criticism of arrangements that could make participants in lower-resource settings bear risks without fair access to the benefits. The fourth edition, published in 2016, is titled *International Ethical Guidelines for Health-related Research Involving Humans*. It aims to facilitate the application of ethical principles in different
Nuremberg Revisited 135 socioeconomic, legal and administrative settings rather than to establish a lesser standard for people living in those settings. [64] This distinction is fundamental. Differences in context may require different methods of explanation, recruitment, follow-up, compensation, data protection and community engagement. They do not make deception, avoidable harm or unfair selection acceptable. Nor does a country’s limited health infrastructure justify treating a weakly resourced research site as a place where sponsors may lower their duties. Context helps determine what safeguards will work; it should not become a rhetorical device for excusing inadequate protection. [44, 64, 65] Social value and scientific validity Research should address a question important enough to justify exposing people to its burdens. In lower-resource settings, the question of social value has at least two dimensions. First, is the study scientifically valid and relevant to the health needs of the population from which participants will be recruited? Second, are there credible arrangements for translating the knowledge into benefits, rather than simply exporting samples, data and labour to a sponsor elsewhere? A study may generate valuable knowledge for the world, yet its ethical justification becomes weaker if the relevant community is treated only as a recruitment pool and has no meaningful role in setting priorities. [64, 70] Scientific value is not the same as immediate local availability of the final product. Some research develops knowledge with broad global importance; not every host country can be expected to adopt every intervention. Nonetheless, investigators should explain why the research question is relevant, how local knowledge has informed the design, and what can reasonably be expected in relation to post-trial access, affordability, infrastructure, training or public-health capacity. The expected benefits should be concrete enough to review, proportionate to what the study can deliver and honest about uncertainty. [44, 64]
Nuremberg Revisited 136 Fair selection and vulnerability The term “vulnerable population” is useful only if it leads to better decisions. Vulnerability can arise from a person’s limited decision- making capacity, but it can also be produced by circumstances: poverty, dependency, discrimination, institutional confinement, emergency, social exclusion or lack of realistic alternatives. A population is not inherently incapable of making decisions simply because it is poor, rural, Indigenous or has limited formal education. Treating such populations as uniformly vulnerable can itself become paternalistic and may exclude people from research relevant to their health. The more useful question is what specific feature of the situation creates a risk of being wronged and what protection addresses it. [43, 44, 64] The 2024 Declaration of Helsinki explicitly cautions that exclusion may perpetuate disparities as well as inclusion may cause harm. CIOMS similarly calls for carefully justified selection, additional safeguards where needed, and meaningful consideration of local circumstances. A study should not automatically exclude women, older adults or people with disabilities merely because they may be more complicated to recruit. Conversely, inclusion is not ethically fair if investigators choose these groups mainly because their access to care is limited or refusal is unlikely. The test is not whether a population is labelled vulnerable but whether its participation is necessary, fair and adequately protected. [44, 64] Community engagement without replacing individual consent Community engagement can improve research questions, explain local concerns, support recruitment that is not misleading, and strengthen the legitimacy of consent procedures. It is particularly relevant where community leaders, family structures or local institutions shape how health information is understood. However, collective agreement does not replace individual consent when an individual has capacity to decide. A chief, village
Nuremberg Revisited 137 council, hospital administrator or family representative may facilitate discussion but should not authorise the enrolment of an individual who declines. Equally, individual signatures do not always exhaust community interests, especially in research that may stigmatise a group or disclose information with collective implications. [44, 64, 71] Good community engagement is not ceremonial consultation after the protocol has been fixed. Communities and prospective participants should have opportunities to raise concerns about acceptable procedures, time burdens, language, specimen storage, privacy, return of findings and future use of data. Engagement can also continue after the trial through accessible summary results, discussion of implications and mechanisms for complaints. The aim is not to guarantee that every stakeholder agrees with every decision. It is to ensure that research does not use distance, complexity or unequal bargaining power to make important decisions invisible to the people most affected. [44, 64] Consent, standards of care and the place of the comparator Consent must be adapted to the circumstances without being reduced to a formal ritual. Translation alone is not enough if the explanation is too technical, the information is delivered under pressure or the participant cannot distinguish experimental intervention from routine treatment. Independent consent staff, teach-back questions, time for family discussion where desired, and clear explanations of alternatives may improve understanding. But a comprehension check should help the research team improve its communication, not turn participation into an examination of the participant. [44, 64, 65] Standards of care remain one of the most difficult areas of international research ethics. An investigator may argue that a trial should compare a new intervention against the care routinely available locally. Critics may reply that the “local standard” is inadequate precisely because people have been deprived of
Nuremberg Revisited 138 effective treatment. Neither claim settles the matter by itself. The protocol must explain why the comparator is scientifically needed, what proven interventions exist, what risks arise from withholding or delaying care, whether a different design is feasible, and whether the resulting evidence will address a meaningful health question. The 2024 Helsinki Declaration retains a general preference for comparison with the best proven intervention, while recognising specified circumstances in which placebo or no intervention may be acceptable. Any exception requires ethical justification; it is not a blanket permission to use inferior care wherever poverty makes it common. [44, 64] Post-trial obligations and benefit-sharing The ethical relationship does not necessarily end at the final study visit. A participant may have benefited from an intervention during the study; a community may need information about the result; a host institution may require training or technical capacity; and a country may face the prospect that a successful product is unaffordable. CIOMS and Helsinki encourage advance planning for post-trial access and fair arrangements for sharing benefits. Such planning should occur before recruitment, not after a positive result makes the question commercially inconvenient. [44, 64] “Benefit-sharing” should not be treated as a single formula. It may include appropriate post-trial access, capacity-building, infrastructure, local authorship and leadership, return of findings, contribution to public-health programmes, or reasonable access to the intervention if it is developed. These measures must fit the study and should not be exaggerated into promises that cannot be honoured. A sponsor cannot make an ethically adequate plan by stating that the host government will decide everything later; responsibilities should be assigned, budgeted, and reviewed. [44, 64]
Nuremberg Revisited 139 Ethical review across borders International collaboration requires oversight in more than one place. Local ethics review is important because reviewers familiar with the host context may notice linguistic, cultural, healthcare or political factors that an external committee misses. Review in the sponsor’s country also matters because the sponsor and its institutions remain accountable for design, funding, safety systems and use of the results. The 2024 Helsinki text expressly requires ethics approval in both sponsoring and host countries for collaborative international research. Dual review should not become a paper exercise in which each committee assumes the other has examined the hard questions. [44, 64] CIOMS is guidance rather than a treaty automatically enforceable in every jurisdiction. National laws, regulatory rules and institutional policies determine formal obligations. Yet the guidelines give a substantive framework for examining whether a cross-border project is ethically defensible even when legal minimums differ. The key question is not merely whether the paperwork has been approved. It is whether the partnership has been designed so that scientific aims, local health needs, participant rights and post-trial responsibilities are taken seriously. [52, 64, 65] The contribution of CIOMS is a shift in emphasis from a narrow model of participant protection to a broader ethic of fair partnership. Protection remains indispensable, but ethical international research must also ask who defined the question, who controls the evidence, who can use the resulting product, what institutions are strengthened, and what remains after the sponsor leaves. A country should not have to choose between refusing research altogether and accepting research under conditions that it would not regard as fair if the participants were its own citizens in a wealthier setting. [44, 64]
Nuremberg Revisited 140 Chapter 24. Ethics Committees and Independent Review: The Gatekeepers Under Pressure Independent review developed because the investigator’s commitment to science, however sincere, cannot be the only safeguard for the people asked to participate. Research teams are invested in their hypotheses, institutions in reputation and funding, and sponsors in product development. Each may have legitimate interests, but none should be the sole judge of whether the burden imposed on participants is acceptable. Research ethics committees (RECs), institutional review boards (IRBs) and, in India, ethics committees operating under national guidance and drug-trial regulations, are intended to provide a structured, independent challenge before and during research. [44, 52, 66, 67] Why independent review became necessary Clinical judgement is not enough because a research participant is not simply a patient receiving an individualised treatment recommendation. In a trial, procedures are often determined by a protocol designed to answer a broader question. A treating clinician may recommend participation, but the clinician’s role can create a conflict when care and research are combined. The ethics committee supplies an additional perspective: it examines whether the question is worth asking, whether the design can answer it, whether risks are proportionate, whether recruitment is
Nuremberg Revisited 141 fair, and whether participants will receive understandable information and appropriate care. [43, 44, 66] The United States system evolved from federal agency policies and later statutory and regulatory requirements. The National Research Act of 1974 strengthened oversight and created the National Commission that produced the Belmont Report. The Common Rule, adopted by multiple federal departments and agencies in 1991 and substantially revised in 2017, provides a shared regulatory framework for covered research. Its current provisions must be read alongside agency-specific rules, including the Food and Drug Administration’s requirements for research within its jurisdiction. It is not accurate to assume that every research activity in every institution is governed by an identical set of federal regulations; coverage depends on the funding, institution, intervention and legal framework. [43, 66] In India, the ethical framework has developed through successive ICMR guidance, including the National Ethical Guidelines for Biomedical and Health Research Involving Human Participants (2017). Clinical trials of new drugs and related regulated activities are also governed by the New Drugs and Clinical Trials Rules, 2019, as amended, and by relevant CDSCO requirements. These are related but not interchangeable instruments: professional ethical guidance is broader, while specific rules govern defined regulated activities, approvals, compensation, and ethics- committee registration. Current regulations and amendments must be checked against the official government texts. [52, 65] What a committee must examine A defensible ethics review begins with the scientific protocol because a study that cannot generate reliable knowledge cannot justify exposing participants to research burdens. The committee should then examine how the scientific plan affects real people: eligibility criteria, study procedures, recruitment, consent, foreseeable harms, expected benefits, clinical care, privacy, payment, research-related injury, follow-up, data management
