2
Approved
4
Declined
67%
Decline rate
4
Absorbed, never registered
1. The Register
Six change requests reached the Development Committee. Two were approved and four were declined — a decline rate of 67%.
| Ref | Change | Raised by | Decided | Outcome | Reasoning | Full form |
|---|---|---|---|---|---|---|
| CR-01 | Add an Asian-population cohort to Phase 2 | Program | 09 Oct 2024 | Approved | Increases N by 50. Supports the later EU and Asia-Pacific strategy without a separate bridging study. | record |
| CR-02 | Add a cardiovascular outcomes sub-study to Phase 3 | Program | 20 Apr 2026 | Approved | Adds $6,400,000 and six months to the Phase 3 window. Pre-empts a post-marketing requirement (D-06). | record |
| CR-D1 | Add a head-to-head comparative tolerability trial | J. Barrington | 2025-11-14 | Declined | ⚠⚠ The most consequential decline in the program. Would have cost roughly $34,000,000 and eighteen months against a fixed patent expiry. Defensible on the information available — and it is the evidence a payer asks for first, which this program will not have. | record |
| CR-D2 | Extend Phase 3 to a third pivotal in an adolescent population | Dr. A. Okoye | 2026-02-03 | Declined | PREA is satisfied by the agreed iPSP with a post-approval deferral. Running it now would add cost and calendar to satisfy an obligation already deferred. | record |
| CR-D3 | Switch the primary endpoint to percent achieving ≥10% reduction | Dr. F. Achterberg | 2026-01-20 | Declined | ⚠ Statistically attractive and agreed with the agency at End-of-Phase-2 as a KEY SECONDARY. Changing a co-primary after EOP2 alignment would have reopened a settled question and invited a new one. | record |
| CR-D4 | Add a second contract manufacturer for drug product | Dr. K. Oyelaran | 2026-04-28 | Declined | Technology transfer, a second validation and a second pre-approval inspection — roughly two years, against a filing date that does not move. Recorded as an accepted single-source risk instead. | record |
A register recording only approvals is a log of things that happened, not a control.
The four declines are what make this evidence that decisions were taken. Each names who proposed it, when it was decided, and why the answer was no — and every one was a reasonable proposal made by a senior person in good faith, which is exactly the category a change control process exists to handle.
The fact base asserts it: if approvals outnumbered declines, the build would fail on the grounds that a body approving most of what it receives is recording intentions rather than exercising control.
The four declines are what make this evidence that decisions were taken. Each names who proposed it, when it was decided, and why the answer was no — and every one was a reasonable proposal made by a senior person in good faith, which is exactly the category a change control process exists to handle.
The fact base asserts it: if approvals outnumbered declines, the build would fail on the grounds that a body approving most of what it receives is recording intentions rather than exercising control.
2. CR-02 — A Change That Was Approved
Full form: CR-02 record
Adding a cardiovascular sub-study, $6,400,000, approved 20 April 2026. Worth walking end to end because it shows what a well-formed change looks like.
| Step | What happened |
|---|---|
| Raised | By the Chief Medical Officer, with the alternative stated: do nothing, and accept a likely post-marketing requirement imposed at approval instead. |
| Impact | $6,400,000, six months added to the Phase 3 window, a new endpoint, a separate analysis and a standalone clinical study report. ⚠ Checked against the traceability matrix: it creates an entirely NEW chain rather than altering an existing one. |
| Routed | Tier 4 — scope and anything touching the submission. Development Committee, not the Program Director. |
| Decided | Approved. Funded from contingency; the ceiling did not move. |
| Baselined | Held as a discrete work package (1.6.5) rather than absorbed into the Phase 3 clinical account. |
| Verified | Sub-study protocol issued to the DMC 12 August 2026, closing gate condition GC-02. |
The reason it could be approved mid-program at all is that it touched nothing existing.
A change that had altered an existing analysis would have required reopening the testing hierarchy — and the hierarchy is shared state, so a local change there has program-wide consequence. CR-02 added a chain instead of editing one, which is why a six-month, six-million-dollar addition was tractable eighteen months into Phase 3.
