The Evidence Chain Moves Faster Than the Molecule

by Priya NairClinical and Regulatory Correspondent, HMND

Novartis’s agreement with Abogen puts a familiar pressure point under a very expensive spotlight. The deal covers an exclusive worldwide license to ABO2203 and options on additional RNA programs, with Abogen receiving $575 million upfront and potentially about $7.2 billion in milestone payments. The scientific asset crosses a corporate boundary in a day. Its evidence may spend years trying to cross the same border.

A global pharma network adds facilities, laboratories, quality units, electronic records platforms, security controls, and regional regulatory expectations. Each part can work exactly as designed while the handoff between them quietly loses the plot. The licensing announcement rarely shows that part. The molecule gets the headline. The identifiers get the audit.

Picture a Tuesday morning in a tech transfer room. The partner’s development group has changed the lipid formulation after a scale up run showed a drift in encapsulation efficiency. The partner calls the change F 17. Novartis’s product lifecycle system calls the receiving version F 17A because its naming service requires a sponsor controlled suffix. The analytical laboratory reports an encapsulation result against sample code ABG 240611 03. The manufacturing execution system records the material as ABO2203 DS 240611, while the batch record was opened at 08:14 China time and signed at 09:14 Swiss time after conversion into coordinated universal time.

Everyone has a record.

The records disagree.

The assay result sits inside the partner’s laboratory information management system. Its method version is AM 2.4, with a calculation template updated two days earlier. The sponsor’s receiving laboratory has transferred the method into its own system as AM 2.4 N, changing the integration parameters and adding a system suitability check. The transfer report describes the method as equivalent. The raw chromatograms remain in the partner’s validated archive, where the sample identifier is linked through a spreadsheet that was uploaded after the run.

A batch review specialist notices that the reported encapsulation efficiency is 91.8 percent in the certificate of analysis and 90.9 percent in the laboratory worksheet. The difference comes from a rounding rule in the exported report. The formulation change appears in the development report as effective on 11 June. The electronic batch record shows the manufacturing instruction becoming effective on 12 June. A quality reviewer places the lot on hold while the teams ask whether this is a calculation discrepancy, an undocumented process change, or a batch made under a formulation that the sponsor never formally accepted.

The software has performed its assigned tasks. The evidence no longer forms a defensible sentence.

The partner owns the first bottleneck

Abogen owns the original process knowledge, source data, development history, and the explanation of what changed in the formulation. The partner also owns the first responsibility for making its records intelligible to an outside quality organization. A validated local system does not transfer meaning by osmosis.

The partner must preserve the link between the formulation version, the manufacturing instruction, the sample taken, the assay method, the raw data, the calculation, and the quality conclusion. That chain needs stable identifiers and an explicit time basis. A document that says “the process was unchanged apart from the lipid ratio” leaves too much work for the receiving team. Senior reviewers then reconstruct intent from filenames and meeting notes, which is an expensive form of archaeology.

The partner’s quality unit owns the original disposition decision. It must show who reviewed the result, which data were available at the time, whether the record was changed, and why the batch was accepted or rejected. Audit trails matter here because a final report cannot reveal whether a failed integration was repeated, deleted, or superseded after a conversation in a messaging application.

Operators know the emotional shape of this problem. Someone remembers the run. Someone else remembers the change control. A third person remembers that the spreadsheet was only meant to bridge the systems temporarily. Temporary bridges acquire surprisingly permanent foundations.

The sponsor owns the receiving bottleneck

Novartis owns the decision to treat partner evidence as sufficient for development, manufacturing, and regulatory use. That makes the sponsor responsible for more than reviewing a document package. It must qualify the interface through which the package becomes sponsor evidence.

Analytical method transfer sits at the center. The sponsor laboratory needs a protocol that defines the method’s intended use, critical parameters, system suitability, acceptance criteria, reference materials, sample handling, and treatment of atypical results. It also needs to compare the partner’s method performance with the receiving laboratory’s performance under known conditions. A successful transfer demonstrates that the receiving laboratory can produce fit for purpose results; it does not amount to a ceremonial signature on a report called Method Transfer Complete.

RNA products sharpen the problem because product quality depends on several linked attributes. Identity, integrity, purity, potency, encapsulation, particle characteristics, residuals, and microbiological controls may each be measured by different methods with different sensitivities. A small change in sample preparation or instrument integration can alter the apparent result without changing the material. The reverse also happens. A stable looking number can conceal a real process shift when the assay lacks the sensitivity or specificity needed for the question.

The sponsor’s analytical development group owns the scientific bridge. Quality owns the decision about whether the bridge is acceptable. Information technology owns the system controls that make the bridge observable. Those responsibilities overlap, which is precisely why they need names attached to them.