Nuremberg Revisited 142 and dissemination. Sponsor funding and investigator conflicts should be disclosed, not hidden in documents that reviewers may overlook. The protocol should explain why each procedure is needed and how the study team will respond if risks change. [44, 64, 67, 70] Consent documents are a visible part of review, but reviewing words on paper is insufficient. The committee should ask who will approach participants, whether that person has a dependent relationship with them, whether the information is understandable, how interpretation will be provided, whether sufficient time for a decision is allowed, and how refusal will be respected. For participants who need additional safeguards, the committee should assess the specific reason and whether proposed protections address it. A generic sentence saying that a population is “vulnerable” is not a risk-management plan. [43, 44, 64] Review also includes decisions that arise after approval. Amendments may change risk; new safety information may alter the balance of benefits and burdens; recruitment may differ from the intended population; and a protocol may be unable to retain participants or maintain data quality. Ethics committees require reliable information to revisit their judgement. The 2024 Declaration of Helsinki states that committees must be able to monitor ongoing research, request changes, withdraw approval or suspend a study, and review amendments before implementation except where an immediate hazard requires urgent action. A committee unable to intervene once recruitment begins offers only partial protection. [44, 67, 68] Independence is more than a membership list A committee can appear diverse on paper and still lack real independence. Members may depend on the institution for employment, facilities, contracts, or professional advancement. A sponsor may fund the research and pay administrative costs. A chair may be influential but overburdened. Reviewers may lack
Nuremberg Revisited 143 sufficient time, training, or access to relevant expertise. Formal rules on conflict of interest help, but independence also depends on institutional culture: can a committee ask a powerful investigator to redesign a study, delay recruitment, or stop a project without punishment? [44, 66, 67] A sound review system therefore needs written procedures, declared conflicts, appropriate quorum, minutes, reasons for decisions, protected deliberation, suitable expertise, administrative support, and channels for escalating concerns. Reviewers should have access to the complete protocol and relevant investigator and sponsor information. Institutions should not measure a committee’s performance simply by the speed with which it approves proposals. A rapid decision can be appropriate when a protocol is clear and low-risk; speed alone is not evidence of quality. [67] Independence also requires balance. Researchers should be able to respond to reviewers’ questions, and committees should explain decisions rather than issue unexplained vetoes. Scientific disagreement should be resolved through reasoned assessment, not deference to institutional rank. External expertise may be necessary for complex methods or specialised populations, and community or public representation can help reviewers recognise how a procedure may be experienced outside the research team. [44, 67] The committee’s limits: review is not surveillance of everything No committee can directly observe every interaction between a recruiter and a participant at every site. It generally depends on submitted documents, site reports, monitoring information, and the integrity of investigators. Large multicentre studies create added challenges: different sites may implement consent differently, report harms at different speeds, or interpret protocol deviations inconsistently. Research teams must therefore be truthful and proactive in reporting, and sponsors must supply
Nuremberg Revisited 144 systems that allow safety signals and misconduct concerns to reach those who can act. [67, 68, 70] Ethics committee review should not be confused with data and safety monitoring. A data and safety monitoring board or committee, where appropriate, periodically examines accumulating data and may advise on continuing, modifying, or stopping a trial. An ethics committee focuses on the acceptability of the research and its continuing safeguards within its mandate. A regulator assesses compliance with legal requirements and may inspect, suspend, or sanction where empowered to do so. These functions overlap in participant protection but are not substitutes for one another. A safety board cannot legitimise invalid consent, and ethics approval does not prove the data are accurate. [52, 67, 68] Nor should the committee become a convenient place for institutions to transfer all moral responsibility. The investigator remains responsible for the conduct of the study; the sponsor remains responsible for its own design, systems, and oversight duties; and the institution remains responsible for the conditions under which research occurs. Committee approval is a necessary safeguard for many studies, not a certificate that future conduct will automatically be ethical. The Helsinki Declaration places responsibility for participant protection on the researchers and says it must never be shifted to participants simply because they consented. [44, 69] Review under Indian conditions For Indian researchers, it is important to distinguish biomedical and health research generally from regulated clinical trials of new drugs. The ICMR 2017 guidelines provide ethical guidance across a broad range of biomedical and health research involving human participants. The New Drugs and Clinical Trials Rules set requirements relevant to ethics committees reviewing clinical trials and related regulated studies, including registration and responsibilities connected with serious adverse events and
Nuremberg Revisited 145 compensation. Institutional committees need to know which framework applies to the specific study and whether additional regulatory approvals are required. [52, 65] The official CDSCO website provides the rules and related notices and records concerning ethics-committee registration, cancellations, and investigators subject to regulatory action. Such public information is relevant to transparency but must be interpreted carefully: a listed cancellation or debarment reflects an official regulatory act as described in the notice, and the reason, scope, date, and legal status should be checked in the underlying order. A registry entry should not be turned into an allegation beyond what the official document states. [52] Practical problems can include insufficient training, heavy workloads, limited access to independent statistical advice, dependence on the institution being reviewed, and difficulty following research after approval. These are systemic risks rather than proof that a particular committee acted improperly. Responses include regular training, transparent standard operating procedures, external review where independence is compromised, conflict-of-interest management, proportionate review pathways, accessible complaint mechanisms, and adequate funding for competent administration. Ethics review needs resources because independence without time and expertise is an aspiration, not a functioning system. [52, 65, 67] From Approval to Accountable Oversight The real test of an ethics committee is not whether a protocol has an approval number. It is whether the committee has considered the right questions, documented the reasons, retained independence, received relevant safety and progress information, and acted when the ethical basis of the study changed. Institutions should examine not only how many proposals are approved, but whether amendments and serious events are handled promptly, whether complaints reach a fair process, whether decisions are
Nuremberg Revisited 146 implemented, and whether participants can obtain answers without fear of retaliation. [44, 67, 68] Independent review cannot guarantee an ethical outcome. It can, however, create a structured point at which assumptions are challenged before human exposure begins and at which the decision to continue can be reconsidered. When research ethics committees are allowed to function as genuine independent guardians, their role is not to obstruct science. It is to protect the moral legitimacy of the scientific enterprise by ensuring that the people whose participation makes research possible are treated as persons rather than operational inputs. [44, 67]
Nuremberg Revisited 147 Chapter 25. Good Clinical Practice: Quality, Transparency, and Regulatory Accountability Good Clinical Practice (GCP) attempts to connect ethical principles with the operational details of a clinical trial. It deals with how a study is designed, staffed, documented, monitored, and reported; how responsibilities are divided between the investigator and the sponsor; and how records allow the reliability of results to be examined. GCP is important because misconduct can occur not only through explicit abuse but also through poorly designed protocols, missing safety information, unreliable records, selective reporting, or a chain of responsibility in which each organisation assumes that another will act. [45, 69, 70] Why operational standards matter Ethical statements establish that a participant’s rights and welfare must be protected. A trial still needs practical systems that turn this obligation into conduct: a clear protocol, adequately trained investigators, appropriately informed consent, secure records, reliable handling of investigational products, timely recording and reporting of harms, and an analysis that corresponds to the question specified before the data were observed. Without these systems, it may be impossible to determine whether the intervention helped, harmed, or made no meaningful difference. Unreliable results can also lead future clinicians to make decisions based on a false picture of benefit and risk. [69, 70]
Nuremberg Revisited 148 The International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) developed a common GCP framework to support the acceptance of clinical-trial data across regulatory regions while protecting participants. The original E6 guideline became an international reference point in the 1990s; a revised version, E6(R2), incorporated expanded expectations around quality management and risk-based monitoring; and E6(R3) has now moved the framework further toward quality by design, proportionality, and modern trial methods. ICH guidance is implemented through regional or national regulatory systems; its practical legal effect depends on the relevant jurisdiction. [45, 69] ICH E6(R3): a current transition As of 15 October 2026, the Principles and Annex 1 of ICH E6(R3) are the current version in the European Union and became effective on 23 July 2025. The consolidated E6(R3) text, including Annex 2, was published by the European Medicines Agency on 15 July 2026, but the page states a legal effective date of 15 January 2027. Annex 2 covers additional design features such as pragmatic clinical trials, decentralised trials and the use of real-world data. The distinction matters: a manuscript dated in October 2026 should not describe the entire consolidated version as already legally effective in the EU. National implementation can have its own details and dates, so local regulatory notices should also be checked. [45] The revised framework emphasises that quality should be designed into a trial, rather than inspected into it at the end. Investigators and sponsors are encouraged to identify factors critical to participant protection and reliable results, focus controls where failure would matter most, and use monitoring proportionate to the nature and complexity of the study. This is not permission to reduce safeguards merely to save money. Proportionality requires a reasoned match between risk and control. A high-risk, first-in-human medicine trial warrants different oversight from a study using routine data with minimal
Nuremberg Revisited 149 additional burden, but both require a scientifically defensible approach and protection appropriate to the people and information involved. [45, 69, 70] Responsibilities of investigators and sponsors The investigator is responsible for conducting the study at the site in accordance with the approved protocol, applicable requirements, and GCP. The investigator must ensure that delegated tasks are performed by suitably qualified personnel, that consent is obtained and documented appropriately, that investigational products are handled as required, and that source records and study data are accurate and accessible for authorised review. Delegation does not erase the investigator’s accountability for the conduct of the trial at the site. [69] The sponsor has a distinct set of responsibilities. These include ensuring appropriate trial design and resources, selecting qualified investigators and sites, establishing quality-management and safety-reporting systems, maintaining essential records, providing investigational products, ensuring appropriate monitoring, and addressing issues that threaten participant protection or data reliability. Some activities may be delegated to a contract research organisation, laboratory, or other service provider, but outsourcing a task does not mean that the sponsor can abandon oversight. Written agreements should make responsibilities clear, and the sponsor should retain effective supervision of critical functions. [69] These duties are complementary rather than competitive. An investigator may identify an unexpected safety concern before it is visible in aggregate data. A sponsor may identify a pattern across sites that an individual investigator cannot see. A monitor may detect source-data discrepancies; an ethics committee may identify a consent or fairness problem; and a regulator may discover systemic failures during inspection. A sound system allows these signals to be recorded, evaluated, and escalated
