Privacy-Preserving Health Insurance Eligibility Verification on Blockchain Using zk-SNARKs

Verifying a patient’s eligibility for insurance reimbursement typically requires disclosing sensitive clinical data—diagnosis codes, laboratory values, and policy details—to third parties, conflicting with the data-minimization principle of modern data-protection regimes. Blockchain platforms improve auditability but, when health records are placed on-chain, amplify the privacy problem. This paper presents a zero-knowledge framework in which a patient proves that an authority-signed health record satisfies a set of policy predicates—a covered diagnosis, a laboratory value within an approved range, and a claim amount within the policy ceiling—without revealing the underlying values. Records are committed to an on-chain Merkle registry, and a record-bound nullifier prevents double claims without linking a claim to the patient. We state the NP relation proved by the circuit and reduce claim soundness, claim unlinkability, and front-running resistance to the knowledge soundness and zero-knowledge of the argument, the collision resistance of the hash, and the unforgeability of the signature scheme. A circom and snarkjs prototype with a Solidity verifier compiles to 10,050 R1CS constraints, of which the eligibility predicates contribute only 2.1%, and produces a constant 256-byte proof verified on-chain by a fixed four-pair check. On-chain cost scales with the number of public inputs rather than circuit size, and we identify the attack surface such designs commonly leave open—proof front-running, issuer-side linkability, and anonymity-set size, which the registry population bounds rather than tree depth.