Should patients be paid when research uses their data?

Follow one medical record on its working day. A hospital licenses de-identified patient data to an analytics company — revenue. The analytics company packages it into a product for a pharmaceutical sponsor — revenue. The EHR vendor that built the pipes takes its fees. The researchers drawing conclusions are salaried; the journal publishing the result charges for access.

One party in that chain contributed the thirty years of symptoms, surgeries, lab draws, and 3 a.m. flare notes that make the record worth anything. That party's compensation, in nearly every case, is zero.

Whether that should change is one of the sharpest open questions in research ethics. The argument for paying patients is easy to state. The arguments against are serious and deserve better than a shrug. Here is both sides, played straight.

The asymmetry is the argument

The case for payment starts with a simple observation: a market in patient data already exists — the only question is whether patients are in it. De-identified health data is bought, sold, and licensed at scale, an economy we map in how health data gets bought and sold. Nobody in that economy works for free except the data's subject.

For chronic-disease patients the asymmetry stings twice, because their contribution is not passive exhaust. Years of consistent symptom tracking, organized records, and community knowledge are work — work that makes the difference between a usable longitudinal dataset and noise. The history of research helping itself to people's biology without asking, from HeLa cells onward, is the backdrop here; we trace it in from Henrietta Lacks to data dividends.

The case against, taken seriously

Three objections carry real weight.

Undue inducement. The classic worry, embedded in research-ethics practice since the Belmont Report: payment large enough to cloud a person's judgment about risk. If $5,000 makes someone ignore a drug trial's dangers, the consent it produced is damaged. Ethics committees have policed payment for decades on exactly these grounds.

Commodification. Some goods change when priced. The worry — descended from Richard Titmuss's famous study of blood donation — is that paying for participation crowds out altruism and reframes the patient-researcher relationship as a transaction, with the poor effectively selling what the rich donate.

Consent quality and equity. Payment lands differently across incomes. A fee trivial to one person is rent to another, and a system that pays for data could systematically draw its cohorts from people with the fewest alternatives — recruiting need rather than consent.

None of these is silly. A platform that waves them off has not thought about them.

Where the objections do and don't bite

Now look closely at what each objection assumes.

Undue inducement is about risk — payment distorting the weighing of physical danger. Sharing already-collected data carries risks (privacy, chiefly), but they are different in kind from a Phase I dose escalation. The inducement concern scales with the hazard being discounted; for data use it argues for modest, proportionate payment, not for zero. Notably, mainstream US guidance treats compensation for participants' time and contribution as acceptable when reviewed and not coercive — the FDA has long said payment for participation is not inherently improper; it is a recruitment incentive to be reviewed, bounded, and disclosed.

Commodification cuts both ways. The transaction already happened — the data is already a commodity, traded daily between institutions. Declining to pay patients does not keep health data sacred; it keeps the proceeds concentrated. If anything corrodes trust, it is patients discovering that "donation" described only their own role in the deal.

Equity is the sharpest objection, and the honest answer is design, not denial: payments small enough that nobody shares data against their own judgment, never conditioned on staying enrolled, never gated on disclosing more than a study needs — and a consent that can be withdrawn without forfeiting anything already earned. Which is precisely why payment terms belong inside ethics review rather than in a terms-of-service footnote. What an institutional review board actually scrutinizes, and why that scrutiny is load-bearing, is the subject of what an IRB actually does. The Common Rule at 45 CFR 46 makes IRB review the place where recruitment and payment practices answer for themselves.

The middle path: participation fees, not purchase prices

Between "patients get nothing" and "data becomes a yard sale" there is a narrow, defensible lane, and it has four properties.

The payment is a participation fee, not a purchase. Nothing is sold; no title transfers. The patient is compensated for contribution and ongoing participation, the way research has compensated participants for a century — while keeping every right they started with, including the right to walk away, which is only meaningful if revocation actually works the way we describe in what revocable consent really means.

The amount is fair-market and modest — calibrated to contribution, not to what desperation would accept.

The trigger is an IRB-approved research-use event. Payment posts when an approved study actually uses the data — not as an inducement to join, not as a bounty for recruitment, but as a settled fact following a reviewed use.

And the whole arrangement is auditable: who was paid, for which use, under which approval, on a record that cannot be quietly edited.

How Mithras builds it

This lane is the one Mithras Health — a patient-mediated research-data platform for chronic disease, currently pre-launch — was designed around. Patients own their records and grant each study a separate, scope-limited, revocable consent. When an IRB-approved research-use event occurs, a fair-market participation fee posts to the patient. No event, no fee; no approval, no event.

The structural complement matters as much as the payment: Mithras never buys or sells data. Revenue comes from research access, with patients paid per use — so the platform's incentive is studies patients willingly join, not data hoarded or flipped. The reasoning behind that line is laid out in why Mithras Health will never buy or sell your data.

The strongest version of the opposing view says research is a commons, and commerce will poison it. But the commons argument has been losing on the facts for twenty years — the commerce arrived long ago, and it simply routed around the patient. The real choice is not between a gift economy and a market. It is between a market that pays everyone except the person in the record, and one with a seat at the table for them. Put that way, the question rather answers itself.