Nuremberg Revisited 150 without being suppressed by concerns about recruitment targets, timelines, reputation, or commercial return. [67, 68, 69] Consent, records and data integrity GCP treats consent as a process that must be properly documented and supported by adequate information. Documentation does not replace comprehension, and GCP compliance cannot rescue a misleading consent conversation. The information should match the current protocol, describe relevant foreseeable risks and alternatives, and explain whom participants can contact about questions or harm. If the protocol changes materially or new information alters the decision to participate, researchers must assess whether participants need to be informed and whether renewed consent is required under the applicable rules and ethics approval. [44, 52, 69] Data integrity means more than a database that passes validation checks. Data should be attributable to their source, legible, contemporaneous, original or appropriately verified, accurate and complete. Corrections should leave an audit trail. Missing values, exclusions, protocol deviations and changes to the analysis should be explained. Fabrication, falsification or selective omission can mislead regulators, clinicians and patients even when participant recruitment and consent were otherwise adequate. Conversely, a data error is not automatically fraud: the investigation must distinguish error, poor practice, negligence and intentional misconduct based on evidence. [69, 70, 72] Electronic records, remote visits, wearable devices and decentralised trial methods create new opportunities and risks. They may reduce travel and expand access, but they also raise questions about identity verification, privacy, digital literacy, device accuracy, cybersecurity, who can see data, and whether a remote participant can obtain help when a serious adverse event occurs. Proportionate GCP should adapt control systems to these methods rather than force every study into a single operational
Nuremberg Revisited 151 template. Innovation is ethically acceptable when its safeguards are designed for the real conditions of participation. [45, 69, 70] Risk-based monitoring and safety signals Monitoring should focus on whether the trial is being conducted as planned, whether participant rights and safety are protected, and whether data are reliable enough to support the study’s conclusions. Risk-based monitoring can include central statistical review, targeted on-site visits, remote review and examination of selected source records. The methods should be justified by the trial’s risks and critical data; using fewer routine visits is not inherently weak oversight, just as conducting many visits is not proof that serious risks will be detected. [69, 70] Serious adverse events and other important safety information require established routes for recording, assessment, and reporting. Investigators, sponsors, ethics committees, and regulatory authorities have different responsibilities and timeframes. The exact reporting threshold and timing depend on the governing law, protocol, and applicable guidance. Unexpected safety signals may require additional analysis, changes to recruitment or consent, new risk minimisation measures, temporary suspension, or termination. A functioning system must be capable of acting before the study’s statistical endpoint has been reached. [45, 52, 68, 69, 70] The ethical obligation is not to report every temporal event as if the investigational product caused it. Causality requires clinical and scientific assessment. But uncertainty should not be used to delay appropriate reporting or to dismiss a possible signal before it can be evaluated. Transparent safety reporting preserves the distinction between an event that happened after an intervention and an event caused by it, while allowing patterns to emerge from multiple participants and sites. [52, 68, 69]
Nuremberg Revisited 152 Registration and disclosure: the public record The integrity of a trial depends on more than what appears in a journal article. If only positive trials are published, clinicians may overestimate efficacy and underestimate harm. If outcomes are changed after results are known without transparent explanation, readers may be shown the most favourable version of the evidence. If a study is never registered, it may be difficult to learn that it existed, identify its original objectives or determine whether its findings remain unavailable. These practices waste participant contributions and can distort subsequent clinical decisions. [57, 58, 72] The International Committee of Medical Journal Editors (ICMJE) requires prospective public registration as a condition for consideration by its member journals, with registration at or before enrolment of the first participant under its current policy. The ICMJE makes clear that ethics approval does not substitute for registration. The World Health Organization’s 2017 joint statement on public disclosure called for prospective registration and timely disclosure of results, including negative and inconclusive findings. Registration policies, laws, and reporting obligations are not identical across every jurisdiction, but they provide complementary safeguards against invisible or selectively reported research. [57, 58] Registration is not a guarantee of transparency. Trial entries may be incomplete, outdated, or difficult to interpret; results may be posted in forms that are not linked to publications; and protocol amendments can make comparison with the final report difficult. Good practice requires the registry to be updated, the final analysis to be explained, and publications to identify the registration number. Where results remain unavailable, institutions and funders should be able to explain why and to require corrective action. The public record should make it possible to follow a study from its intended question to its final conclusions. [57, 58, 72]
Nuremberg Revisited 153 GCP and the Indian regulatory setting In India, clinical trials of new drugs and specified regulated activities are governed by the New Drugs and Clinical Trials Rules, 2019, as amended, together with the applicable provisions of the Drugs and Cosmetics framework, CDSCO directions, and other relevant requirements. The Rules address matters such as permission to conduct regulated studies, ethics committee functions and registration, informed consent, and the treatment and compensation of participants harmed in relation to clinical trials. Investigators and sponsors should consult the current official version and subsequent amendments rather than rely solely on a secondary summary. [52] The ICMR National Ethical Guidelines for Biomedical and Health Research Involving Human Participants (2017) apply more broadly across biomedical and health research. They complement rather than replace statutory rules where those rules apply. The distinction matters for investigator-initiated research, public- health research, studies using biological material or data, and trials regulated as drug development. Researchers should establish at the design stage which approvals, registrations, consent requirements, privacy controls, and reporting duties are relevant. [52, 65, 71] Regulatory accountability includes more than prior permission. Sponsors and institutions need mechanisms to detect breaches, correct deficiencies, report safety concerns, preserve essential records and cooperate with inspections. Public regulator notices may record registration cancellations, suspensions, debarments, or other actions. Such records are important indicators of enforcement, but they should be interpreted using the underlying order, its legal basis and procedural status. Regulatory action is not a licence for commentators to state additional allegations as proven facts. At the same time, an enforcement system that issues rules but does not publish meaningful information about compliance will be difficult for participants and the public to evaluate. [52]
Nuremberg Revisited 154 GCP cannot replace ethical judgement GCP is necessary, but the most important ethical questions cannot all be answered by a checklist. A study can meet technical requirements and still ask a question of little social value, recruit a population unfairly, offer a comparator that is ethically difficult to justify, or fail to provide a credible plan for post-trial access. A sponsor may have complete documentation yet negotiate publication arrangements that constrain academic independence. A consent form may be compliant but incomprehensible. Quality systems must be used to realise substantive ethical commitments, not as substitutes for them. [44, 64, 69, 70, 72] The 2021 revision of ICH E8(R1) reinforces the importance of patient protection, robust science, patient input and quality by design across clinical development. Together, E8(R1) and E6(R3) place greater emphasis on identifying what matters most to a study’s validity and participant protection. This approach can reduce waste by preventing avoidable design failures before a trial begins. It should also encourage sponsors and investigators to ask participants and communities what burdens may be overlooked in a protocol written primarily from the perspective of operational efficiency. [45, 69, 70] The broader lesson is that accountability must run through the entire research lifecycle: from selection of the question to protocol design, review, recruitment, consent, conduct, safety monitoring, analysis, registration, publication, and post-trial responsibilities. No single institution can perform every function. Ethical research depends on a system in which each actor knows its obligations, records its decisions, reports problems, accepts independent scrutiny, and provides a route to remedy when the system fails. This is how a good clinical practice framework becomes more than good paperwork. [44, 45, 52, 57, 58, 67, 68, 69, 70, 72]
Nuremberg Revisited 155 Synthesis: The Architecture of ethical accountability The instruments discussed in Part V have distinct roles. They should not be collapsed into a single “international code”, nor should their differences obscure their shared purpose. The table below is a practical orientation for readers; it is not a legal opinion and does not replace the text of any instrument. Instrument or mechanism Primary contribution Important limitation Nuremberg Code (1947) Foundational post-war principles, especially voluntary consent and investigator responsibility. Not a comprehensive, globally enforceable regulatory system by itself. [62] Declaration of Helsinki (1964; current text 2024) Professional ethical framework for medical research, independent ethics review, consent, risk assessment and post-trial duties. A declaration, not automatically a uniform statute in every jurisdiction. [44] Belmont Report (1979) Three principles and their applications: consent, risk-benefit assessment and fair subject selection. Framework requiring interpretation; rooted in a particular US historical and legal context. [43] CIOMS Guidelines (2016) Detailed international guidance for applying ethical principles across varied contexts, including lower- resource settings. Guidance; national laws and local implementation still matter. [64] Ethics committees / IRBs Independent protocol review and continuing oversight. Effectiveness depends on independence, competence, resources and timely information. [44, 66, 67] ICH GCP (E6(R3)) Operational standards for trial conduct, quality, responsibilities, Regional/national implementation varies; compliance
Nuremberg Revisited 156 Instrument or mechanism Primary contribution Important limitation records and participant protection. does not resolve every ethical question. [45, 69] National laws and regulations Legally enforceable permissions, duties, inspections and remedies in a defined jurisdiction. Coverage, enforcement capacity and remedies vary; a rule on paper is not proof of compliance. [52, 66] Trial registration and results disclosure Public visibility of planned studies and their findings, including negative or inconclusive results. Effectiveness depends on completeness, timeliness, linkage and enforcement. [57, 58] Six questions for testing the system 1. Scientific value: Is the question important, the design valid, the comparator defensible and the analysis planned before results are known? 2. Voluntary participation: Does the person understand what is research, what is uncertain, and how to refuse or withdraw without reprisal? 3. Fair selection: Are recruitment and exclusions justified scientifically, rather than by convenience, dependency or weak bargaining power? 4. Independent oversight: Is review genuinely independent, adequately resourced and able to respond to new information after approval? 5. Operational accountability: Are the investigator, sponsor, institution, monitor, ethics committee and regulator clear about their responsibilities? 6. Public trust: Is the trial prospectively registered, are results—including negative findings—disclosed, and are participants and communities informed about what was learned?
Nuremberg Revisited 157 The history of research ethics shows that safeguards become meaningful only when they change behaviour. A code can express a principle; a report can explain it; a committee can challenge a protocol; a regulator can enforce a rule; and a registry can make a study visible. Each mechanism is incomplete when used alone. An ethical research culture is one in which investigators welcome scrutiny, sponsors budget for participant protection, institutions support independent judgement, regulators act on evidence, journals insist on transparency, and participants can ask questions without being treated as obstacles to progress. Part V therefore marks a transition in the book. The next part examines COVID-19, a period in which urgency, uncertainty, public expectation, and commercial development placed research and public-health decisions under exceptional pressure. The same distinctions established here will be essential: research participation is not the same as public-health policy; regulatory authorisation is not itself proof of a trial’s ethical adequacy; and disagreement with a policy is not by itself evidence of misconduct. Each claim must be evaluated using its own documentary record, scientific context, and applicable standard.
