From Henrietta Lacks to data dividends: research without asking

In 1951, a thirty-one-year-old mother of five named Henrietta Lacks went to Johns Hopkins with what turned out to be an aggressive cervical cancer. During her treatment, a sample of her tumor was taken and passed to a laboratory. The cells did something no human cells had reliably done before: they kept dividing, indefinitely. They became HeLa, the most widely used human cell line in the history of biology, present in polio-vaccine development, cancer research, and tens of thousands of papers.

Henrietta Lacks was never asked. She did not know the sample had been taken, her family was not told for more than twenty years, and neither she nor they shared in the value her cells generated. The science was extraordinary. The consent was nonexistent. Understanding the distance between those two facts is the work of modern research ethics — and it is not finished.

The pattern, not the exception

The HeLa story is usually told as a singular injustice. It is more useful to read it as the first clear instance of a recurring pattern: biological material or data collected in one context, then reused in others the person never agreed to, with the value flowing away from them.

That pattern does not require malice. The Hopkins researchers in 1951 were not villains; they operated inside a norm that simply did not treat a patient's permission as something owed. The harm came from the structure — a default in which a sample, once taken, belonged to whoever could make use of it. Rebecca Skloot's account of the Lacks family decades later made the human cost legible, but the structural lesson is the durable one: when consent is not built into the system, good science and grave wrongs can share the same petri dish.

Havasupai: the same default, a different decade

The Havasupai case shows the pattern surviving long after HeLa, and after the rules supposedly changed. In the early 1990s, members of the Havasupai Tribe in Arizona gave blood for a study of the diabetes devastating their community. They consented to diabetes research. That is not what their blood was used for.

Over the following years the samples were used for studies on schizophrenia, on inbreeding, and on ancestral migration patterns — including work suggesting origins that contradicted the tribe's own account of itself, a profound cultural injury. The donors had consented to one purpose; their material was quietly repurposed for several. The dispute settled in 2010, with the university paying the tribe and returning the samples. The lesson echoes HeLa's exactly: a sample given for a named purpose is not a blank check, and "broad" permission can paper over uses the person would have refused. We compare the broad-versus-specific approaches directly in broad consent vs. dynamic consent.

deCODE and the population scale

Iceland's deCODE genetics moved the same question to an entire nation. In the late 1990s, legislation authorized a centralized Health Sector Database intended to combine Icelanders' medical records, genealogy, and genetic data for research, with participation based on presumed consent — you were in unless you opted out.

The controversy was immediate and constitutional. Critics argued that presuming a whole population's agreement to commercial genetic research inverted the meaning of consent, and that opt-out placed the burden on citizens to defend data they never volunteered. Iceland's Supreme Court ultimately found the scheme's privacy protections inadequate, in a case brought by a teenager who wished to prevent her deceased father's records from being included. deCODE did valuable science. But it demonstrated that scaling up a consent shortcut does not dilute the problem — it multiplies it by the size of the population.

What each scandal forced

These cases were not absorbed quietly. Each one pushed the ethics infrastructure forward, and the rules we now take for granted are, in large part, scar tissue.

The foundational response in the United States is the Belmont Report of 1979, which named respect for persons, beneficence, and justice as the pillars of ethical research and entrenched the institutional review board — created by the National Research Act of 1974 — as the gatekeeper — the body whose actual job we describe in what an IRB actually does. For HeLa specifically, the National Institutes of Health reached an agreement with the Lacks family in 2013 governing access to the cells' genome data, with family members on the review committee — a belated acknowledgment, sixty years on, that the people behind the data had a say. The NIH treats that arrangement as a precedent, not a one-off.

The data-dividend answer

The data-dividend movement is the present-day extension of this arc, and it makes a sharper claim than "protect people." It argues that when personal data generates economic value, the people who produced it should share in that value — through direct payment, collective licensing, or other mechanisms — rather than have it extracted for free under the banner of consent they never meaningfully gave.

The idea has surfaced in mainstream politics, from a US governor floating a consumer "data dividend" to advocacy projects organizing people to collectively bargain over their data. The critiques are worth taking seriously — that pricing data could pressure the desperate, or commodify something that should stay protected — and we weigh them honestly in should patients be paid when research uses their data. But the underlying intuition is hard to dismiss after HeLa: if everyone in the chain is compensated except the person at its source, "for the good of science" starts to sound like a description of who gets to keep the proceeds. The modern data-brokerage economy, which we map in how health data gets bought and sold, is that imbalance running at industrial scale.

Mithras Health is, in a sense, an attempt to build the lesson of these cases into machinery so it cannot be forgotten. Consent is requested per study and is revocable, so no sample becomes a blank check the way the Havasupai blood did. The audit trail is append-only and permanent, so reuse cannot happen in silence the way HeLa's did. And patients are paid fair-market participation fees that post when research actually uses their data — the data-dividend principle, made concrete and tied to consent rather than offered as charity. The aim is narrow and unglamorous: to make "research without asking" not merely against the rules, but impossible to perform. Seventy-five years after a sample left a hospital in Baltimore without a word to the woman it came from, that still counts as progress worth engineering.