An EDMS manages electronic documents; an eTMF manages the electronic trial master file and its essential records. Choose based on the work you need to control. Document drafting, approval and reuse can belong in an EDMS. Evidence of a particular trial's conduct needs an appropriate TMF arrangement, which may include records held in other controlled systems.
These functions can coexist within one platform or across connected applications. A PDF is not automatically an eTMF record, and a Word document is not automatically an EDMS-only document. The decisive questions are what the record proves, who controls it, which version applies, and how someone will retrieve it later.
For a sponsor with both workloads, start by designing the handoff rather than assuming one purchase replaces the other. This comparison provides a record ownership matrix and a worked amendment example for clinical operations, quality and IT teams. It addresses human medicinal-product trials with U.S. and EU examples; other trial types need their own applicability assessment.
Compare the responsibilities before comparing software
In this article, EDMS means an electronic document management system, not an engineering data management system. The comparison below describes two intended uses. It is not a claim that every product with either label includes every function.
| Decision | Controlled-document EDMS use | eTMF use | Purchasing consequence |
|---|---|---|---|
| Main question | Which approved document should the team use? | What evidence reconstructs this trial's conduct? | Demonstrate both if both are in scope |
| Organizing context | Document identity, owner, revision and lifecycle | Trial, country, site and relevant record category | A folder hierarchy alone does not demonstrate the required relationships |
| Change | Develop and approve the next revision | Preserve the relevant history and distinguish later evidence | Test a revision alongside a previously filed record |
| Completeness | Required document fields and workflow steps | Expected records for the actual trial and milestones | Define what is expected before trusting a completeness percentage |
| External collaboration | Authors, reviewers and other authorized contributors | Sponsor, CRO and investigator responsibilities | Test separate access roles and handover arrangements |
| End of use | Supersession, archival and controlled disposition | Continued retrieval through trial closeout and retention | Include contract exit and archival retrieval in acceptance criteria |
The underlying records need not all be copied into one application. ICH E6(R3), Appendix C, describes essential records maintained in or referred to from sponsor and investigator repositories, with their locations recorded. It also recognizes records relevant to multiple trials that may reside outside trial-specific repositories. These provisions support a documented repository map; they do not excuse a broken link or inaccessible history. See FDA's September 2025 E6(R3) guidance, Appendix C.2.
An eTMF configuration might support drafting and approval. An EDMS configuration might support trial classification and records management. Require the proposed configuration to demonstrate your workflow instead of accepting the category name as proof.
Use this record ownership and transfer matrix
Bring the following worksheet to a joint session with the trial manager, document owner, investigator representative where relevant, records manager and system owner. The entries are illustrative design decisions, not a mandatory filing index. Replace them with the trial's agreed arrangements.
Before completing a row, record the trial identifier, applicable countries and regulatory framework, sponsor/CRO responsibilities, relevant site, record purpose and confidentiality restrictions. Assign a named person behind each owner role. A department name alone does not tell an uploader who can resolve an exception.
| Record and purpose | Accountable decision owner | Proposed authoritative location | TMF relationship and transfer trigger | Evidence to preserve | Stop condition |
|---|---|---|---|---|---|
| Protocol master and approved revision | Sponsor protocol owner | Controlled authoring repository | File or reference the approved revision for the trial after release checks | Document ID, revision, approval evidence and relevant applicability dates | Approval or trial identity unresolved |
| Country/site protocol implementation evidence | Trial manager with site responsibility confirmed | Agreed sponsor/site repositories | Associate the local evidence with the exact protocol revision | Site, country, relevant approval and implementation records | Sponsor release mistaken for permission to implement at every site |
| Company SOP used across trials | Process owner and records manager | Corporate controlled repository | Reference the applicable revision where relevant to the trial | Version-specific location and evidence of required availability | Link always resolves to the newest SOP |
| Investigator-generated confidential record | Investigator/institution | Agreed investigator-controlled location | Document authorized access or permitted transfer, if needed | Record identity, responsible party and access restrictions | Sponsor upload requested without established authority |
| Monitoring report and follow-up | Owner identified in sponsor/CRO agreement | Agreed trial repository | File after the defined review step | Report identity, site/visit context and related correspondence | CRO platform exit would remove sponsor access |
| Trial-specific system configuration evidence | Responsible system owner and trial quality lead | Validation repository or agreed trial repository | Map relevant evidence to the trial configuration | Configuration identity, approval and retrievable supporting evidence | Generic supplier certificate is the only evidence offered |
| Clinical study report | Sponsor report owner | Controlled report repository | Identify the final trial record and its relationship to submission copies | Final version, approval and destination identities | Submission rendition overwrites the retained trial record |
For each row, add a second, short routing record using these fields:
Record key; source ID/revision; destination ID; transfer method; sender; receiving reviewer; transfer date; acceptance result; exception reference; retention schedule/reference; archive owner.