Nuremberg Revisited 158 PART VI COVID-19: ETHICS UNDER EMERGENCY CONDITIONS Chapters 26–30
Nuremberg Revisited 159 Chapter 26. When Emergency Becomes a Justification Speed, Scientific Value and Ethical Review The Moral Urgency to learn In an emergency, research is not a luxury postponed until the crisis passes. It can be a moral obligation. When clinicians face a new disease and must choose among interventions of uncertain value, the failure to generate trustworthy evidence can itself cause harm. Ineffective treatment can expose patients to adverse effects, consume scarce resources, and displace effective care. A well- designed trial can answer questions faster and more reliably than repeated small, uncontrolled studies. The ethical challenge is therefore not whether research should continue during an emergency. It is how to accelerate useful research without treating urgency as permission to weaken the protections that make research legitimate. WHO's March 2020 guidance distilled established standards for research in public- health emergencies, while its May 2020 guidance advised ethics committees on rapid review procedures. The documents recognised the need for speed, but they did not present emergency review as an exemption from scientific validity, independent scrutiny, informed consent, fair participant selection or public accountability [75, 76]. This distinction matters because speed and quality are not opposites. A coordinated, multicentre trial with a shared protocol can recruit efficiently and reduce duplicated effort. A poorly
Nuremberg Revisited 160 designed study can recruit quickly yet leave the central question unanswered. Its participants have still accepted risk, investigators have still used public or charitable resources, and patients may still be influenced by its claims. A negative or inconclusive study is not necessarily a failure; a study that could not answer its stated question is a more serious waste. Rapid review is still review Ethics committees had to review new protocols while hospitals were crowded, staff were ill or redeployed, and ordinary meetings were difficult. Regulators and funders faced simultaneous applications from many research groups. WHO and the Pan American Health Organization recommended practical measures such as emergency standard operating procedures, remote meetings, prioritisation, coordination among review bodies, and clear pathways for continuing oversight [75, 76]. India's ICMR issued national guidance for ethics committees reviewing biomedical and health research during the COVID-19 pandemic, along with a template for adapting procedures [77]. A defensible rapid review asks the same essential questions in a more organised way. Is the scientific question important? Is the design capable of answering it? Does the expected value justify the foreseeable risks and burdens? Are the inclusion and exclusion criteria fair? Is the consent process comprehensible and voluntary? Are there adequate arrangements for clinical care, safety reporting, privacy, compensation where applicable, and withdrawal? Who will monitor accumulating evidence and stop the study if the balance changes? The committee need not reproduce every administrative step used in routine conditions if an equivalent, documented safeguard can be achieved by a faster method. It must, however, retain the substance of independent review. A meeting held within hours may be rigorous; a meeting held after weeks may be perfunctory. The ethical measure is not the interval alone but the quality, independence, documentation, and follow-up of the decision.
Nuremberg Revisited 161 The pressures that threaten independent judgment Pandemics magnify institutional pressures. Investigators may fear that a protocol delayed for review will miss the epidemic peak. Hospitals may view research participation as a route to scarce equipment, drugs or funding. Sponsors may face commercial or geopolitical deadlines. Governments may want quick results to reassure the public or justify policy. Media outlets may turn an interim finding into a headline before the data are mature. An ethics committee can be technically independent yet operate in an environment where saying 'not yet' is treated as obstructive. These pressures are not proof that a specific committee acted improperly. They are predictable systemic risks and should be managed prospectively. Emergency procedures should disclose conflicts of interest, record reasons for accelerated decisions, protect committee members from undue pressure, require amendments to be reviewed, and define how adverse events and new external evidence will be assessed. A decision should be traceable after the emergency, not merely remembered by its participants. The review must also examine scientific equipoise: genuine uncertainty among the relevant expert community about the comparative merits of the interventions being studied. A treatment's theoretical mechanism, laboratory activity or success in another disease does not establish clinical benefit in COVID-19. Equally, an intervention should not be rejected simply because it is unfamiliar. The question is whether the available evidence and the study design create a defensible reason to expose participants to the proposed intervention or comparator. Consent in crowded wards and through a screen Obtaining consent during respiratory isolation, severe illness, delirium, breathlessness, or fear presented practical and moral difficulty. Remote consent, electronic forms, telephone explanations, and authorised representatives could sometimes preserve participation while reducing infection risk. But they
Nuremberg Revisited 162 could also create barriers: patients may struggle with small screens, poor connectivity, language, literacy, hearing, or the cognitive effects of illness. A signature or recorded click cannot establish comprehension by itself. Researchers should distinguish a patient capable of making a decision from one temporarily or persistently unable to do so. Where consent from a legally authorised representative is permitted, the applicable rules and the participant's own wishes should still guide the process. Waivers or alterations of consent may be ethically and legally permissible in narrowly defined circumstances, particularly for certain minimal-risk activities, but they require an appropriate authorisation and justification. Emergency conditions do not make every study minimal risk and do not automatically permit the enrolment of unconscious patients in any available protocol. Community engagement is similarly more than a public-relations exercise. In communities where research has a history of exploitation, explanations need to be culturally and linguistically accessible, responsive to local concerns and clear about what is known and unknown. Engagement may identify concerns that are invisible to investigators, including practical costs of participation, mistrust arising from past conduct, or the fear that declining a study will affect ordinary care. The obligation does not end at enrolment Research ethics extends beyond the signature and the last study visit. Safety monitoring must continue; protocol deviations must be recorded; and new findings should be communicated to participants when they may affect decisions. Trial registration helps others distinguish a genuine absence of evidence from evidence that has never been disclosed. The WHO's 2025 guidance on reporting summary results in trial registries again emphasised that results reporting remains incomplete and that unpublished trials can leave the evidence base distorted [78]. This
Nuremberg Revisited 163 is a post-pandemic reminder that a crisis does not excuse silence after the crisis has passed. Emergency studies can also leave participants with continuing needs. If an intervention is shown to be beneficial, arrangements for post-trial access, referral, follow-up and explanation of results should be addressed before the first participant is enrolled. Not every successful experimental intervention can immediately be supplied to every participant indefinitely; availability, safety, regulation and health-system capacity matter. The ethical obligation is to plan transparently and fairly rather than promise benefits that cannot be delivered. A Practical Emergency Review Test The COVID-19 experience suggests a concise test for future emergencies: 1. Necessity: Is the question important enough to justify human participation now? 2. Validity: Can the design produce an interpretable answer, and is the protocol aligned with the current evidence? 3. Proportionality: Are foreseeable risks minimised and reasonable in relation to anticipated social and scientific value? 4. Voluntariness and fairness: Is consent meaningful, and are participants chosen for scientifically relevant reasons rather than convenience or vulnerability? 5. Independence: Are review, safety monitoring and stopping decisions protected from sponsor, political and institutional pressure? 6. Transparency: Will the protocol be registered, methods disclosed, results reported and changes explained? 7. Continuity: Are follow-up, new safety information, compensation and post-trial arrangements addressed?
Nuremberg Revisited 164 A public-health emergency can justify accelerating the route to a decision. It cannot justify abandoning the reasons why the decision deserves public confidence. The lesson is not that ethics must slow science. It is that sound ethical systems help science move quickly in a direction worth following.
Nuremberg Revisited 165 Chapter 27. COVID-19 Treatment Trials Hope, uncertainty, and the cost of bad evidence The rush from possibility to prescription At the start of 2020, clinicians confronted a disease for which no treatment had yet been proved to reduce mortality. Physicians tried existing medicines, laboratory findings generated hypotheses, and early case reports spread rapidly through professional and public networks. Such steps are understandable when patients are dying. They become ethically dangerous when the level of certainty is exaggerated or when experimental treatment is offered as established care without a clear explanation of its uncertain benefit and potential harms. A drug can have antiviral activity in a laboratory and still fail in patients because the dose, timing, distribution, immune response, or disease mechanism is different. A treatment used successfully for another condition may have a different risk-benefit balance in acutely ill patients. Mechanistic plausibility is a reason to investigate, not a substitute for clinical outcomes. A small uncontrolled series can generate a hypothesis; it ordinarily cannot establish whether the observed recoveries were caused by treatment rather than natural history, selection, supportive care or chance. This was not a problem unique to COVID-19, but the pandemic accelerated it. Political leaders, public figures, commentators, and
Nuremberg Revisited 166 sometimes clinicians promoted particular medicines before robust clinical evidence was available. Some patients sought them because no alternative seemed satisfactory. Others may have interpreted preliminary results as a guarantee. The ethical duty of investigators and clinicians was to state uncertainty plainly, avoid false reassurance, monitor toxicity, and wherever feasible enrol eligible patients in well-designed trials capable of resolving the uncertainty. Hydroxychloroquine and the fragility of an evidence system Hydroxychloroquine became one of the best-known examples of a treatment moving rapidly from preliminary hope to public controversy. Laboratory hypotheses and small clinical reports were followed by intense political and media interest. A multinational observational analysis published in The Lancet in May 2020 reported associations between hydroxychloroquine or chloroquine-based regimens and adverse outcomes. Serious concerns were raised about the data supplied by Surgisphere and the ability to validate the underlying records. The paper was retracted in June 2020 after independent reviewers could not obtain the data needed to complete their assessment [79, 80]. The retraction should be described accurately. It demonstrated a serious failure in the evidentiary and publication process for that paper; it does not, by itself, establish every allegation later made about the authors, sponsor, journal, or wider medical system. Nor does the failure of one article show that all peer review is worthless. Its lesson is that peer review cannot replace access to credible source data, independent verification, transparent methods, and accountability when doubts arise. Randomised trial evidence subsequently answered the clinical question more reliably. WHO announced in June 2020 that the hydroxychloroquine arm of its Solidarity trial would stop enrolling participants after evidence from Solidarity, the UK's RECOVERY trial, and other evidence showed no reduction in mortality among
