Consent at National Scale

A national genomic programme cannot obtain specific consent for every future study. The studies do not exist yet. Some will be designed by researchers who are currently students, using methods that have not been invented. The alternative to broad consent is not better consent. It is no biobank.

So broad consent it is, which means the ethical work moves from the moment of signature to the governance that follows it. This is well-trodden ground in African research ethics, and a Nigerian programme starting now has the advantage of not needing to invent any of it.

What broad consent actually is

Broad consent is consent to the use of samples and data for unspecified future studies, subject to conditions. The conditions are the substance. They can restrict the categories of research permitted, the diseases studied, who may access the material, and the circumstances of transfer.

The H3Africa framework, developed through consultation with research ethics committees across the continent, concluded that broad consent is ethically appropriate in most African contexts when used responsibly, and identified where the responsibility sits: in explaining genomic research and biobanking, and their implications, properly during the consent process.

That word "properly" is where programmes fail. A form that a participant signs without understanding what a biobank is has produced a signature, not consent, and the deficiency surfaces years later when a use arises that the participant would not have chosen.

Community engagement is not outreach

In many African contexts, decisions about participation are taken in consultation with family and community rather than in isolation. Genomic research also produces findings that implicate families and populations, not only individuals. A variant frequency finding is a statement about a group.

The H3Africa framework therefore treats community engagement as a core component of governance rather than as communications activity: informing, consulting, and actively involving the communities with a legitimate interest in the research.

For a national programme this has an operational implication. Engagement is a staffed, budgeted, continuing function, not a launch event. It is also the function most likely to be cut when budgets tighten, and the one whose absence is least visible until trust fails.

Withdrawal has to be real

Broad consent is only defensible if withdrawal works. In practice that means:

The participant can find out what they consented to, years later, in language they can read. Withdrawal can be initiated without travelling to the original collection site. Withdrawal propagates to derived datasets and to external recipients who received data under access agreements, which requires those agreements to have contained propagation obligations from the start. And the limits of withdrawal are stated honestly at the outset, since results already published cannot be unpublished.

Most of that is engineering, and it is engineering that must be designed in before the first participant is enrolled. Retrofitting withdrawal propagation into a system that did not plan for it is close to impossible.

Consent conditions should be machine-readable

The gap between what a participant agreed to and what a data access committee remembers is a real operational risk at scale. With tens of thousands of participants recruited under multiple consent versions, no reviewer can hold the matrix in their head.

The Data Use Ontology encodes consent conditions in a machine-readable form that access systems can enforce directly. Combined with a data access committee working to published criteria, it makes the connection between what was promised and what is permitted auditable rather than institutional memory.

Benefit sharing is a governance question, not a gesture

Among the ethical concerns most consistently raised in the African genomics literature are ownership of samples and data, fairness in access, and the capacity asymmetry between scientists in high-income and low-income countries. Those are structural concerns and they need structural answers.

For a Nigerian programme the credible ones are concrete: that analysis capability is built domestically rather than outsourced, that Nigerian researchers have priority access to Nigerian data, that authorship and intellectual property arrangements are agreed before collaboration rather than after, and that capacity building is a funded programme obligation rather than an aspiration in a memorandum.

The failure mode to design against

Consent governance is not primarily protection against malice. It is protection against ordinary institutional failure: a company running out of money, an investor exercising rights the participants never knew existed, a partnership ending badly.

The recent history of Nigerian genomics contains a direct example. A biorepository holding samples and clinical data from roughly 130,000 West African participants ended up as a contested asset in a corporate dispute, with reporting describing genomic data from 100,000 Nigerians offered for sale at three million dollars, and a Federal High Court in Lagos issuing an injunction in 2025 to prevent the sale pending resolution. The company's holding entity was incorporated abroad while the biobank sat in Nigeria.

No consent form drafted in that programme contemplated that scenario, because consent forms do not contemplate insolvency. What would have contemplated it is governance: a structure specifying where samples sit if the operator fails, participant representation in the body making that decision, and an entity constituted under the law of the country whose citizens provided the material.

That is the standard a national programme should hold itself to. Not because failure is expected, but because consent obtained on any lesser basis is consent obtained on an incomplete description of the risk.