Computer scientists from the University of Science and Technology of Oran Mohamed Boudiaf and the CEDRIC Laboratory at the Conservatoire National des Arts et MĂ©tiers have developed and experimentally evaluated a decentralized cross-chain public key infrastructure architecture designed to address longstanding security, interoperability, and fraud vulnerabilities in electronic prescription management. The transition from physical paper scripts to electronic prescriptions has offered clear administrative advantages across clinical healthcare, including improved coordination between prescribing physicians and dispensing pharmacists, lower rates of medication errors, and streamlined traceability of patient therapies. However, centralized public key infrastructure frameworks, which represent the conventional approach to managing cryptographic credentials and securing digital health communications, introduce a critical single point of failure into the healthcare ecosystem. These centralized designs hinder cross-domain interoperability, fail to offer robust technical safeguards against prescription duplication or double-dispensing, and expose sensitive healthcare systems to medication errors, illicit drug diversion, and fraudulent reimbursement claims that collectively impose billions of dollars in annual economic damages across international healthcare networks.

To overcome these structural limitations without compromising operational efficiency, the research team engineered a multi-tiered architecture that integrates four complementary distributed technologies into an integrated operational framework: permissioned blockchains executing the Practical Byzantine Fault Tolerance consensus algorithm, three functionally segregated InterPlanetary File System storage nodes, a formally verified Lock and Mint cross-chain synchronization protocol operating across a dedicated Ethereum relay chain, and single-use optical quick response codes with constructive entropy guarantees. Unlike existing systems that attempt to record large clinical documents directly onto immutable ledgers or rely on single centralized databases, the newly proposed architecture establishes a strict separation between a lightweight control plane and an off-chain distributed storage plane. The control plane, maintained across the permissioned ledgers, anchors only fixed-size content identifiers, pseudonymous cryptographic digests, state flags, and transaction timestamps. Meanwhile, the actual clinical payloads, signed digital records, and cross-chain audit proofs reside across dedicated storage nodes, ensuring that ledger expansion remains strictly decoupled from the volume or clinical complexity of individual medical prescriptions.

The operational framework divides the healthcare environment into two distinct professional domains, dedicating an independent permissioned blockchain to medical practitioners and another to dispensing pharmacists. The doctor domain operates on the BCAD blockchain, while the pharmacy domain runs on the parallel BCAPH ledger, with each network maintaining validator nodes that run Practical Byzantine Fault Tolerance consensus among accredited institutional participants such as medical licensing boards, pharmacist associations, healthcare institutions, and statutory regulatory authorities. Digital certificates governing healthcare practitioners follow an explicit five-state lifecycle spanning initial, active, suspended, revoked, and expired states, with consecutive certificate updates cryptographically bound together via prior content identifiers. Rather than placing trust in an individual central certificate authority, any modification to a practitioner certificate, including initial generation, periodic public key renewal, or administrative revocation, requires approval by a two-thirds supermajority of consensus validator nodes before becoming immutably committed to the ledger, thereby preventing unilateral authority compromise from destabilizing credential validation across the network.

When a physician issues an electronic prescription, the patient presents an identification credential such as the Chifa biometric card, allowing the physician to associate the patient identifier with prescribed drug regimens and dosages. The complete prescription file is digitally signed with the physician’s private RSA-2048 key, and the resulting file is transmitted to the dedicated physician storage node, IPFSD, which divides the document into blocks and generates a unique content identifier derived from the cryptographic SHA-256 digest of the data. This content identifier, alongside pseudonymous physician and patient hashes and an issuance timestamp, is anchored on the BCAD blockchain following consensus verification. The physician then generates a version-adaptive secure quick response code bearing error-correction level H, which encodes a structured uniform resource identifier containing the storage content identifier, the transaction block hash, hashed subject identifiers, a validity timestamp establishing a three-month operational window, and a 128-bit cryptographically secure pseudorandom nonce. This constructive entropy mechanism delivers an average Shannon entropy of 1,235.8 bits while preventing low-entropy plain identifiers from appearing verbatim in the optical payload.

At the pharmacy, dispensing workflow begins when the pharmacist scans the patient quick response code using a secure terminal application, initiating an automated sequence of local and ledger-level verification routines. The terminal inspects the embedded timestamp, automatically rejecting any expired prescription. If the validity window remains active, the application retrieves the signed prescription from the cross-chain synchronized storage node, IPFSCOM, using the embedded content identifier to verify content integrity. The pharmacist application extracts the physician digital signature and validates it against the practitioner public key retrieved from the local BCAPH ledger replica. Because practitioner certificate transitions are synchronized to BCAPH via the Lock and Mint protocol, the pharmacist verifies the active credential status through a single local query without inducing cross-chain network latency. Upon successful signature and credential verification, the pharmacist dispenses the medication and submits a consumption transaction to BCAPH containing prescription, practitioner, and patient identifiers, which is validated by consensus and immutably recorded to prevent subsequent re-dispensing.