Nuremberg Revisited 167 hospitalised patients compared with standard care [81]. This sequence illustrates how a transparent trial platform can permit a treatment to be tested, stopped, and de-prioritised without treating every early hypothesis as misconduct. Solidarity: coordination at international scale WHO's Solidarity trial was designed to evaluate candidate treatments across many countries using a common protocol. Its broad, pragmatic design sought to produce results relevant to hospital practice while avoiding a proliferation of small, isolated studies. In the interim results published in 2021, remdesivir, hydroxychloroquine, lopinavir and interferon regimens showed little or no effect on mortality among hospitalised patients, or on initiation of ventilation or duration of hospitalisation, in the comparisons reported [82]. These findings were specific to the regimens, populations, outcomes, and period studied; they were not a universal statement about every use of every drug in all circumstances. Large international trials can improve statistical power and generalisability, but they are not automatically equitable. Host hospitals need capacity to consent patients, deliver the protocol, document outcomes and report harms. Local investigators should participate in design and analysis, not merely recruitment. Researchers must also explain to participants that randomisation allocates treatment by chance and that the trial is designed to learn which approach works better, rather than to guarantee access to the treatment a participant prefers. RECOVERY and the value of an honest negative result The UK's RECOVERY platform provided a counterexample to the notion that emergency trials were necessarily weak or unproductive. In June 2020, preliminary results indicated that dexamethasone reduced mortality in hospitalised patients who were receiving oxygen or invasive mechanical ventilation, but not
Nuremberg Revisited 168 in patients who did not require respiratory support. The peer- reviewed report described a randomised comparison of dexamethasone with usual care, and WHO welcomed the findings because they provided evidence of a treatment that could save lives in specified clinical groups [83, 84]. The distinction by severity was ethically and clinically important. Dexamethasone was not established as beneficial for every person with COVID-19 at every stage. Administering it indiscriminately could obscure the evidence-supported indication and create avoidable harm. The result also demonstrates why research should report effect estimates, confidence intervals, study populations and subgroup limits rather than compressing a nuanced result into 'the drug works'. The same principle applies to ineffective interventions. A negative trial can protect future patients from unnecessary exposure and allow resources to be redirected. Refusing to publish a negative result because it disappoints a sponsor, investigator or political constituency would betray the social value used to justify enrolling participants in the first place. Compassionate use, off-label treatment, and clinical trials Four activities are often blurred in public discussion. Routine clinical care applies treatments supported by the standard of care. Off-label prescribing uses a licensed medicine outside the terms of its product authorisation, subject to local law and professional duties. Compassionate or expanded access may provide an investigational product to a patient who cannot enter a trial and for whom appropriate alternatives are lacking. A clinical trial follows a protocol intended to generate generalisable evidence. These activities may overlap in real practice, but their ethical purposes and safeguards are not interchangeable. Compassionate access can be humane when a patient faces serious illness and no satisfactory options. Yet it has limits: there may be little information about dose, interactions or toxicity; access can
Nuremberg Revisited 169 be unequal; and a product's availability can create the mistaken impression that it is effective. Where a trial is feasible and clinically appropriate, steering potential participants toward unstructured access can impair research needed by future patients. Conversely, a trial should not be used to deny necessary standard care or to promise a therapeutic advantage it cannot guarantee. Research integrity under pressure COVID-19 also tested the integrity of trial registration, data management, publication, and correction. An article in a major journal is not self-authenticating. The Surgisphere episode showed how a paper using a large, difficult-to-verify dataset could influence discussion before doubts made a full validation possible. The journal's subsequent editorial reflections described changes to declarations and data-sharing processes [79]. The appropriate response is stronger systems for data provenance, authorship accountability, methods transparency, and prompt correction— not automatic cynicism about all findings. For future emergencies, treatment trials should be registered prospectively, use clinically meaningful outcomes, publish protocols and amendments, provide independent safety monitoring where appropriate, make results public even when negative, and give readers enough information to judge applicability. Public-health communication must also distinguish peer-reviewed results from preprints, press releases, interim analyses, and anecdotes. The speed of a news cycle should not dictate the certainty of a medical recommendation. The ethical lesson The history of COVID-19 treatments is not a simple story of bad actors followed by good science. It is a case study in the difficulty of making decisions when evidence is incomplete and the costs of delay are visible. Some early hypotheses were rational to test; some claims outran their evidence; some highly visible analyses proved unreliable; and well-designed trials identified both
Nuremberg Revisited 170 ineffective interventions and a treatment with demonstrable benefit. The ethical standard is not that investigators must predict the answer. It is that they ask valid questions, protect participants, speak honestly about uncertainty and allow evidence to correct mistaken beliefs.
Nuremberg Revisited 171 Chapter 28. COVID-19 Vaccine Trials Acceleration, authorisation and continuing obligations What “unprecedented speed” did—and did not— mean The rapid development of COVID-19 vaccines became both a scientific achievement and an ethical test. Research advanced in parallel: platform technologies had been developed before 2020; governments and organisations invested at risk in manufacturing capacity; regulators communicated with developers during development; and large trials enrolled participants while transmission was occurring. These factors compressed timelines. They do not support the claim that every vaccine bypassed clinical testing, nor do they establish that every decision was beyond criticism. The central ethical question is more exacting: What evidence was available at each authorisation point, what uncertainties remained, what information did participants and recipients receive, and what systems were in place to learn from subsequent use? The answer varies by vaccine, jurisdiction, date and regulatory pathway. A book of research ethics must avoid discussing all vaccines or every national authorisation as if the same data package and decision rule applied to each. The pivotal randomised trials for the first mRNA vaccines enrolled tens of thousands of participants and used prespecified outcomes. The published BNT162b2 trial reported about 95%
Nuremberg Revisited 172 efficacy against symptomatic COVID-19 in its primary analysis; the mRNA-1273 trial reported 94.1% efficacy against symptomatic infection in its primary analysis [85, 86]. These were important results, but their primary efficacy endpoints were defined clinical outcomes during a specified follow-up period. They did not establish permanent protection, zero risk of infection, elimination of transmission, or the absence of every rare or long-term adverse event. Emergency authorisation is not a synonym for full approval Regulatory terminology matters. An emergency use authorisation, conditional authorisation, provisional approval and full marketing approval arise under different legal systems and may impose different evidence and follow-up requirements. An emergency pathway can be justified in a serious public-health threat when the evidence is sufficient for the applicable standard and the expected benefits are judged to outweigh known and potential risks. It also makes the disclosure of uncertainty and the continuation of evidence generation particularly important. The US Food and Drug Administration explains that manufacturers receiving an EUA are expected to continue their clinical trials to obtain additional safety and effectiveness information and pursue licensure. It also describes the obligation to inform recipients of the emergency status, known and potential benefits and risks, uncertainties, option to accept or refuse, and available alternatives [87]. WHO's 2022 statement for health professionals similarly explains that emergency or provisional decisions commonly used at least about two months of follow-up after the final vaccine dose in pivotal trial participants, with longer follow-up required for full authorisation and continuing population surveillance after rollout [88]. These are descriptions of regulatory practice, not claims that two months can reveal every possible rare or delayed event.
Nuremberg Revisited 173 A safety database of several tens of thousands of trial participants can identify common adverse events and many clinically important signals, but may be too small to detect events occurring very rarely. That statistical limitation is familiar in vaccine development. It creates a reason for active and passive post- authorisation surveillance, transparent signal assessment and continued follow-up, not a reason to assume in advance that a rare event either must or cannot be related to vaccination. What trial efficacy does—and does not—tell us Efficacy is not a single, context-free property. It depends on the outcome measured, the follow-up interval, population characteristics, circulating variant, prior infection, dose interval, and the trial's statistical plan. A trial designed primarily to detect symptomatic disease may provide limited direct evidence about asymptomatic infection or transmission. A study with a few severe cases may have wide uncertainty about protection against rare outcomes even when its primary endpoint is met. Later variants and changing background immunity can alter real-world performance without retrospectively making the original trial result fraudulent. This is why informed consent should explain the trial's purpose and limitations in language participants can understand. Participants need to know whether the candidate vaccine is experimental, what comparator is used, which outcomes the investigators will measure, what safety monitoring is planned, how randomisation works, and what clinical care remains available outside the trial. Investigators should not promise that participation will prevent infection or benefit the individual personally when those claims are not supported. India: Covishield, Covaxin and the problem of transparency India's experience illustrates the importance of distinguishing a regulatory decision from the later publication of a trial report. On
Nuremberg Revisited 174 3 January 2021, the Government of India's public statement on the Drugs Controller General of India described restricted emergency approvals for Covishield and Covaxin following recommendations of the Subject Expert Committee [89]. The evidentiary histories were not identical. Covishield drew on safety, immunogenicity, and efficacy data from overseas clinical studies presented by its manufacturer. Covaxin's authorisation drew public scrutiny because Phase III efficacy results were not yet available as a full peer-reviewed journal publication at the time of the initial decision. The Phase III BBV152 report was later published in The Lancet in November 2021. It described a randomised, double-blind, placebo-controlled multicentre trial conducted at 25 Indian sites, with an interim efficacy estimate of 77.8% against symptomatic COVID-19 in the specified adult analysis and a median of 99 days of follow-up for the study population at the data cutoff reported [90]. The report also disclosed funding and author affiliations, including roles at Bharat Biotech and ICMR. These details are essential to the interpretation of the publication; the existence of funding or an institutional affiliation does not, by itself, establish misconduct. The ethical issue is not resolved by saying either that the vaccine was authorised and therefore every concern was answered, or that its authorisation preceded a peer-reviewed publication and therefore the vaccine had no evidence. The proper questions concern the data package available to the regulator at the time; the strength and limits of preliminary data; the reasoning and conditions attached to the decision; what information was publicly disclosed; whether uncertainty was communicated accurately; and whether subsequent evidence was collected and made accessible. Those questions require the contemporaneous regulatory record, not hindsight alone.