Keeping it as a discrete work package is the other half. Absorbed into the clinical account it would have been invisible within one reporting cycle, and at closure nobody could have answered what the decision cost.
A change that had altered an existing analysis would have required reopening the testing hierarchy — and the hierarchy is shared state, so a local change there has program-wide consequence. CR-02 added a chain instead of editing one, which is why a six-month, six-million-dollar addition was tractable eighteen months into Phase 3.
Keeping it as a discrete work package is the other half. Absorbed into the clinical account it would have been invisible within one reporting cycle, and at closure nobody could have answered what the decision cost.
3. CR-D1 — The Decline That Mattered
Full form: CR-D1 record
A head-to-head comparative tolerability trial, proposed by Market Access in November 2025. Declined. It is the largest what-if in the program.
| What it would have provided | ⚠ The comparative evidence a payer asks for first, and the clinical claim the label strategy is built on — generated directly rather than inferred from an indirect comparison. |
| What it would have cost | Roughly $34,000,000 and eighteen months, against a patent that expires on a fixed date. |
| Why it was declined | The eighteen months came off the commercial life at the far end, the cost was outside contingency, and the tolerability endpoint was already positioned in the Phase 3 hierarchy — so the evidence would be generated anyway, if the endpoint cleared. |
| What actually happened | ⚠⚠ The endpoint sits fourth in a six-position hierarchy. It is generated, and it is claimable only if three endpoints above it succeed first. |
The reasoning was sound and the exposure it created is now the largest single unknown in the
program.
On the information available in November 2025 the case against was strong: a dedicated trial cost eighteen months of patent life to generate evidence the pivotal program was already positioned to produce. ⚠ What the decision did not fully price is that positioned to produce and able to claim are different things, and that the difference is decided by a testing hierarchy nobody was weighing commercially at the time.
Whether that pricing error matters will not be known until the pivotals read out in 2028. What can be said at this status date is narrower and still worth saying: the program's commercial differentiation now rests entirely on an endpoint sitting fourth of six, and the decision that put it there was taken on schedule and cost grounds.
Recording the decline with its reasoning is what makes that visible now. Had the register held only approvals, the question — why didn’t we run a head-to-head? — would have no answer on file, and the honest answer is better than an absence.
On the information available in November 2025 the case against was strong: a dedicated trial cost eighteen months of patent life to generate evidence the pivotal program was already positioned to produce. ⚠ What the decision did not fully price is that positioned to produce and able to claim are different things, and that the difference is decided by a testing hierarchy nobody was weighing commercially at the time.
Whether that pricing error matters will not be known until the pivotals read out in 2028. What can be said at this status date is narrower and still worth saying: the program's commercial differentiation now rests entirely on an endpoint sitting fourth of six, and the decision that put it there was taken on schedule and cost grounds.
Recording the decline with its reasoning is what makes that visible now. Had the register held only approvals, the question — why didn’t we run a head-to-head? — would have no answer on file, and the honest answer is better than an absence.
4. The Lifecycle, and Where Requests Are Routed
| Step | Who | What must happen |
|---|---|---|
| Raise | Anyone | A change request states the change, the driver, and the option of NOT doing it. ⚠ A request with no stated alternative is a proposal, not a request. |
| Impact | PMO with the affected owners | Cost, schedule, scope, quality, risk. Checked against the traceability matrix for what else it touches. |
| Route | PMO | To the authority tier the impact demands, never to the tier that is easiest to reach. |
| Decide | The named authority | Approve, decline, or defer with a stated review date. ⚠ Deferral is a decision and is recorded as one. |
| Baseline | PMO | An approved change updates the scope, schedule and cost baselines together, or none of them. |
| Verify | Control account owner | The change is confirmed implemented before the request is closed. ⚠ Approved is not done. |
| Impact | Authority | How it is recorded |
|---|---|---|
| Within a work package, no cost or date impact | Control account owner | Recorded in the control account; no register entry |
| Across work packages, within the control account envelope | Control account owner | Change note, notified to the PMO |
| Cost impact within contingency, no date or scope change | Program Director | Contingency draw recorded WITH the rejected alternative |
| Scope, milestone dates, or anything touching the submission | Development Committee | Change request; CR-01 and CR-02 are the two raised |
| The authorized ceiling | Board | Ceiling is not a Committee matter at any size |
| Anything affecting participant safety | Immediate — no approval required | Safety bypasses commercial escalation entirely |
Routing to the tier the impact demands, never to the tier that is easiest to reach.