The transfer method should say exactly what happens: retain in place with a version-specific reference, transfer a copy while retaining the original, or replace an original under an approved process. Those are different decisions. “Upload to eTMF” is too imprecise to authorize deletion from the source system.
Use four acceptance results consistently:
- Accepted: identity, context, authority and retrieval checks passed, with evidence recorded.
- Rejected: an observed defect prevents acceptance; identify its owner and corrective action.
- Unknown: necessary context or evidence is unavailable; route to the responsible person and block final acceptance.
- Not applicable: the named owner documents why this transfer is unnecessary; the original record's obligations remain separately addressed.
These results assess the handoff, not the trial's overall compliance. An accepted transfer does not mean a protocol can be implemented, a site can open, or a source original can be destroyed.
Worked example: preserve the version that actually applied
Consider a fictional sponsor running trial ST-204. Its EDMS contains protocol PRO-204 revision 2, approved on March 3. The sponsor's eTMF record TMF-771 identifies that revision. A later revision 3 is approved on April 18. Site 12 has separate implementation evidence for the relevant revision and local context.
The unsafe shortcut is to make TMF-771 a link to “latest approved protocol.” Once revision 3 is released, the same link displays different content. An investigator looking back at March activity can no longer establish the applicable version from that link alone.
EMA's TMF guidance specifically discusses retaining superseded final documents to reconstruct the trial, including version and implementation evidence in the investigator's records. That is a stronger requirement for the handoff design than simply displaying a current approved file. See EMA/INS/GCP/856758/2018, section 3.5.2.
Use the following completed routing example. Dates and identifiers are fictional; the sequence illustrates a control, not an approval timeline.
| Routing field | Revision 2 record | Revision 3 record |
|---|---|---|
| Record key | ST-204 / protocol / rev 2 | ST-204 / protocol / rev 3 |
| Controlled source | PRO-204, revision 2 | PRO-204, revision 3 |
| Sponsor approval | March 3 | April 18 |
| Trial filing identity | TMF-771, version-specific reference | TMF-884, separate version-specific reference |
| Local application | Link Site 12's relevant evidence for revision 2 | Pending verification for Site 12 |
| Receiving check | Retrieved revision 2 and matched approval identity | Retrieved revision 3; local context unresolved |
| Disposition | Accepted as identified historical trial record | Source filing accepted; Site 12 implementation status Unknown |
Complete the remaining routing fields too. In this fictional example, document owner Dana sends the revision 2 reference on March 4; TMF reviewer Lee accepts it that day. The method is retain original and register a version-specific reference, with no source deletion. Trial records lead Morgan owns the archive arrangement under approved schedule RS-ST204. That schedule must contain the applicable trigger and duration; the label alone is not a retention decision. For revision 3, Lee records exception EX-19, assigned to the site coordinator, for the missing Site 12 context. No disposition date is inferred while the schedule's trigger has not occurred.
The final row matters. Filing revision 3 at sponsor level and confirming its implementation at Site 12 are separate outcomes. Do not make a single green status imply both.
For this example, the handoff acceptance test is straightforward: approve revision 3, then retrieve TMF-771 as an authorized reader. It must still identify revision 2 and its retained evidence. Retrieve TMF-884 separately and confirm that the unresolved Site 12 status remains visible. Ask a reader to reconstruct the sequence without help from the uploader.
“Freeze the filed version” means preserve that identity and history. It does not mean prohibit every later correction. A misclassified record may need correction through the agreed traceable process. A later amendment may need a new record or a linked version according to the configured filing model. The acceptance criterion is successful reconstruction, not one universally mandated screen layout.
Boundary case: an attachment with no trial or site context
A CRO sends protocol_final_signed.pdf with no reliable revision, trial identifier or country/site context. A familiar filename and a signature image are insufficient to complete the worksheet.
Place the item in the agreed restricted intake process, assign the sender or record owner to resolve the missing information, and mark final filing acceptance Unknown. Do not infer ST-204 from a nearby folder, overwrite a current protocol, or count the record as satisfying a site-specific expectation. After clarification, preserve the correspondence resolving its identity and repeat the acceptance checks.
An automated classifier or AI assistant may propose a match. Treat that as a suggestion until the required evidence resolves the ambiguity. This is especially important when similar protocols, translations or site versions share titles. The same stop rule should apply to a person and an automated uploader.
Sponsor and investigator records need separate access decisions
Centralizing software does not make the sponsor the appropriate recipient of every investigator record. A sponsor portal and an investigator repository may share infrastructure while retaining different authority and access arrangements.
EMA's TMF guidance separates sponsor and investigator responsibilities, discusses investigator control and continuous access, and identifies records that should be segregated. Its examples include the subject identification code list in the investigator TMF. Use that distinction before designing a bulk transfer into a sponsor-managed workspace. See sections 3.1 and 4.1.3 of the EMA TMF guideline.