Nuremberg Revisited 175 Placebo recipients and the changing standard of prevention Before a vaccine is proven and available, a placebo-controlled trial can be scientifically efficient and ethically acceptable if equipoise remains, participants give valid consent and risks are reasonable. Once an effective, authorised vaccine becomes locally available to an enrolled participant who is eligible to receive it, the ethical balance changes. Continuing to keep that participant on placebo solely to preserve a pristine comparison may deny access to an intervention that is now part of the locally available prevention standard. WHO's November 2021 policy brief addressed precisely this problem. It advised that, where authorised vaccines are available locally and participants meet the relevant eligibility criteria, trial teams should inform participants of their option to receive those vaccines and consider progressive unblinding of placebo arms. It also discussed alternative trial designs for modified and next- generation vaccines [91]. Such transitions can reduce the statistical power and duration of the original placebo comparison; they may nevertheless be ethically required. The scientific problem created by an appropriate ethical decision should be addressed through design, long-term follow-up, immunobridging or other valid methods—not by ignoring participants' interests. Safety surveillance: a signal is a question, not a verdict Post-authorisation pharmacovigilance is essential because widespread use exposes a vaccine to populations larger and more diverse than the pivotal trial. Different systems detect different patterns. Passive reporting systems can raise a signal quickly but are affected by under-reporting, stimulated reporting, incomplete clinical details and the absence of a suitable comparison group for many analyses. Active surveillance and linked health data can improve signal evaluation, while clinical review and
Nuremberg Revisited 176 epidemiological methods assess timing, background rates, plausible mechanisms and alternative explanations. The US CDC emphasises that a report to VAERS is not proof that a vaccine caused the reported event and that VAERS data alone cannot establish causation or usually determine incidence [92]. This methodological point is not a reason to dismiss reports. A credible safety system receives reports, examines patterns, communicates uncertainty and investigates signals. Dismissing every report would be as unsound as counting every temporal association as vaccine-caused injury. The same standard of openness applies when evidence favours vaccination. Regulators and sponsors should make protocols, analysis plans, relevant trial results, safety updates and reasons for major changes accessible, subject to legitimate protections for privacy and confidential personal data. Public trust is weakened if authorities ask people to accept a risk-benefit judgment while withholding the reasoning or appearing to minimise uncertainty. What participants are owed after the trial Trial participants contribute more than a vial of blood or a data point. They contribute time, inconvenience, biological risk and trust. Sponsors and investigators should plan for continued safety follow-up, access to relevant findings, care and compensation under applicable law, and appropriate access to an intervention shown to be beneficial when the relevant ethical guidelines and feasibility considerations require it. The 2024 revision of the Declaration of Helsinki reinforces the importance of research ethics review, transparency and post-trial arrangements; these are continuing obligations rather than ceremonial text [44]. The pandemic demonstrated both what can be achieved by parallel development and what should not be sacrificed: truthful consent, independent oversight, accessible evidence and a credible system for learning after authorisation. Acceleration is not the enemy of ethics. Unexplained uncertainty, opaque decisions and failure to follow the evidence are.
Nuremberg Revisited 177 Chapter 29. Mandates, Bodily Autonomy, and Public- Health Ethics A different question from research consent Why the concepts must not be collapsed A research participant's consent to enter a clinical trial and a citizen's decision about an authorised vaccine are related by respect for bodily integrity, but they are not the same ethical or legal question. A clinical trial is organised to generate knowledge; participation ordinarily requires a voluntary, informed decision under the applicable research framework. A public-health programme applies an intervention to protect health, and the state or an employer may debate conditions for access to particular settings or activities. Such rules raise questions of public-health necessity, proportionality, equality and law. One cannot settle the ethics of a mandate merely by citing a rule about clinical-trial consent, nor can a public-health justification erase consent duties inside a vaccine trial. This distinction is important in a history of research ethics because debates about COVID-19 sometimes moved rapidly between different kinds of claim. The question whether a vaccine was tested ethically differs from whether its authorisation was justified; both differ from whether vaccination should be required for a health-care worker, a school attendance condition or general access to public services. Evidence relevant to one decision may
Nuremberg Revisited 178 inform another, but the standards and burdens of justification are not interchangeable. The ethical case for and against mandates Public-health arguments for vaccination requirements include reducing severe disease, protecting people at high risk, preserving the functioning of essential services and, where evidence supports it, reducing transmission in a particular setting. A mandate may be defended where voluntary uptake is inadequate, the intervention is sufficiently safe and effective for the intended outcome, the risk to others is substantial, and less restrictive approaches are unlikely to achieve the same goal. Such an argument is strongest when the benefit is clear, the targeted group faces a substantial risk, supply is reliable and implementation is fair. Arguments against mandates emphasise bodily autonomy, the possibility of medical contraindications, uncertainty about benefits for particular populations or outcomes, mistrust arising from poor communication, unequal access and the risk that penalties will burden people with less power. Mandates can also backfire if they are framed as punishment rather than justified protection. A policy that imposes the same burden on groups with very different baseline risks or different access to vaccination may be formally uniform but substantively unfair. Neither position can be resolved by slogans. 'Individual choice' does not eliminate the government's responsibility to protect others from preventable harm. 'Public health' does not make every restriction proportionate. The ethical task is to establish the relevant objective, examine the evidence for achieving it, consider the severity and distribution of burdens, evaluate alternatives and make the decision open to review as circumstances change.
Nuremberg Revisited 179 Necessity, proportionality and least-restrictive means WHO's 2021 policy brief on mandatory COVID-19 vaccination identifies ethical considerations and caveats for policy-makers. It emphasises the need to assess whether a mandate is necessary and proportionate, whether vaccine supply is sufficient and reliable, whether affected people have been engaged, and whether policies discriminate or disadvantage vulnerable groups [93]. These tests are particularly important when the disease, intervention and epidemiological context are changing. Necessity asks whether the public-health goal is sufficiently important and whether a mandate is likely to help achieve it. Proportionality compares the expected benefit with the restriction and burden imposed. Least-restrictive means asks whether education, convenient access, testing, ventilation, masks in high- risk settings, paid sick leave, or role-specific precautions can achieve much of the intended protection with less coercion. Not every alternative will work equally well in every environment; the analysis should be empirical and context-sensitive. The evidentiary basis must also match the mandate's objective. Evidence that an intervention reduces severe disease may support a policy designed to protect high-risk people or preserve clinical capacity, but may be insufficient by itself to support a claim that the intervention reliably prevents all infection or transmission. When protection against infection wanes or varies by variant and time since vaccination, a policy justified on transmission prevention needs reassessment. Conversely, breakthrough infections do not by themselves negate evidence of reduced risk of severe outcomes. Ethical policy describes precisely which benefit is being pursued and which evidence supports it. India and the Supreme Court judgment of 2 May 2022 In Jacob Puliyel v. Union of India, the Supreme Court of India considered challenges relating to COVID-19 vaccine clinical-trial
Nuremberg Revisited 180 data, transparency, adverse-event information, and mandates [94]. The Court recognised bodily integrity and personal autonomy under Article 21 and stated that no individual could be forced to be vaccinated. It also concluded that the Union's overall vaccination policy could not be called unreasonable or manifestly arbitrary on the material presented, including expert evidence of vaccination's benefits against severe disease. The judgment separately examined restrictions imposed by State Governments and Union Territories on unvaccinated people. Given material before the Court concerning transmission and the epidemiological situation at that time, the Court considered those restrictions disproportionate and suggested that authorities review such orders while infection rates remained low, subject to new research or developments that could justify reasonable and proportionate restrictions. This was a context-specific assessment, not a timeless declaration that any public-health restriction related to vaccination is unlawful in all circumstances. The case is relevant for more than its outcome. The petition raised concerns about segregated trial data, the reasoning behind regulatory decisions, committee records, and adverse-event reporting. The Court's analysis demonstrates that regulatory expertise is important but not inherently beyond scrutiny; at the same time, courts must examine evidence rather than substitute unsupported scientific assertions for it. A serious research-ethics system requires both competent expert assessment and procedures through which decisions can be explained and challenged. Consent, refusal, and the risk of coercion A decision can be formally voluntary and still occur under strong practical pressure. A person who must choose between vaccination and loss of livelihood may experience a very different decision environment from a person with a secure job and easy access to medical advice. Research ethics has long recognised undue influence and vulnerability; public-health policy should
Nuremberg Revisited 181 similarly examine how financial penalties, exclusion from services or professional consequences affect meaningful choice. This does not mean that any inducement or requirement is automatically coercive or unethical. Vaccination policies may lawfully set conditions in some settings, and the ethical case may be stronger when a worker's role creates a predictable risk to vulnerable patients. But any restriction should have a clear legal basis, an evidence-based objective, exceptions for recognised clinical contraindications, fair procedures, reasonable accommodation where possible, and periodic review. A mandate that cannot be explained publicly or adjusted when its premises change is especially vulnerable to loss of legitimacy. Trust is an outcome of the process Trust cannot be demanded as an act of obedience. It is built when authorities disclose what is known and unknown, correct mistakes, listen to affected groups and apply consistent standards. When early statements prove too absolute, correction is more ethical than defending the original wording for fear of appearing uncertain. Scientific knowledge changes; transparency about that process is a sign of integrity, not necessarily weakness. The research lesson is to keep the categories clear. A person must not be enrolled in a trial through deception or threat. A population policy may impose certain limits only if the relevant legal and ethical tests are met. A vaccine's benefits and risks should be evaluated from the best available evidence rather than political identity. Respect for autonomy is not a veto on all collective action, but collective action must be justified to the people who bear its burdens.
Nuremberg Revisited 182 Chapter 30. Global Inequality in the Pandemic Response Who Participated, Who Benefited, and who waited A global emergency with unequal starting positions A pathogen crosses borders, but the capacities to respond do not begin equally. Wealthier states could finance research, reserve vaccine supplies through advance purchase agreements, support manufacturing and use large procurement systems. Many lower- income countries faced more limited fiscal space, fragile health infrastructure, competing health burdens and delayed access to vaccines, diagnostics and therapeutics. These differences were visible well beyond vaccine allocation: oxygen, protective equipment, laboratory capacity, genomic surveillance, cold-chain logistics and trained staff were also unevenly distributed. Inequality is not itself proof that a specific trial was unethical. Nor does every delay in access mean that research participants were exploited. Ethical analysis must investigate the relevant decisions: how trials were placed, whether local health priorities informed the research agenda, whether participants were protected, whether local investigators shared leadership and intellectual credit, and whether the population could reasonably expect benefits from the knowledge it helped generate.
