What this page is
ICH E6(R3) is the GCP guideline. It puts weight on quality by design, risk-proportionate monitoring and the integrity of data across its whole life. This page says, topic by topic, what Trialion does for the sponsor, the investigator and the data, and what it does not. Each row names the page that holds the detail and the screens; the questions an auditor would ask, with a status, are on those pages and are collected in the framework matrix.
Nothing here is a statement that a trial is GCP compliant. That is a property of how a trial is run.
Quality by design
| Expectation | What Trialion does |
|---|---|
| Critical-to-quality factors identified up front | The verification plan decides which forms and fields a monitor must verify against source, so effort goes where the protocol says it matters (Monitor and verify data) |
| A design that is controlled and versioned | The whole study design is one versioned configuration, tested in a UAT cycle and released only after three sign-offs by three different people, each bound to the exact design approved (Review and release a study, Sign off and go live) |
| Validation before use | UAT is a full working workspace with its own participants and records, kept apart from Live by the database, in which a failed test is kept as evidence with what was done about it (Test a version in UAT) |
| Changes under control | An amendment is a new version, tested and approved the same way, and participants move to it one at a time, after re-consent where the amendment asks for it (Amend a live study) |
Investigator responsibilities
| Expectation | What Trialion does |
|---|---|
| Supervision of the site's staff | A site's staff get no access until the principal investigator signs their delegation of authority with a fresh code; nobody signs their own entry (Staff a site) |
| A delegation log | The site's staff page is the log: who, which tasks, from when, signed by whom; entries are ended and amended by new entries, never deleted |
| Eligibility decided by the investigator | Only the investigator can record an eligibility determination, and enrollment is refused without one (Screen a participant, Enroll a participant) |
| Informed consent before any procedure | A participant cannot be recorded without a consent date and an ICF version; a changed protocol is followed by a re-consent record that moves the participant to the new version (Record consent and re-consent) |
| Accurate source records and corrections | Saved values are read-only; a change needs a query from someone else and a reason, and the earlier value is kept (Correct data) |
| Attestation of the data | The investigator's signature on a record locks it, with the meaning, the time and a fingerprint of the values; a visit and a casebook can be signed where the study asks (Sign records) |
| Withdrawal and early exit documented | Withdrawal, lost to follow-up, screen failure and the other exits each need a reason from a standard list, a date and, for a withdrawal, a record of whether a safety follow-up is owed (Change a participant's status) |
Sponsor oversight and monitoring
| Expectation | What Trialion does |
|---|---|
| Risk-based source verification | A plan by form and field, a monitor's queue of what is pending, verified, out of date or discrepant, and a verification that no longer stands if the verified value is later changed |
| Verification independent of data entry | Only the monitor role records a verification; the person who sets the plan, an administrator and the site cannot |
| Follow-up of data quality | Queries with separation of duties: the person who raised a query cannot answer it, and the person who gave the latest answer cannot close it, enforced by a database trigger as well as the screen (Work with queries) |
| Protocol deviations managed | Deviations are recorded with category, severity, whether they are important or reportable, status, site and participant, and reviewed (Report a deviation) |
| Central monitoring and key risk indicators | Not built. There are no risk indicators, central statistical monitoring or automated risk signals |
| Monitoring-visit reports | Not built |
| Oversight that cannot alter the site's data | The Site Monitor and the Auditor read and raise queries and change nothing; the Project Manager runs the study and cannot alter the data it collects (Site Monitor, Project Manager) |
Ethics and site conduct
| Expectation | What Trialion does |
|---|---|
| Ethics approval before enrollment | A site cannot be activated without an approval reference and date and a GCP training confirmation, and only an active site can enroll (Open a site) |
| Training | The confirmation at activation is enforced. The training complete flag on a team assignment blocks nothing; there is no per-person training record with versions and expiry |
| Adherence to the protocol | Visit windows (entry before a window opens is refused unless a person records an exception with a reason; late entry is allowed and reported), mapped forms and the study's rules encode the protocol (Run a visit) |
| Fresh ethics approval for an amendment at each site | Not built. A version goes Live for the whole study; the site's approval details on a version are recorded, not a gate |
Data governance and the data lifecycle
| Expectation | What Trialion does |
|---|---|
| Data integrity across the lifecycle | See ALCOA+ |
| Computerised systems fit for purpose | Tested and released under gates (Testing and quality); no validation package is supplied (GAMP 5 and CSV) |
| Controlled changes to data | A correction needs a reason and an authorizing query; every change is in the audit trail (Audit trail and evidence) |
| Access limited to those who need it | Roles with separation of duties, site confinement, and refusals recorded. A blinded study cannot be run: the arm is returned to every role that may read participants (Not built yet) |
| Database lock before analysis | Two people sign a request and a confirmation, each with a fresh code, on the same fingerprinted list of what was open; every write is refused afterwards (Close and lock a study) |
| Essential records kept and retrievable | A document register with versions, checksums, approvals and the version in force on a date (Manage documents). It is not a TMF: no reference model, no completeness checklist |
| Certified copies | CSV, Excel and ODM-XML exports, the casebook and the audit trail (Export data) |
| Retention | Nothing governed is deleted; the retention period, backup and restore are the hosting's (Running it) |
Safety
| Expectation | What Trialion does |
|---|---|
| Capture of adverse and serious events | On-demand records, individually numbered, never deleted (Record an adverse event) |
| Awareness time and the reporting clock | The time the site became aware is recorded and a due time derived from the version's setting |
| Follow-up and independent medical review | Follow-up reports, a medical review with causality and expectedness (a related and unexpected event is flagged as a suspected unexpected serious adverse reaction), and a reconciliation record (Review a serious event) |
| Expedited reporting to authorities | Not built. No E2B file, no gateway, no automatic reconciliation with a safety database, and nobody is told when a deadline nears |
Trial supply and randomization
Not built. No randomization list or allocation (the arm is typed in from the randomization system), no blinding or unblinding, no drug supply or accountability, and no medical coding. All are listed with what to do instead on Not built yet.
Compliance Q&A
In part. A verification plan targets source data verification by form and field, a monitor's queue shows what needs attention, and a verification stops standing when its value changes.
Limits: there are no key risk indicators, no central statistical monitoring, no automated risk signals and no monitoring-visit reports (Not built yet).
Evidence: server/src/modules/study/sdv (plan, verification ledger, monitor queue)no key risk indicator or central monitoring code in server/src
What this means for a monitoring plan
Trialion supports risk-based source verification, delegation of authority, ethics gating, controlled change and data integrity well enough to build a monitoring plan around. It does not replace your safety system, your randomization and supply system, your training records or your TMF, and it does not by itself satisfy the expectation that a computerised system is validated for its intended use.
Related
Framework matrix · ALCOA+ · Audit trail and evidence · GAMP 5 and CSV · Not built yet