For procurement, demonstrate three roles with the same sample record: the authorized site reader, an authorized sponsor reader, and someone who must not see it. Repeat the demonstration after the trial closes and after the operational portal account would normally expire. Record who grants access, who can export, and how access survives the service agreement ending.
Do not treat a successful administrator demonstration as proof that investigators have suitable access. Test the actual roles and the expected operating conditions, including a closed site or departing CRO. If the arrangement cannot explain who receives the retained records and how they open them, the archive design is unfinished.
Apply clinical-system guidance and retention rules to the actual scope
For U.S. clinical investigations, FDA's October 2024, Revision 1 electronic-systems guidance addresses electronic records, systems and signatures. It is nonbinding guidance and discusses Part 11 in relation to applicable underlying record requirements. It does not make every file in every clinical department subject to identical controls. Its Q3 discusses certified copies when replacing originals; Q5 addresses retention of records and associated metadata; Q7 discusses risk-based validation. See the FDA final guidance.
For this matrix, that means choosing “replace original” requires a separate evidence review. A readable PDF alone does not establish that the replacement preserves the required information and context. Include the intended use and transfer boundary in the system assessment; our computerized system validation guide provides broader planning background.
The current GCP sources also have different regional dates. FDA issued its final E6(R3) guidance in September 2025. EMA lists the principles and Annex 1 as effective from July 23, 2025. As checked on October 6, 2026, EMA lists Annex 2 as adopted with an effective date of January 15, 2027. Do not describe that EU effective date as already reached. See FDA's E6(R3) status page and EMA's current E6 page.
Retention likewise needs a scoped decision:
| Context | Rule or boundary | Matrix consequence |
|---|---|---|
| Trials under EU Clinical Trials Regulation 536/2014 | Article 58 requires TMF content retention for at least 25 years after trial end, unless other Union law requires longer; subjects' medical files follow national law | Record trial scope, end-date basis and applicable exceptions |
| U.S. IND investigator records | Under 21 CFR 312.62(c), two years after marketing approval for the investigated indication; if no application will be filed or it is not approved for that indication, until two years after the investigation is discontinued and FDA is notified | Do not start a generic two-year clock from last patient visit |
| U.S. IND sponsor records | Under 21 CFR 312.57(c), two years after marketing approval for the drug; if no application is approved, until two years after investigational shipment and delivery stop and FDA has been notified | Confirm the sponsor's trigger separately from the site's |
| Unknown or mixed jurisdiction | No single duration is established by the filename or software category | Resolve applicable schedules before authorizing disposition |
The cited requirements are in Article 58 of the consolidated EU regulation, 21 CFR 312.62(c) and 21 CFR 312.57(c). These examples do not establish a universal retention period for all medical-device investigations, all countries or all corporate documents.
Which investment should come first?
Scenario A: authoring is controlled; trial evidence is fragmented. A sponsor can identify approved protocol versions, but its CRO handover has unresolved site records, incomplete expected-record definitions and uncertain archival access. Prioritize the TMF operating model and evaluate an eTMF configuration against it. Another authoring repository will not answer those specific gaps. An existing EDMS could still be suitable if its demonstrated configuration fulfills that trial workload.
Scenario B: trial filing works; approved sources are unreliable. Clinical operations can retrieve accepted trial records, but report authors exchange attachments and cannot reliably distinguish the approved source from a later draft. Prioritize controlled authoring and the EDMS-to-TMF handoff. Replacing the working eTMF may add migration work without fixing the source approval process.
These are opposite investment priorities, not opposing regulatory conclusions. If the evidence changes - such as discovering that Scenario A's current platform already provides the needed trial functions - the decision should change too. Buy missing capability only after distinguishing software gaps from configuration, ownership and procedural gaps.
Compare total operating effort across each proposal: authoring administration, TMF configuration, migration reconciliation, interface exception handling, external-user support, validation work, archive services and contract exit. Allocate an owner and quoted scope to each. Do not assume one platform eliminates those tasks, or that two applications necessarily cost more overall. Without comparable proposals, there is no defensible cost winner.
Finish with a demonstrated handoff
Ask the proposed operating team to run the ST-204 example, the unidentified attachment and the restricted-site record through its configuration. Accept the design only when the owners can show the correct version, explain unresolved statuses, enforce the intended access and retrieve the retained evidence. Repeat the retrieval using the planned archive or export arrangement, rather than relying entirely on a live supplier account.
A clinical report may subsequently become part of a regulatory submission, adding another destination with its own identity and lifecycle. Our eCTD Module 5 guide explains that downstream organization. Submission preparation should not silently replace the trial's retained record.
For an Assyro evaluation, bring the completed matrix to a workflow discussion and identify the specific document-management or regulatory handoff you want assessed. This comparison does not establish Assyro as an eTMF replacement. Confirm the proposed scope against your trial responsibilities before changing the system that holds the records.
About the author
Assyro Team
Expert regulatory operations consultants helping pharmaceutical companies navigate complex compliance challenges.