Nuremberg Revisited 183 The Ethics of where research is conducted Multicountry trials can be scientifically valuable because disease incidence, participant characteristics, and health systems differ across locations. They can improve the generalisability of results, enrol diverse populations and build research capacity. But they may also exploit a gap in bargaining power if sponsors choose locations primarily for speed or cost while the host population has little influence over the research question or little chance of accessing the eventual product. A defensible partnership should begin with a question relevant to the health needs of the host population or provide a compelling reason why the research is important there. It should invest in local scientific leadership, trained staff, ethics review, laboratory and data capacity, and community engagement. Agreements should clarify authorship, data access, ownership of inventions, use of stored samples, publication rights, safety reporting and the handling of incidental findings. Capacity-building should be measurable rather than described only as an aspirational benefit. Participants are not merely a resource that a foreign sponsor may use in exchange for reimbursement. Nor should research in lower- income settings be rejected simply because a sponsor is international. The relevant question is whether the collaboration is fair, scientifically sound, respectful of local authority and designed to leave the host system better able to protect and benefit its population. COVAX and the limits of allocation without control of supply The Access to COVID-19 Tools Accelerator and its vaccine pillar, COVAX, were created to accelerate development and expand equitable access to tests, treatments and vaccines. WHO's 2022 lessons-learned review of COVAX described an unprecedented global rollout at scale while identifying significant challenges to equitable access and lessons for future pandemic preparedness [95, 96, 97]. The programme should be assessed in both respects:
Nuremberg Revisited 184 it helped organise a multilateral response, yet its ability to secure timely supplies depended on manufacturing capacity, financing, export decisions and the purchasing power of states. A global allocation mechanism cannot distribute doses that do not exist or that manufacturers have already committed elsewhere. High-income states' advance purchasing, export restrictions, production bottlenecks and differing regulatory decisions affected the pace and timing of supply. Donations could fill some gaps but often arrived with limited shelf life or short notice, making deployment difficult. These are not identical failures and should be attributed carefully, but together they show why fairness requires more than a formula for allocating a fixed stock. Equity must be assessed against need and practical ability to deliver, not only the number of doses delivered to each country. A shipment without syringes, trained vaccinators, reliable cold storage, public communication, paid time off or systems for adverse-event surveillance may not produce the intended benefit. Countries and communities also need diagnostic capacity, access to treatment and the ability to generate and interpret their own data. Fair distribution is a system-wide task. Intellectual property, technology transfer and manufacturing capacity The pandemic revived disputes about intellectual property, voluntary licensing, technology transfer and temporary waivers of selected trade rules. Supporters of broader access argued that existing legal and commercial arrangements could prolong scarcity and that public financing created a strong claim for wider production and affordable supply. Opponents or more cautious commentators pointed to the complexity of biological manufacturing, the need for quality control, know-how, regulatory capacity and the risk that a patent waiver alone would not produce safe, effective vaccines at scale. The ethical debate should not be simplified into a choice between patents and public health. Intellectual property can support
Nuremberg Revisited 185 investment and innovation, but exclusive control over a publicly important technology may also impede fair access if no effective mechanisms exist for licensing, technology transfer and affordable supply. Conversely, formally removing a patent without sharing process knowledge, materials, technical assistance or regulatory expertise may have limited immediate effect. Future agreements should plan for manufacturing diversity, voluntary or compulsory licensing pathways where appropriate, transparent public- financing conditions, and regional capacity before the emergency peaks. Benefit-sharing is also an issue of governance. If public funds underwrite early scientific work or absorb commercial risk, agreements should consider how the public will benefit through access, affordability, supply commitments, independent evaluation, data transparency and meaningful accountability. Public funding does not automatically entitle every citizen to a particular product free of charge, but it creates a legitimate question about the obligations attached to that funding. India as manufacturer, trial host, and decision- maker India's role was complex. It hosted clinical research, developed and manufactured vaccines, supplied doses domestically, and participated in international supply efforts. The country had major manufacturing capacity, yet domestic needs, successive waves of disease, supply constraints, and policy choices shaped what could be exported and when. This is a reminder that a country may be both a producer and a setting where access remains uneven. National manufacturing capacity does not eliminate internal differences in access, trust, health literacy, or the ability to obtain medical advice. The ethical requirements for India and other vaccine-producing countries are therefore twofold. First, research and regulation must protect participants and provide reliable, transparent evidence. Second, public-health distribution should identify high-
Nuremberg Revisited 186 risk groups, reduce avoidable barriers, and explain the rationale for prioritisation. Research governance and deployment governance are distinct, but they meet in the question of whether evidence leads to a fair health benefit. Vaccine nationalism and the moral limits of priority Governments have a special obligation to protect their own populations, and no global allocation framework can simply erase domestic political responsibility. The ethical problem begins when national priority becomes unlimited hoarding or when contracts are concluded without regard for catastrophic scarcity elsewhere. A defensible strategy must balance domestic protection with global risk reduction and obligations to people whose exposure or vulnerability is high but whose countries have little purchasing power. The language of solidarity is meaningful only when it is translated into mechanisms: predictable financing; diversified and geographically distributed manufacturing; transparent allocation criteria; regulatory cooperation; pooled procurement where advantageous; fair licensing and technology transfer; reliable delivery; and systems for monitoring safety and effectiveness across diverse populations. Philanthropic and multilateral initiatives can contribute, but accountability must remain visible. Partnerships should publish objectives, conflicts of interest, governance arrangements, financing commitments and assessments of what was delivered. A fair-benefit framework for future pandemics Future clinical-trial and product-development agreements should be negotiated before a crisis becomes a scramble. At a minimum, investigators, sponsors, governments, and host institutions should document the following: 1. Local relevance: the health question, intended use and population benefit, including why the study is being conducted in the host setting.
Nuremberg Revisited 187 2. Local leadership: meaningful roles in protocol design, analysis, authorship, governance and regulatory dialogue. 3. Participant protections: independent ethics review, understandable consent, clinical care for research-related injury and fair access to relevant findings. 4. Capacity transfer: training, laboratory capability, data systems, pharmacovigilance and durable research infrastructure where feasible. 5. Access commitments: realistic arrangements for affordability, supply and post-trial access to interventions shown to be beneficial, consistent with applicable ethics and law. 6. Data governance: clear arrangements for privacy, sharing, storage, secondary use, authorship and access by local researchers. 7. Public accountability: transparent terms, conflict-of-interest declarations, timely trial registration and public reporting of all results. The COVID-19 response showed scientific cooperation on an extraordinary scale and also revealed how much global solidarity depends on supply, finance, and political willingness. A country that contributes participants, researchers, data, or manufacturing capacity should not be left with only the risks and costs while benefits are determined elsewhere. At the same time, ethical judgement should focus on evidence about particular arrangements rather than assuming that every international partnership is extractive or every multilateral programme is equitable. Conclusion: Knowledge must travel with accountability A pandemic transforms research into a global public good only when reliable knowledge is generated and its benefits can reach the people who need it. The moral claim is not that every country must receive identical quantities at every moment, irrespective of need or feasibility. It is that allocation should be justified by transparent and relevant principles, that avoidable inequalities
Nuremberg Revisited 188 should be addressed, and that the people who enable research should not be forgotten when the product reaches the market or the data are published. COVID-19 should therefore be remembered neither as a complete triumph of science nor as proof that global research is inherently exploitative. It was a stress test of institutions. Some systems generated useful evidence rapidly; other processes failed to provide adequate transparency or fair access. The enduring lesson is that emergency innovation and justice must be designed together. When accountability is postponed until after the emergency, the promise of equal concern is most likely to be broken precisely when it matters most.
Nuremberg Revisited 189 Table. Five ethical questions that must remain distinct A working guide for the COVID-19 chapters Question Primary ethical test Evidence that should be examined Was the research ethical? Valid design, independent review, voluntary informed consent, fair selection and risk minimisation. Protocol versions, ethics committee decisions, consent materials, monitoring records, trial registry. Was a treatment effective? Reliable comparative evidence with clinically meaningful outcomes. Randomised trial results, confidence intervals, adverse events, systematic reviews and applicability. Was an intervention authorisation justified? The evidence and legal standard required for that regulatory pathway at that time. Regulatory decision, data package, review rationale, conditions and post-authorisation commitments. Was a mandate proportionate? Necessity, legality, benefit to the stated objective, fair burdens and least-restrictive means. Epidemiology, intervention effectiveness for the relevant outcome, access data, legal rules and alternatives. Was the response equitable? Fair distribution of risks, benefits, access, voice and long-term capacity. Procurement and allocation records, trial locations, access agreements, delivery data and manufacturing arrangements.
Nuremberg Revisited 190 PART VII THE FUTURE: ETHICS, TECHNOLOGY, AND ACCOUNTABILITY Chapters 31–33
Nuremberg Revisited 191 CHAPTER 31 Genomics, Artificial Intelligence, Biobanks, and Digital Surveillance From the Sampled Body to the datafied person The New Research Subject: A Body, a dataset and a prediction The history of research ethics began with the body that could be examined, sampled, injected, observed, or operated upon. Contemporary research often adds a second subject: the record assembled from that body. A blood sample becomes a genome sequence; the sequence is linked to clinical notes; the notes are combined with prescriptions, images, location histories or information collected from phones; and a statistical or artificial- intelligence model uses these data to estimate risk, classify people or predict future outcomes. The person may never again meet the research team, yet the research may continue to affect the person and their relatives for years. This change expands the meaning of consent. A participant may understand why blood is taken today but may not anticipate every future use of a sequence that can be analysed with technologies not yet developed. A patient may agree to care without agreeing to the use of clinical records to train a commercial algorithm. A person may volunteer for a study of a particular disease without expecting the sample to be used later for ancestry inference, behavioural prediction or research into another condition. It is
Nuremberg Revisited 192 not enough to ask whether a signature exists. The question is whether the permission is sufficiently informed, appropriately scoped and supported by governance that can respond to future uses. UNESCO’s International Declaration on Human Genetic Data emphasises prior, free, informed and express consent, protection against discrimination and stigmatisation, and appropriate sharing of benefits from research [99]. Its importance lies in treating genetic information not simply as another laboratory result, but as information with implications for individuals, relatives and sometimes communities. The Declaration is an international normative instrument, not a substitute for domestic law, but it offers a durable ethical framework. Genomic privacy: why removing names may not be enough A genome is highly identifying and relational. It contains information about the person sampled, but also about biological relatives who may never have consented to be part of the study. Removing a name and replacing it with a code reduces direct identification; it does not guarantee anonymity. In a widely discussed 2013 study, Gymrek and colleagues demonstrated that individuals could be identified by combining information from publicly accessible genealogical and genetic resources with other data [100]. The result did not mean that every genomic dataset can always be re-identified. It did demonstrate that the risk depends on what other information is available and can change as databases grow. A responsible consent process should therefore avoid promises of absolute confidentiality when absolute confidentiality cannot be guaranteed. Participants should be told what data will be collected, how identifiers will be separated, which institutions or repositories may receive data, whether commercial access is possible, and what limits exist on withdrawal once data have been distributed or used in completed analyses. Researchers should
Nuremberg Revisited 193 describe realistic risks in plain language, not bury them in a technical paragraph. The US National Institutes of Health genomic data-sharing framework provides one example of how broad scientific sharing can be paired with limits based on consent, controlled-access repositories, institutional certification, and data-use agreements [101]. Such a model is not universally applicable; domestic law, community commitments and the conditions of original collection differ. The general lesson is that responsible data sharing is neither unrestricted publication of every record nor a blanket refusal to share. It is governed access linked to the permission under which the data were obtained. Broad consent, dynamic consent, and future use Biobanks cannot ask a participant to anticipate every future project. Broad consent is one answer: the participant agrees that samples and associated information may be used for a defined range of future research, subject to independent oversight and stated limits. Broad consent is not a licence for any conceivable activity. Its ethical acceptability rests on trustworthy governance, clear boundaries, transparent access decisions and a practical way to raise concerns. A participant should be able to understand whether commercial research, international sharing, family- related findings or linkage to other datasets may occur. Dynamic consent uses digital or other continuing communication to give participants information about new uses and, where feasible, to obtain further choices. It can strengthen control, but it is not a universal solution. Frequent requests can overwhelm people, digital exclusion can reproduce inequality, and a clicking interface may create the illusion of meaningful consent while people still do not understand the consequences. The best approach is context-dependent: layered information, accessible summaries, community engagement and genuine options may be more important than the technology used to administer consent.
