Why protocols multiply, and why copies drift
A clinical program that is working produces more protocols. The Phase 1 study that reads out well becomes a Phase 2 protocol with the same product, the same dose rationale and most of the same safety language. A compound that works in one indication is tried in a second, with eligibility criteria that differ in three lines out of forty. A large study grows a sub-study that runs alongside it at a subset of sites, sharing its visits and its criteria. A program can also need two versions of the same protocol in parallel, one per region or population, agreeing on everything except what genuinely differs.
In a Word-based process each of these is a copy, and a copy starts drifting the day it is made. The source protocol gets an amendment; nobody re-reads the sub-study against it. The derived protocol’s safety section is tightened in review; the parent keeps the old wording. Six months later the two documents disagree in ways nobody chose, and finding the differences means reading both end to end. The cost is not the copying, it is the keeping track.
Trelice replaces the copy with a derivation: a recorded, audited act of creating one protocol from another, with a living link between them.
Creating a protocol from an existing one
Every protocol card carries a Create from this button, and the same action sits in the editor’s menu. It opens a short guided flow. First the reason: sub-study, new indication, next phase, or another reason in your own words. The reason goes into the audit trail with the new protocol’s record of where it came from. Then the source: the live draft, or a numbered version frozen at approval, which is usually the safer choice because you copy exactly what was approved. Then the destination study, which can be any study in your organization, including one of a different study type, in which case sections that do not exist for that type are skipped and the summary says so.
The fourth step is the heart of it: picking the content. The screen shows the source protocol’s sections as a tree, everything selected by default. Untick a section to leave it out, or expand it and choose field by field. Lists expand into individual items, so you can copy three of five study arms, or only the screening and baseline visits. A counter above the tree keeps score, for example 14 items across 6 sections.
Some entries are locked and never copied, because the new protocol must get them fresh: the protocol number, version, date, amendment history and the signature page. Fields that the protocol states in several sections, such as the planned number of subjects, are ticked and unticked together, so the copy cannot leave one section saying 100 and another blank. The Schedule of Assessments brings its own logic: copying an assessment automatically brings the visits it refers to, because an assessment without its visits is not a schedule.
The final step summarizes the whole choice in one sentence and creates the protocol. A results panel lists every section with its outcome, says how many fields will be watched for drift, and opens the new protocol’s editor. The copied sections start as drafts and go through the normal section review, exactly like hand-written content. The full flow is in the guide: creating protocols from existing ones.
Derivation templates: decide once what a sub-study copies
Organizations that run the same kind of study repeatedly make the same selection repeatedly: the eligibility criteria, the study design, the schedule, never the statistics. A derivation template remembers that choice. On the pick-content step, Save as template stores the selection under a name, for example Sub-study standard, with an optional description and a default reason. From then on the template is offered at the start of every derivation in your organization, and applying it preselects the tree.
A template stores what to copy, never any protocol content. That is what makes it reusable across different source protocols: applied to a protocol with seven visits instead of five, a template that said all visits selects all seven, and a template that named two visits finds them by name. Anything the template names that the new source does not have is simply skipped, and the screen says what was skipped and why.
Templates are managed on their own settings page, where each card shows what it copies, section by section, who saved it and when. A template can be edited, switched off without deleting it, or deleted outright; protocols already derived with it keep its name in their history either way. The derivation itself records the template used, so an auditor can see not just that a protocol was derived, but by which organizational standard. The details are in the derivation templates guide.
Drift alerts: when linked protocols stop agreeing
Copying is only half the story. Every copied field stays linked to its source, and the link works in both directions. Suppose the source protocol’s planned number of subjects changes from 100 to 150 while the sub-study still says 100. Trelice raises a drift alert on the sub-study. Editing the sub-study alerts the source side the same way. Nothing is ever changed automatically: a drift alert is a notice that two protocols which used to agree no longer do, addressed to a person who decides what that means.
Drift is hard to miss. The section’s assignee gets a notification and an entry in pending tasks. The affected section shows a marker in the editor’s outline. The created-from bar at the top of the protocol turns amber and counts the drifted fields. Beside the field itself sits a chip labeled Drifted. On the dashboard, a drift column counts open alerts per study, so a program lead sees at a glance which studies have unreviewed divergence.
When a source protocol is amended, the alert says so: the drift is stamped with the version it appeared in, so the sub-study team knows the parent changed at amendment 2, not in a casual edit.
Resolving drift: three honest answers
Opening a drifted field shows both values side by side, with the differing words highlighted and the protocol where the change happened named. There are exactly three answers, because there are only three true states. Keep mine says my value is right for this protocol; the alert closes but the field stays linked, so future changes still alert. Take theirs writes the linked protocol’s value into mine through the normal edit path, so review status and amendment rules apply. Unlink field says these two protocols are meant to differ here, permanently; no further alerts for that field.
A Drift review panel gathers every open alert in one place, grouped by section, each row with both values and the same three buttons, so a writer returning from two weeks of parent-protocol amendments can work through the whole backlog in one sitting. Every resolution is recorded in the audit trail with who decided and how.
Drift participates in the approval gates. A section cannot be approved while it has unresolved drift from its source: the reviewer is asked for each linked field to be resolved first, the same way golden threads block approval within one protocol. If taking a linked value would touch an already-approved section, Trelice asks before reopening it. When two protocols have permanently diverged, the whole link can be dissolved with Unlink protocols; the protocol’s origin stays visible in its history, marked unlinked.
Drift control does not stop at the protocol
Each derived protocol is a full protocol, so it generates its own downstream documents: its own consent forms, visit worksheets, eligibility screener and site documents, all from its own approved content. When the protocol changes, those documents flag for regeneration, the same amendment cascade every Trelice protocol has. So the chain holds end to end: a change in the parent protocol raises drift on the derived protocol, resolving the drift changes the derived protocol through its normal review, and that change flags the derived protocol’s own documents. At no point does a person have to remember which documents depend on which protocol.
Product facts have their own upstream anchor. When the studies share an investigational product, the Investigator’s Brochure holds the product facts once, prefills them into each protocol, and raises its own alerts when a new edition changes them. Derivation links carry the study facts between protocols; the brochure link carries the product facts into all of them.
What is recorded, and what is protected
Every derivation is recorded on both protocols: the source knows its children, the child knows its parent, the reason, the source version and the template used, even after the template is deleted. Every drift alert and every resolution is audited with the deciding user. Deleting or archiving the source study does not touch the derived protocol: it keeps its content and remembers where it came from.
Derivation respects the same boundaries as everything else in Trelice. Only studies of your own organization are offered as destinations, every query is scoped to your organization, and roles decide who can derive at all. The people who cannot create protocols do not see the button.
Getting started
Derived protocols and derivation templates are part of protocol authoring; there is nothing separate to buy or switch on. Open any protocol card, click Create from this, and the flow walks you through it. The two guides cover every step and every badge: creating protocols from existing ones, and drift alerts and derivation templates.