Consent: When It Arrived, and What It Replaced

The rule is older than the story

The account most people carry is that modern medical ethics began at Nuremberg: that the trial of the camp physicians produced the Code in 1947, and that research on human beings has been governed by consent ever since.

Both halves are wrong, and they are wrong in the same direction. The rule is older than 1947, and compliance is younger than 1947 by decades.

On 28 February 1931 the German Reich Minister of the Interior issued a circular setting out guidelines for new therapies and for scientific experiments on human beings. It required unambiguous consent. It took particular care over subjects under eighteen. It prohibited exploiting the poor and the socially disadvantaged. It was issued in response to a disaster — seventy-two infants had died in Lübeck after being given a tuberculosis vaccine contaminated at the laboratory — and it was binding law throughout the remaining existence of the German Reich.

It was never repealed. The experiments tried at Nuremberg were carried out while it was in force.

That is the shape this whole page keeps finding. The rule arrives, it is written clearly, and it does not reach the people doing the work.

What Nuremberg actually was

The Code was not legislation. It is ten points set out in the judgment of a military tribunal trying twenty-three individuals, the first of which says that the voluntary consent of the human subject is absolutely essential. Sixteen defendants were convicted and seven were executed.

A judgment binds the parties to it. The Code bound nobody else, and the timeline above shows what that meant in practice. The United States Public Health Service was running an inoculation study in Guatemala in August 1947, when the judgment came down. The study did not stop. The same service was running a study in Alabama that would continue for another twenty-five years.

The sequence of instruments that followed is familiar and looks like progress, which it partly is: the Declaration of Helsinki in 1964, the National Research Act in 1974, the Belmont Report in 1979, the Common Rule in 1991. What the timeline shows is that each of them arrived after something, not before it — and that the thing it arrived after was usually already covered by the previous instrument.

The single worst month in the record

If one fact belongs in this piece it is this one, and it comes from the files of the agency itself rather than from its critics.

In April 1947 the Atomic Energy Commission adopted a formal patient-consent policy for human experiments. In the same month it circulated a memorandum recommending that human experimentation not be made public, on the grounds that publicity might have an adverse effect on public opinion or lead to legal suits.

The rule and the reason for ignoring the rule were written by the same body in the same month. When the advisory committee examined the eighteen patients injected with plutonium between 1945 and 1947, it found that with a single exception the record suggests they were not told they were subjects of an experiment from which no medical benefit was expected. Consent is documented once.

This is why the page is built on the written standard rather than on the harm. The usual argument about these studies is conducted as though the people running them did not know better, and the documents mostly show that they did.

What Tuskegee was, and was not

The Alabama study is the best known entry here and the most often described loosely, in ways that make it both worse and less instructive than the record does.

What happened is bad enough to state plainly. 399 men with late-latent syphilis and 201 controls were followed for forty years, told they were being treated for 'bad blood' under a government health programme, and given meals, transport and burial insurance. No evidence exists that informed consent was obtained. Penicillin became the standard treatment in the late 1940s and was not offered. The study ended in 1972 after a reporter wrote about it, on the recommendation of an ad hoc advisory panel, and a settlement of $10 million followed in 1974.

What the study did not do is infect anyone. The men had syphilis when they were enrolled. The wrong was the withholding of treatment, information and the ability to leave. The deliberate-infection case is a different row on the timeline — Guatemala, run by the same service with some of the same personnel, where more than a thousand people were deliberately exposed and which nobody found until a historian went through a dead doctor's papers in 2010.

Conflating the two is common and it costs something. The Alabama study produced the American regulatory system; the Guatemala study produced an apology sixty-two years late and a commission report. They are different wrongs with different lessons, and the fact that the second was invisible for six decades is the part most relevant to a reader of this archive.

Consent is not one thing

The Willowbrook row is here because it resists the frame, and the section's house rule is that the row which resists is the one to put in front.

Children with intellectual disabilities at an overcrowded state institution were deliberately infected with hepatitis. Their parents signed consent forms. The institution was so hard to get into that enrolment in the study functioned as a route of admission, which is why critics have argued for sixty years that the consent was not freely given. The researchers' defence was that infection there was close to universal anyway, and the work contributed to the hepatitis vaccines that followed.

Every element of that is true at once. The Nuremberg Code's first principle requires the subject to have legal capacity to consent, which these subjects could not — and that is exactly why the question moved to their parents, where a different kind of pressure was available. A rule that locates consent in the right place can still be satisfied in form by people acting under constraint.

When the profession finally said this out loud it was not a regulator or a court that did it. Henry Beecher's 1966 paper in the New England Journal of Medicine set out twenty-two published studies, drawn from the literature rather than from leaks, in which consent had been absent or inadequate. Willowbrook was one of them. The studies were not secret. They had been published, peer-reviewed and read.

What this timeline cannot establish

It is not a measure of harm. The advisory committee on radiation research found no evidence that most subjects were injured by the exposures they received, and said so; its findings are about consent and candour. Reading the two as one is the error that turns a documented wrong into an exaggerated one, and the exaggeration is what lets the whole thing be dismissed.

It is also selected. These are the cases that surfaced, which means they are the ones where a reporter, a historian, a committee or a court got there. The research that was conducted properly leaves no comparable trace, and neither does the research that was not conducted properly and was never found — so nothing here supports an inference about how common either was.

And it does not tell you the rules are adequate now. What it establishes is narrower and harder to argue with: that in every row but the first, a written standard requiring consent already existed when the work was done, and in the sharpest case the agency wrote the standard and the reason for burying the work in the same month. The question worth carrying out of it is not whether the rules exist. It is who the rules reach.