An experimentally tested electronic prescription architecture returns to a practical question for pharmacies: how can they confirm that a prescription is still valid and has not already been used? Researchers at the University of Science and Technology of Oran Mohamed Boudiaf and the CEDRIC Laboratory at the Conservatoire National des Arts et Métiers have developed a decentralized approach that covers both credential verification and dispensing records. The approach connects security checks with the point at which medication changes hands.
Taking Another Look at the Dispensing Decision
Electronic prescribing already helps clinical services coordinate care and keep track of medication use. Yet the research team identifies a weakness that remains in conventional centralized public key infrastructure. The credentials used to establish trust between professionals depend on a central authority.
A valid digital signature answers an essential question about who authorized a document. Safe dispensing requires additional answers. Is that practitioner’s credential still active? Has another pharmacy already supplied the medicine? Can the dispensing system obtain current information without delaying the patient?
Those questions explain why this development deserves attention now. The prototype brings credential status, document integrity, and prescription consumption into one experimentally evaluated workflow. It moves the discussion beyond digitizing the prescription toward maintaining a shared, usable account of its status.
For healthcare systems evaluating new infrastructure, that changes the assessment. Administrative convenience remains useful, but it cannot stand in for evidence that separate organizations can make consistent authorization decisions.
Separating Clinical Records From Verification Records
The researchers’ design gives prescribers and pharmacists separate permissioned blockchains, BCAD and BCAPH. Accredited institutional participants validate changes rather than allowing one certificate authority to control the process alone. Updates to professional credentials require approval from a two-thirds majority of validators.
Clinical documents sit outside those ledgers in functionally separate InterPlanetary File System storage nodes. The chains retain compact references, cryptographic identifiers, timestamps, and status information.
This separation has a practical purpose. A more complex prescription should not require a correspondingly larger blockchain entry. The signed clinical content stays in the storage layer; the ledger establishes which document and status are authoritative.
At the pharmacy, a secure application reads the patient’s quick response code and checks its validity period. It then retrieves the prescription and verifies the prescriber’s signature and credential. After dispensing, it records consumption on BCAPH. A relay mechanism then communicates that change back to the prescribing domain.
For clinical services, the useful innovation is the connection between authorization and fulfillment. A prescriber can receive confirmation that the prescription has been consumed, while the dispensing record is intended to prevent another supply against the same authorization.
Local Verification Depends on Timely Synchronization
The team uses a Lock and Mint protocol to move verified status assertions between the two professional domains through an intermediate blockchain module, BCCOM. Despite terminology borrowed from asset bridges, this mechanism carries healthcare state information rather than financial tokens.
Its operational advantage is that a pharmacy can query a local ledger replica for the practitioner’s credential. That avoids waiting for a cross-domain request during each verification. The researchers also formally verified protections against repeated proof execution, reentrancy, and failed operations leaving incomplete changes.
However, local verification depends on how quickly the local copy becomes current.
In the attack evaluation, the team reported two missed certificate-forgery attempts during an 89.8-millisecond interval before relay commitment completed. That result makes synchronization delay a clinical workflow consideration, even when the interval appears brief. An authorization decision can occur while an update is still moving between domains.
Deployment assessments would therefore need to address:
- How a terminal handles credential changes that have not yet propagated.
- Whether dispensing can proceed when ledger verification is unavailable.
- How services resolve competing presentations of an unused prescription.
These are operating-policy questions as well as software questions. The answers govern what a pharmacist may do when the infrastructure cannot give a definitive answer.
Fast Prototype Results Do Not Establish Clinical Readiness
Across thirty complete workflows, the researchers measured average issuance time at 74.02 milliseconds and average consumption time at 70.46 milliseconds. Most of that processing time came from Ethereum block anchoring rather than cryptographic operations.
Those findings support the feasibility of the prototype’s execution path. They do not establish performance across a distributed healthcare network. The evaluation used one physical host, single-threaded PyEVM execution, and mocked storage held in memory. Its reported throughput of roughly 12 transactions per second must be read within those constraints.
Security results also require a differentiated reading. The system detected 89.0 percent of two hundred generated attacks, while all sixty legitimate control prescriptions proceeded without rejection. Reentrancy attempts were fully detected in the test set. Detection was lower for cloned codes and replay attempts, at 80.0 percent and 75.0 percent respectively.
The researchers linked missed cases to clones presented before first use and degraded offline operation in which ledger checks were bypassed, followed by reconciliation after reconnection. A secure code therefore does not, by itself, resolve every dispensing risk. The verification procedure remains decisive.
The Next Evidence Should Come From Service Conditions
Healthcare technology teams can examine a concrete architecture in this work, rather than rely on a general promise of improved trust through decentralization. They can also see which decisions would need to be settled before clinical implementation.
The next useful evaluation would test distributed operation, realistic storage, and interruptions to connectivity alongside the prescription workflow. Clinical readiness would require evidence about what happens when records disagree or a terminal cannot verify current status.
Organizations considering this infrastructure should begin by assessing credential management alongside dispensing authorization. Procurement and pilot requirements should spell out how the pharmacy counter handles uncertain status. That includes when dispensing must wait for verification and how unresolved events reach the responsible clinical team.
