Expected evidence to filed artifact. Completeness, timeliness, quality, provenance, inspection, and archive in one trial record.

An electronic trial master file is not a document repository with numbered folders. It is the controlled evidence that a clinical trial was planned, conducted, monitored, analyzed, reported, and closed appropriately across sponsor, country, site, vendor, and time.

Seal models expected evidence before it arrives and preserves every artifact’s provenance, classification, quality review, approval, filing, relationship, replacement, timeliness, access, and archive state.

Model name, version, zones, sections, artifact types, definitions, required metadata, milestones, responsible parties, retention, supersession, and change notes define the reusable structure.

A model upgrade is assessed and adopted; it does not silently reclassify historical evidence.

Protocol, phase, design, countries, sites, committees, vendors, systems, products, procedures, milestones, responsibilities, filing conventions, QC strategy, timeliness targets, inspection approach, archive plan, and approved exceptions define the study-specific file.

Study start, country activation, site selection, approval, initiation, amendment, enrollment, monitoring, safety event, vendor deliverable, database lock, closeout, submission, and archive can create evidence expectations.

Expected, not applicable, substituted, waived, overdue, filed, superseded, and missing remain distinct states with rationale.

An artifact’s study, country, site, investigator, committee, vendor, product, system, subject scope where permitted, milestone, and effective dates determine where it belongs and what it covers.

One global document can satisfy several expectations without being duplicated into unrelated files.

Source system, sender, channel, transfer, file name, format, checksum, received time, created time, document date, author, organization, signature state, source identifier, duplicate candidate, and original package remain captured.

Email, portal, e-signature, CTMS, EDC, safety, laboratory, vendor, scan, and bulk migration sources remain distinguishable.

Artifact type, zone, section, study, country, site, milestone, document date, version, language, author, organization, subject, confidentiality, expected-artifact match, confidence, and reviewer define filing.

Automatic suggestions accelerate work but cannot conceal uncertainty or overwrite an approved classification.

Correct document, complete content, readable pages, signatures, dates, version, metadata, filing location, scope, duplicates, redactions, language, certification, replacement relationship, and inspection usability form the QC decision.

Failed QC retains findings, return path, corrected artifact, resubmission, reviewer, and final outcome.

Authoring, review, approval, execution, electronic signature, wet-signature certification, filing QC, and finalization are separate states with assigned roles and evidence.

A document can be correctly filed but not yet approved—or approved but filed in the wrong scope.

Draft, final, executed, corrected, translated, certified copy, redacted copy, superseded version, amendment, addendum, duplicate, and replacement remain related to one artifact family.

The effective document for a country or site resolves by scope and date without deleting the earlier state.

Required artifacts, applicability, milestone due dates, received content, QC, approvals, supersession, substitutions, waivers, and approved exceptions calculate completeness by study, zone, country, site, vendor, milestone, and owner.

Event or document date, availability date, receipt, submission, QC start, QC completion, filing, rejection, resubmission, finalization, target, pause reason, and exception define timeliness.

The metric distinguishes late creation, late transmission, late QC, and late filing.

QC findings, rejections, metadata errors, duplicates, missing signatures, wrong versions, unreadable scans, late documents, filing corrections, owners, vendors, sites, artifact types, root causes, actions, and recurrence support quality oversight.

Message, sender, recipients, date, subject, thread, attachment, decision, commitment, follow-up, confidentiality, filing rationale, artifact type, scope, QC, and relationship to formal documents define relevant correspondence.

The system does not file every email; it preserves the reason a communication is essential.

Expected deliverables, transfer schedules, format, metadata, responsible party, access, submissions, rejects, resubmissions, completeness, quality, timeliness, reconciliations, final transfer, acceptance, and archive define external contribution.

Sponsor, CRO, site, vendor, auditor, inspector, study, country, site, zone, artifact type, confidentiality, action, start, expiry, approval, periodic review, access history, revocation, and export remain controlled.

Scope, expected artifacts, completeness, overdue content, QC failures, unapproved items, access, redactions, translations, audit trails, retrieval tests, open actions, owners, and readiness decisions remain visible before an inspector asks.

Inspection workspaces expose approved evidence and record exactly what was viewed or exported.

CTMS milestones, sites and monitoring; EDC subjects and lock; safety cases; vendor inventories; quality events; document systems; and archive manifests are compared against expected TMF evidence.

Matched, missing, extra, discrepant, pending, accepted, and resolved states retain the reconciliation rule and reviewer.

Final expectations, reconciliation, outstanding actions, certified-copy checks, access removal, inventory, manifest, checksums, retention basis, archive transfer, receipt, readability, restore test, ownership, and approval define closure.

Seal is strongest where trial operations, countries, sites, vendors, clinical data, quality, documents, access, inspections, and archive intersect. It owns the expected-evidence and artifact-state model while composing CTMS, EDC, clinical operations, DMS, training, and quality.

The first implementation should follow one protocol through study setup, country approval, site selection and activation, staff qualification, initiation, amendment, monitoring, safety correspondence, vendor evidence, database lock, closeout, inspection request, final reconciliation, and archive.

Include a wrong-country filing, missing signature, translated replacement, late vendor transfer, duplicate artifact, approved substitution, revoked inspector access, and an archive restore test. The system must show what was expected, what satisfied it, and why.