Nuremberg Revisited 194 Researchers must also decide how to handle incidental or secondary findings. A genomic analysis may reveal a risk of a condition unrelated to the original research question, information about family relationships, or a result that is scientifically uncertain. Returning every possible finding can cause harm; withholding a highly actionable finding may also be troubling. Protocols should define in advance what categories will be assessed, who will confirm analytical validity, whether clinical confirmation is required, how participants express preferences, and which findings will not be returned. Research results should not be represented as diagnoses unless they have been appropriately validated for clinical use. Community interests and the right not to be stigmatised Genetic information can be used to support diagnosis and better treatment, but population-level findings can also reinforce stereotypes. A finding about a particular allele in a small population may be misrepresented as a claim about all members of an ethnic, caste, tribal or geographic group. Researchers should distinguish biological ancestry from social identity and avoid treating broad social categories as genetically uniform. Sampling designs, data labels and public communication can themselves create harm if they make a group appear diseased, inferior or responsible for a social problem. The ethical relationship is not exhausted by individual consent when research has foreseeable consequences for an identifiable community. Community engagement can help researchers interpret local history, explain how findings might be used, plan benefit-sharing and decide how results should be communicated. Community agreement should not replace individual consent or grant leaders an unlimited right to veto every individual’s participation. The aim is layered accountability: respect for each participant alongside serious engagement with communities that may bear collective consequences.
Nuremberg Revisited 195 Artificial intelligence: bias can enter before the algorithm begins AI systems learn from choices made throughout the research process: which patients entered the dataset, which variables were recorded, how disease was defined, which outcomes were measured, and which observations were missing. If rural patients, older adults, people with disabilities, or under-resourced hospitals are under-represented, an apparently accurate model may perform less well for precisely those groups. A high overall accuracy score can conceal clinically important errors in a minority subgroup. Bias is not restricted to intentional discrimination. A dataset may reflect unequal access to care; a model trained to predict historical spending may learn to underestimate need among people whose care was historically restricted. A model may perform well in one hospital but fail elsewhere because equipment, patient mix, prevalence, or clinical practice differs. These are questions of external validity and justice, not merely engineering details. WHO’s 2021 guidance on AI for health identifies human autonomy, well-being and safety, transparency, accountability, inclusiveness, and responsiveness as core governance concerns [102]. Its 2024 guidance on large multimodal models warns that systems capable of processing different types of input and generating varied outputs may be useful but can also produce inaccurate or biased information and be difficult to evaluate [103]. The Indian Council of Medical Research’s 2023 ethical guidance for AI in biomedical research and healthcare is especially relevant for Indian institutions because it addresses ethical review, consent, governance, and stakeholder responsibilities [104]. These guidance documents should be read together with applicable law and professional standards. AI-assisted research must retain accountable human decision- making. A model should not be treated as a neutral authority because it is mathematically sophisticated. Investigators should document the intended use, training data, inclusion and exclusion
Nuremberg Revisited 196 criteria, performance across relevant subgroups, known failure modes, version changes and procedures for monitoring after deployment. High-impact systems need independent validation and the ability to be suspended when performance deteriorates. Where an AI system changes during a study, the protocol should specify whether new validation or renewed review is required. Transparency and informed consent in AI-enabled studies An AI study may use data to build a model, validate it, compare clinicians with algorithms, or test a model that directly influences care. These are distinct activities with different risks. Consent materials should explain whether participants’ data will be used to train future models, whether a commercial company will obtain access, whether outputs may affect clinical decisions, and what will happen if the system is wrong. A statement that data are used “for research” may be too vague where future use is broad or consequential. Explainability is important, but it is not a magic requirement that every complex model provide a simple causal story. In some systems, explanations are approximate. Ethical review should ask whether clinicians and participants receive information adequate for the decision at hand, whether uncertainty and limitations are communicated, and whether a human can challenge a consequential output. A model that is explainable but poorly validated is unsafe; a model that performs well on average but cannot be monitored or challenged may still be ethically unacceptable for a high-stakes use. Digital surveillance: public benefit without permanent exceptionalism Digital technologies can support outbreak detection, contact tracing, health-service planning, and research. They can also expose sensitive relationships, create records of movement, enable discrimination, or be repurposed for policing, immigration
Nuremberg Revisited 197 control, employment decisions, or commercial profiling. The ethical problem is often not whether data could prove useful, but whether the means of collection are necessary and proportionate, whether less intrusive alternatives exist, and whether uses remain within the stated purpose. Emergency surveillance needs defined scope, access controls, retention limits, independent audit, public explanation, and a meaningful end point. Temporary arrangements should not silently become permanent infrastructure. Data minimisation is a practical ethical principle: collect no more than the purpose requires, retain it no longer than necessary, and separate research datasets from operational decisions wherever feasible. Security safeguards matter, but security alone does not make an unnecessary collection legitimate. India’s Digital Personal Data Protection Act, 2023 and the Digital Personal Data Protection Rules, 2025 form an important part of the contemporary data-governance landscape [105, 106]. Because commencement and implementation provisions may be phased, researchers should verify which obligations apply at the relevant date and how the legal framework interacts with sector-specific requirements, institutional policies, and research-ethics guidelines. A legal basis for processing data is not by itself an ethical endorsement of every research question or secondary use. India’s commencement notification of 13 November 2025 staged the Act: specified institutional provisions took effect on publication; provisions concerning Consent Managers are scheduled for one year later (13 November 2026); and most core personal-data-processing provisions are scheduled for eighteen months after publication (13 May 2027). The Rules also have phased commencement. As of 15 October 2026, it would be inaccurate to describe all substantive duties as already operative; researchers should check the exact provision, the Rules’ commencement clauses, and any subsequent corrigendum before relying on a particular obligation [105, 106].
Nuremberg Revisited 198 A governance test for genomics and AI research Before approving a project involving genomic material, linked health records, AI or digital monitoring, an ethics committee should be able to answer the following questions in concrete terms: • Purpose: Is the question scientifically and socially valuable, and is the proposed data collection necessary to answer it? • Permission: Does the proposed use fit the information provided and the consent or other lawful basis under which the data were collected? • Privacy: What is the realistic re-identification risk, including risks to family members and communities? • Equity: Which groups are missing from the data, who may be misclassified, and who benefits from the resulting tool? • Control: Who can access, copy, combine, retain or commercialise the data, and under what enforceable terms? • Accountability: Who must investigate a breach or harmful output, notify affected people, correct the record and suspend the system? The objective is not to freeze genomic science or AI until all uncertainty disappears. It is to make governance proportionate to what the system can do, including what it may reveal or infer beyond the researcher’s immediate intention. When data persist longer than the original project, governance must persist too. Data sovereignty, reciprocity, and the risk of extraction Large-scale biological and health datasets may be collected in one country, analysed in another and converted into intellectual property in a third. This division of labour can be productive, but it may also leave the institutions and populations that supplied samples without meaningful influence over priorities or benefits. Ethical partnership requires agreements made before collection that identify who controls access, who can approve secondary use, how intellectual contributions will be recognised, how results will
Nuremberg Revisited 199 be returned, and how disputes will be resolved. Promises should be realistic and recorded; vague assurances of future benefit are not a substitute for enforceable responsibilities. Data sovereignty does not mean that every dataset must remain physically within the borders of the country where it originated, nor that all cross-border research is exploitative. It means that decisions about human data must respect applicable law, the consent under which data were collected, legitimate community expectations and the capacity of local institutions to participate in governance. Secure federated analysis, controlled-access repositories, joint oversight and local analytic capacity may sometimes reconcile useful international collaboration with stronger control. The appropriate arrangement depends on the data, the research purpose and the affected populations. The same principle applies to artificial intelligence. If a model is trained on data from a population but is neither validated in that population nor made available to its health system at a reasonable cost, the research relationship deserves scrutiny. Researchers should not assume that publication alone is a sufficient return for participation. Training, technology transfer, accessible summaries, local research leadership and plans for responsible deployment can provide more durable benefits. The governing question is not whether every participant receives a product, but whether the overall arrangement is fair, transparent and responsive to the people whose data made the work possible.
Nuremberg Revisited 200 CHAPTER 32 Human Challenge Studies, Gene Editing, and Emerging Technologies When Technical Possibility Outruns Social and ethical permission Research at the Edge of what can be done Emerging technologies repeatedly create a dangerous mismatch between technical possibility and justified use. A new procedure may become feasible before its risks are adequately understood, before oversight systems have clear jurisdiction, and before affected communities have had a meaningful opportunity to debate its purpose. The question is not only whether an experiment can be conducted, but whether the proposed knowledge is valuable enough, the risks sufficiently controlled, and the conditions of participation sufficiently fair. Human challenge studies and human genome editing illustrate two very different ways in which researchers deliberately cross a boundary. In a challenge study, volunteers are exposed under controlled conditions to an infectious agent or other challenge in order to answer a scientific question. In genome editing, a researcher changes genetic material in cells; depending on the target and method, the change may affect only the treated person or may become heritable. In both cases, ethics cannot be reduced to a signature. It must address scientific value, risk, alternatives, monitoring, future consequences, and the distribution of benefits.