The pressure runs one way. A change routed to a control account owner is decided this week; the same change routed to the Committee waits for a quarterly meeting. Every incentive favors under-classifying, and the classification is made by the people who want the change.
The defense is that the tiers are defined by impact rather than by size or urgency, and the PMO routes rather than the requester. It is imperfect — the impact assessment is itself a judgment — and it is considerably better than letting the requester choose their own audience.
The pressure runs one way. A change routed to a control account owner is decided this week; the same change routed to the Committee waits for a quarterly meeting. Every incentive favors under-classifying, and the classification is made by the people who want the change.
The defense is that the tiers are defined by impact rather than by size or urgency, and the PMO routes rather than the requester. It is imperfect — the impact assessment is itself a judgment — and it is considerably better than letting the requester choose their own audience.
Two tiers are worth noting for what they refuse. The ceiling is never a Committee matter at any size, so no amount of Committee enthusiasm can authorize an overrun. And anything affecting participant safety requires no approval at all — it is implemented immediately and reported afterward, because a control that could delay a safety action would be worse than no control.
5. What Never Reached the Register
The honest limit of any change control process: it governs the changes that are recognized as changes.
| Never reached the register | Why, and what it cost |
|---|---|
| Protocol clarifications issued as letters, not amendments | ⚠ Legitimate where genuinely clarifying, and the boundary is thin. A 'clarification' that changes what a site does is an amendment wearing a cheaper name. |
| Additional monitoring visits at underperforming sites | Within the CRO's contracted scope, so no change order. It consumed monitoring capacity budgeted for elsewhere and appears in no register. |
| Extra analytical method qualification runs after the transfer failure | Absorbed into the CMC control account envelope under tier 2 authority. Correct by the rules, and it means the true cost of I-02 is spread across two records. |
| Growing scope of the market access evidence package | ⚠ Driven by GC-03 rather than by a change request. A gate condition can expand work without ever passing through change control. |
Every item above was handled correctly under the rules, and together they mean the register
understates how much this program actually changed.
Extra qualification runs went into the CMC envelope under tier 2 authority — correct, and it means the true cost of the method transfer failure is split across two records. The market access evidence package expanded under a gate condition, which can grow work without ever passing through change control. Additional monitoring sat inside the CRO's contracted scope and consumed capacity budgeted elsewhere.
None of this is a defect to be fixed by tightening the tiers. A process that routed every clarification and every extra monitoring visit to a quarterly Committee would stop the program. The useful response is to know the register is a floor rather than a complete account, and to read it alongside the cost variance, which sees what the register does not.
Extra qualification runs went into the CMC envelope under tier 2 authority — correct, and it means the true cost of the method transfer failure is split across two records. The market access evidence package expanded under a gate condition, which can grow work without ever passing through change control. Additional monitoring sat inside the CRO's contracted scope and consumed capacity budgeted elsewhere.
None of this is a defect to be fixed by tightening the tiers. A process that routed every clarification and every extra monitoring visit to a quarterly Committee would stop the program. The useful response is to know the register is a floor rather than a complete account, and to read it alongside the cost variance, which sees what the register does not.
The first row is the one to watch. A protocol “clarification” that changes what a site does is an amendment wearing a cheaper name, and the boundary is genuinely thin. It is policed by the medical monitor rather than by the PMO, because the judgment required is clinical rather than administrative — and because the person best placed to spot it is the one who would otherwise benefit from the cheaper route.