The rare disease data market

Pharmaceutical companies developing rare disease therapies pay for natural history data. Regulators require real-world evidence to support post-market surveillance. Insurers need outcomes data to make coverage decisions. Medical food and assistive device manufacturers need usage data to inform product development. Academic researchers need cohort access for hypothesis-driven studies.

Each of these parties pays separately, builds separately, and rarely shares with the others. The result is duplicated cost, fragmented data, and an evidence base that is uneven across conditions and across questions. The pattern is so wasteful that it is hard to defend except as the inertia of the institutional structure that produced it.

A patient-controlled data trust changes the supply side of the equation. The data is held in one place under fiduciary governance. The buyers come to the trust for the access they need, on terms the affected community establishes. The trust does not sell data. The trust sells access to data the patient has consented to make available, under conditions that protect patient identity and respect contributor preferences.

What the buyers are actually buying

Pharmaceutical companies need natural history data for trial design and external control arms. The current cost runs into the tens of millions of dollars per condition for sponsor-funded studies. The trust's offering replaces that cost with an access fee that is much lower because the underlying data collection cost is amortized across multiple buyers and conditions.

Regulatory agencies need real-world evidence to evaluate post-market outcomes. The current source is sponsor-reported data with known biases. The trust's offering provides patient-attributed real-world data with provenance the agency can audit. Several FDA real-world evidence frameworks have been refined over the past five years to accept data of the kind the trust would produce. The infrastructure exists; the data does not.

Insurers and payers need outcomes data to make coverage decisions. The current source is clinical trial data plus a patchwork of registries and academic studies. The trust's offering provides longitudinal cohort data that supports the case-level and policy-level analyses described in coverage discussions.

Academic researchers need access to characterized patient cohorts and longitudinal data. The current cost is the time and grant funding required to build single-institution datasets, plus the intellectual property and access negotiations required to combine them. The trust's offering replaces that cost with a research-access fee structured to support academic budgets.

Manufacturers of medical foods, assistive devices, and consumer rare disease products need usage data to improve products. The current source is sales data and customer surveys. The trust's offering provides structured outcomes data that supports the comparative analyses the manufacturer's R&D team can act on.

What the trust does not sell

The trust does not sell raw data. The data stays under patient control. Buyers receive access under governance rules that include consent verification, reidentification prohibitions, audit logs, and use-restriction terms. The technical implementation can include differential privacy, secure multiparty computation, and federated analysis where appropriate. The contractual implementation includes access agreements, breach penalties, and ongoing compliance monitoring.

The trust does not sell to entities the contributing patients have not consented to share with. The granular consent model allows a contributor to permit research access without permitting commercial access, or to permit specific commercial uses without permitting others. The infrastructure has to support that granularity.

The trust does not sell to the insurer for the purpose of making coverage decisions about specific contributors without those contributors' consent. The data that informs aggregate coverage analysis is not the same data that determines individual coverage. The architectural separation matters for the contributor's interest in not being adversely selected against.

What sustains the model

The trust becomes self-sustaining when access fees and partnership revenue exceed the cost of data infrastructure operation. The unit economics are favorable when the data is broad enough to serve multiple buyers across multiple conditions and deep enough that each buyer extracts substantial analytic value.

The mature state of the model is comparable to the role of utilities in other infrastructure markets. The road network sustains itself through the economic activity it enables, not through tolls on individual trips. The data trust sustains itself through the research, development, and clinical improvement it enables across many conditions, not through transactional sales of patient data.

The contrast with the current state is the contrast between an infrastructure model and a transactional model. The current data marketplace, where it exists, is transactional: sponsors pay for studies they own, registries license data on terms that vary by registry, intellectual property attaches to specific datasets. The infrastructure model treats the dataset as a long-lived asset, owned by the affected community, that supports many uses over time.

What the patient gets

The patient who contributes data to the trust receives several things that the current data market does not deliver.

The patient retains custody of their data. The contributing patient can withdraw, access, audit, and govern the use of their contributions. The contemporary alternative, in which the patient signs a consent form for a specific study and the data flows out of their control thereafter, is replaced with a continuing relationship.

The patient receives a share of the value the data produces. The infrastructure can support direct compensation, foundation funding for advocacy organizations, support for natural history research the affected community wants pursued, or other forms of return-of-value. The specific mechanism is a governance choice. The principle that the value generated by the data flows in part back to the contributors is the architectural invariant.

The patient is part of the research that uses their data. The trust's governance gives contributors standing in decisions about access, use restrictions, and priority research questions. The historical pattern in which patients are subjects of research conducted by others, with no continuing role after enrollment, is replaced with a model in which patients are co-investigators with governance rights.

The infrastructure question is whether the trust can be built. The technical answer is yes; the components exist and have been demonstrated in adjacent contexts. The governance answer is the harder problem, and is the work that defines whether the model serves the affected community as it claims to.