Disease donation · V2

Your donation, your disease,
your value.

Donate for research on YOUR disease. Not everything. Just what helps. And you can withdraw at any time.

The cybernetic loop

Patient → Research → Better AI → Patient.

Virtuous, transparent, opt-in loop. Your data improves the models that serve you.

Patient consent

You enable donation for one specific disease.

Disease research

Pseudonymized cohort made available to researchers.

Outcomes captured

Study results flow back into the research lake.

AI models retrained

CamemBERT-bio + Mistral fine-tuned to better serve you.

Improved recommendations

More precise suggestions for YOUR specific disease.

↻ Patient → Research → Better AI → Patient.

How it works

Four steps, all reversible.

Choose

You enable research for one specific disease (diabetes, hypertension, breast cancer, etc.) — not your full record.

Pseudonymization

Your data is k-anonymized (k=5 minimum) before entering the research lake. No re-identification possible.

Contribution

Researchers can query the cohort without ever seeing your identity. Audit-log on every access.

Revoke

1 tap removes your data from the lake within 24h. No justification required. No penalty.

Why it's different

Three models, only one respects you.

Three models, only one respects you.
 ConsentTransparencyHostingValue returned
GAFAM (Apple Health, Google Fit)All or nothingOpaqueHosted US · Cloud ActNo patient feedback
State platforms (Mon Espace Santé)CentralizedLimitedHosted FRNo patient feedback
My Data My CareDisease-by-disease opt-inTransparent · audit-logFR / EU · HDS v2Patient royalties on pharma studies
Patient royalties

Your value, recognized.

If a pharma study uses your pseudonymized data, you receive a share of the royalties.

5 to 15% of the contract

Target: 5 to 15% of the study contract redistributed to contributors, prorated to their actual contribution.

No marketing leverage

No upsell. No premium pricing. No pressure. Just your value, recognized.

Annual public audit

Royalties paid published annually. Independent audit firm. Auditable ledger.

Technical security

Three layers, zero leak.

E2EE patient vault

End-to-end encryption on the patient side. Keys held by the patient (HKDF). MDMC CANNOT read your raw data.

Research Lake isolated VPC

Lake hosted HDS v2, VPC isolated from the rest of the infra, researcher access via audited API. K-anonymous pseudonymization + direct identifier suppression (FFI L.1110-4).

Federated learning V3

Models trained directly at partner hospitals without PHI ever leaving. Only gradients (model parameters) circulate — not the data.

Understand the architecture

See how each technical layer protects your privacy while feeding the progress loop.

See the architectureTechnical sovereignty