The Trial You Can't Run at Home: Ethics Dumping, Explained

The situation

This week, an ethics committee at the University of Southern Denmark approved a trial that will enroll 14,000 newborns in Guinea-Bissau. Half will receive the hepatitis B vaccine at birth. The other half will have it withheld for six weeks — long enough to test a hypothesis, long enough for some of them to catch a virus that can cause lifelong liver damage and cancer. The researchers are Danish. Part of the funding comes from the US Centers for Disease Control. The trial has been publicly praised by Robert F. Kennedy Jr., the US health secretary and a longtime vaccine opponent. Four members of the Danish ethics committee have resigned in protest.

The obvious question: why is a study no one would run in Denmark being run in West Africa? The answer is an old pattern with a name.

The model: ethics dumping

Ethics dumping is the practice of carrying out research in a lower-income country that would not be permitted in the researcher's own country — exporting the study to where the rules, the scrutiny, or the bargaining power of the subjects are weaker.

The term was coined around 2015 by a European Commission–funded project (TRUST), which went on to write a formal Global Code of Conduct for Research in Resource-Poor Settings. The project didn't invent the concept; it named a documented recurring harm. The landmark case is Pfizer's 1996 trial of the antibiotic Trovan during a meningitis outbreak in Kano, Nigeria: testing on children with contested consent, in conditions no US review board would have approved, followed by years of litigation. Ethics dumping is what you call the pattern Kano exemplifies.

The core mechanism is a mismatch of power. Research needs two things from a population: bodies and permission. Where a population is poor, where health systems are underfunded, and where a foreign institution is deeply embedded in the local system — as the Danish-run Bandim Health Project has been in Guinea-Bissau for decades — permission becomes easier to secure and harder to refuse. The science stays identical. Only the location of the risk moves.

The mechanism, step by step

A standard of care exists. The World Health Organization recommends a birth dose of the hepatitis B vaccine; decades of evidence show it works. In Denmark, every pregnant person is tested and infected mothers' babies are vaccinated.
A trial needs an unvaccinated group. To study whether the vaccine has effects beyond preventing hepatitis B — the speculative claim driving this study — you need babies who don't get it. In a country that already vaccinates at birth, withholding the shot would be illegal and unethical.
So you need a gap. Guinea-Bissau planned universal newborn vaccination for 2027, then pushed it to 2028. That delay is the opening. The trial runs in the window where "no vaccine" is still, technically, the local default.
You need an ethics committee that will say yes. The Danish national committee that oversees medical research had no jurisdiction, because the trial won't take place in Denmark. A university committee that had never reviewed a medical study changed its own rules to vote on this one — and its members with relevant expertise were outvoted.

Each step is individually defensible. Together they produce a result that, as one resigning reviewer put it, enrolls newborns "in a lottery," where the losers get hepatitis B "by design."

Why the usual safeguard failed

The Declaration of Helsinki — medicine's foundational ethics code since 1964 — requires that international research be approved by ethics committees in both the host and the sponsor country. The idea is redundancy: two independent gatekeepers, each able to say no.

Here, redundancy became a loophole. The sponsor-country review fell to a committee with no track record in medical ethics, operating outside the national system built for exactly this job. The host-country approval came from a government whose health minister had earlier said the trial was suspended after international uproar. When the expert gatekeepers are routed around rather than overruled, the two-signature rule stops being a safeguard and becomes a formality. This is the general failure mode of redundancy: two checks only help if they are genuinely independent and genuinely competent.

The other view, stated fairly

The researchers' case deserves its strongest form. A randomized controlled trial — where subjects are split by chance into a treatment group and a control group — is the most reliable way to isolate cause from coincidence; observational data can't do this cleanly. The team argues their revised protocol reduces harm: mothers are now tested for hepatitis B, and babies of positive mothers are vaccinated and excluded. In their framing, they are studying real safety questions in a population that genuinely has high hepatitis B rates.

The critics' rebuttal is concrete, not moral hand-waving. The hepatitis B test has roughly a 5% false-negative rate — one in twenty infected mothers wrongly cleared. The virus spreads casually: it survives on surfaces for up to a week and passes through shared toothbrushes, washcloths, nail clippers. In prior research, half of infected children caught it after birth, from the community — exactly the exposure a six-week delay leaves open. And the thing being tested is, in one Danish statistician's words, "highly speculative," while the thing being withheld is known to work. That asymmetry — large known benefit sacrificed to chase an uncertain hypothesis — is what distinguishes ethics dumping from a hard but fair trial.

Fact: the trial is approved; the design withholds a proven vaccine from newborns for six weeks; four committee members resigned; US and private money is behind it. Interpretation: whether the revised safeguards make the risk acceptable. Unknown: how many enrolled children will actually contract hepatitis B.

Where else this pattern lives

Ethics dumping is not confined to medicine, and recognizing its shape is the useful part. The general form: a risky activity migrates to the jurisdiction least able to refuse it. Electronic waste shipped to West African scrapyards rather than processed under EU rules. Garment work moved to factories exempt from the safety codes of the brands that commission it. Data harvesting routed through countries without privacy law. Carbon-intensive production relocated to states with no emissions caps. In each case the output is sold into the high-standard market while the hazard is borne by the low-standard one.

The tell is always the same mismatch: who decides, who benefits, and who bears the risk are three different populations. When those three line up, ethics works tolerably well. When they split — deciders in Denmark and Washington, beneficiaries in future science, risk-bearers in Guinea-Bissau's maternity wards — the ordinary brakes slip.

What to watch

Two signals will tell you which story is winning.

First, whether the trial proceeds to enrollment. The Africa CDC has opposed it, and "ongoing discussions" between the US and Guinea-Bissau are underway. A quiet pause, as happened earlier in 2026, would suggest the political cost is outrunning the funding.

Second, whether Denmark's national ethics system reasserts jurisdiction. If the national committee or regulators move to claim oversight that a university body took on, that is the Helsinki safeguard trying to repair itself.

Expectation: ~60% that the trial is either halted or materially redesigned before full enrollment, given the depth of institutional dissent — four resignations, Africa CDC opposition, a host government that already wavered once. The pressure points here are unusually public, and ethics dumping depends on not being watched.

Distilled from The Guardian World