Cross-domain synchronization is maintained by the dedicated Lock and Mint communication protocol embedded in the intermediate BCCOM blockchain module. While adapted from the nomenclature of blockchain asset bridges, this protocol transfers verified healthcare state assertions rather than economic tokens, eliminating the financial incentives and structural failure points characteristic of conventional bridges. Traditional bridges depend on external validator committees or continuous block header verification, but the proposed Lock and Mint architecture relies directly on cryptographic proofs signed by the same consensus validators that authorized the source transaction. When a practitioner credential changes state or a prescription is dispensed, the source domain ledger locks the event and submits a signed proof to the BCCOM smart contract. The contract verifies validator signatures over the proof stored in IPFSCOM, records the status transition, and coordinates relaying the update to the destination domain. A second synchronization cycle transmits dispensing events from BCAPH back to BCAD, ensuring that issuing physicians can monitor whether prescriptions have reached consumed status.

The researchers demonstrated that the Lock and Mint relay protocol formally guarantees three critical operational properties: reentrancy freedom, replay idempotency, and atomic rollback of failed operations. Reentrancy vulnerabilities, which have historically caused over 350 million dollars in decentralized finance exploits, are neutralised at the smart contract level through a non-blocking mutex, adherence to the Checks-Effects-Interactions programming pattern, and a dedicated processed-proof registry that immutably logs every executed proof digest to reject repeated calls prior to state mutations. To evaluate these guarantees empirically, the authors built an end-to-end prototype using Python 3.12, executing three independent Ethereum chains through the web3.py PyEVM backend alongside three functionally separated InterPlanetary File System storage nodes. The PBFT consensus layer operated with four validator nodes and one-fault tolerance under deterministic pseudorandom seed conditions, while cryptographic operations employed RSA-2048 with SHA-512 signatures and SHA-256 content identifiers aligned with National Institute of Standards and Technology guidelines.

In empirical performance benchmarks across thirty complete prescription issuance and consumption workflows, end-to-end issuance completed in an average of 74.02 milliseconds, while consumption finalized in an average of 70.46 milliseconds. An analysis of the processing pipeline revealed that Ethereum block anchoring represented approximately 89 percent of issuance latency, averaging 65.99 milliseconds, and 88 percent of consumption latency, averaging 62.06 milliseconds, due to PyEVM sequential execution. The underlying cryptographic primitives proved exceptionally fast on the critical execution path, with RSA-2048 signing averaging 1.27 milliseconds and verification requiring only 0.25 milliseconds. Storage performance remained similarly responsive, with document storage on IPFSD averaging 6.64 milliseconds and retrieval via IPFSPH averaging 8.15 milliseconds. Certificate lifecycle operations demonstrated total round-trip latencies of 105.2 milliseconds for physician issuance at 32,400 gas, 96.3 milliseconds for public key updates at 30,720 gas, and 72.6 milliseconds for credential revocation at 27,400 gas, achieving a 100 percent commit success rate across the evaluated lifecycle.

The security resilience of the architecture was evaluated against a rigorous suite of two hundred generated attack attempts spanning four distinct threat categories: reentrancy exploits, certificate forgery, quick response code cloning, and quick response code replay, alongside a control baseline of sixty legitimate prescriptions that all processed without false rejections. The system achieved a precision of 1.00 and an overall attack detection rate of 89.0 percent across the two hundred malicious attempts, yielding an aggregate F1 score of 0.942. Reentrancy attacks were completely neutralized with a 100 percent detection rate across fifty attempts. Certificate forgery scenarios achieved a 96.7 percent detection rate across sixty attempts, with two false negatives occurring exclusively within an 89.8-millisecond relay commit window before cross-chain propagation finalized. Quick response cloning and replay attacks exhibited detection rates of 80.0 percent and 75.0 percent respectively, with the remaining false negatives corresponding to pre-first-use clone presentation and degraded offline terminal operations where ledger verification was temporarily bypassed but subsequently reconciled upon reconnection.

While the experimental results validate the core cryptographic logic and functional viability of the decentralized public key architecture, the authors outlined key technical and operational limitations that define future research directions. Because the prototype operated on a single physical host using single-threaded PyEVM execution and an in-memory storage mock, the measured transaction throughput of roughly 12 transactions per second reflects local backend scheduling rather than distributed production capacity, and reported latencies do not capture wide-area network propagation delays. Furthermore, practical deployment across national healthcare systems requires formal compliance testing with HL7 Fast Healthcare Interoperability Resources data standards, rigorous legal verification under European Union General Data Protection Regulation and United States Health Insurance Portability and Accountability Act data privacy mandates, and large-scale multi-node clinical testbeds featuring live distributed InterPlanetary File System nodes, variable validator counts, and real-world clinical workflow integration.

Subject of Research: Decentralized cross-chain blockchain architecture for secure medical prescription management

Article Title: Blockchain-based PKI infrastructure for efficient management of medical prescriptions

Article References: Benkara Mostefa, A. F., Hidouri, A., & Bouzefrane, S. (2026). Blockchain-based PKI infrastructure for efficient management of medical prescriptions. Frontiers in Blockchain, 9, Article 1947022. https://doi.org/10.3389/fbloc.2026.1947022

Image Credits: AI Generated

DOI: 10.3389/fbloc.2026.1947022

Keywords: blockchain, public key infrastructure, electronic prescriptions, healthcare cybersecurity, smart contracts, IPFS, PBFT consensus, medical fraud prevention, Blockchain-based, infrastructure, efficient, management

News Source: Denise Maddox. (November 7, 2026). Decentralized Blockchain Architecture Protects Electronic Prescriptions Against Duplication and Forgery. Scienmag.