Scale up creates new evidence, not larger evidence

A process that behaves well in a partner’s development suite may change character when it enters a global manufacturing network. Mixing energy, hold times, equipment geometry, raw material lots, filtration capacity, freezing conditions, and shipping exposure can alter an RNA product or its delivery system. The manufacturing team owns the process performance evidence. The sponsor’s process development group owns the rationale for scale dependent parameters and the comparability plan.

Comparability should begin before the first commercial sized run. The team needs predefined tests for the formulation change, the new equipment, the new site, and the analytical method used to judge similarity. Statistical comparison helps, but statistics cannot repair an identifier that points to the wrong sample. A confidence interval around an orphaned result is still orphaned.

The quality unit owns the conclusion that the post change material remains suitable for its stage of development. Regulators will expect the conclusion to connect process knowledge, analytical results, and manufacturing history. They will also ask whether the methods could detect a meaningful difference. A thick report with thin traceability invites more questions than a shorter report with a clean chain.

Data integrity lives in the seams

Data integrity becomes visible when a reviewer follows one result backward. The result should lead to the calculation, the raw data, the instrument state, the method version, the sample identity, the analyst, the review, and the decision. Each link requires an attributable, legible, contemporaneous, original, accurate, complete, consistent, enduring, and available record.

The database does not establish that chain by itself. A laboratory information management system may preserve an audit trail while a file transfer process drops the relationship between the sample code and the manufacturing lot. An electronic document management system may control the approved report while the source spreadsheet remains editable in a shared folder. A manufacturing execution system may record the batch step while the quality system stores the deviation under a different product code.

Audit trails are useful when people review them as evidence. They become theatre when a procedure says they will be reviewed and no one can explain which events would change a release decision.

The partner and sponsor need a cross system identity map. The map should cover product, batch, formulation, sample, method, instrument, result, deviation, change control, and quality decision. It should record the source identifier, sponsor identifier, transformation rule, owner, effective time, and reason for any change. A controlled mapping table is not glamorous. Neither is a chain of custody. Both are more persuasive than a polished portal.

System validation has to include the handoff

Computerized system validation usually concentrates on the system boundary that a vendor can draw neatly. The regulatory question follows the evidence across the boundary.

The sponsor should validate the transfer workflow, not only the receiving application. That means testing file formats, metadata, time zones, identifier transformations, rejected records, retransmissions, duplicate samples, late corrections, audit trail preservation, access permissions, and reconciliation between source and destination counts. The test should include the failure scene described above. A formulation version should be changed, a result should be corrected through an approved process, a batch record should arrive late, and the quality system should still expose the relationship among them.

The infrastructure layer matters. Identity management must distinguish a partner analyst from a sponsor reviewer without turning the partner into an anonymous service account. The security boundary must permit necessary data exchange while preserving segregation of duties. Interface engines need monitored queues, error handling, and retained transaction logs. Time synchronization needs an agreed reference. Data lakes need schemas that preserve source context rather than flattening every record into a friendly but meaningless table.

User experience decides whether these controls survive contact with work. If an analyst must copy a twelve character sample code across three applications, someone will eventually transpose two characters. If the receiving reviewer cannot see the source method version beside the imported result, the reviewer will trust the report because the screen looks official. A convenient interface can reduce risk. A decorative dashboard can hide it.

The regulator meets the interface

Regulatory confidence depends on the interface because the product is judged through connected evidence. The individual laboratory system may be validated. The manufacturing system may be validated. The document system may be validated. A quality decision formed from mismatched records can still be unreliable.

The FDA’s process and product expectations, and the European GMP emphasis on data integrity and computerized systems, converge on a practical question. Can the sponsor demonstrate what happened, when it happened, who performed it, which data supported it, and why the decision followed? A global filing turns that question into a multi site exercise involving translations, local procedures, different clocks, different system owners, and different assumptions about what counts as the original record.

That is why a partner to sponsor transfer needs an interface owner with authority across quality, manufacturing, analytical development, regulatory operations, and information technology. The partner owns the provenance of its evidence. The sponsor owns the acceptance of that evidence. The system and integration teams own the controls that preserve it. Quality owns the decision to stop when the chain cannot support a conclusion. Nobody owns the interface by default.

Novartis and Abogen have made the commercial promise visible. The harder promise will be quieter: that ABO2203 and future RNA programs can move from a platform laboratory into a worldwide development and manufacturing network without forcing reviewers to decide which clock, code, or calculation deserves belief.

The paperwork will multiply. Some of it will even be useful. The useful part will carry a stable identity from formulation change to batch record to assay result to quality decision, with enough history for a tired human to verify the path.

If a quality or engineering team has a handoff like this open on its screen, hello@example.com is a good place to put the first concrete failure on the table.