12
Criteria
10
Must-meet
5
Assessable today
24
Months to the gate
Contents
1. Tracker, Not Gate
Gate 5 will be held on 30 September 2028, 24 months from the status date. This is not that decision, and it does not pre-empt it. It is the tracker that runs between the two gates so that when the Committee sits down, nothing on the table is new.
| This assessment | The gate decision | |
|---|---|---|
| When | Continuously, from Gate 4 (Jun 2026) onward | Once, on 30 Sep 2028 |
| Who | Program Director, against criteria fixed at Gate 4 | The Development Committee |
| Produces | A position, and an early warning where one is available | A decision, minuted, with conditions |
| Authority | None. It cannot pass or fail anything. | Releases the Stage 5 tranche and authorizes the submission |
| Value | ⚠ Entirely in removing surprise. A gate that surprises the Committee has been governed badly for the two years preceding it. | The decision itself |
The criteria were fixed at Gate 4 and are not renegotiable by the people who have to satisfy
them.
That is the rule this whole artifact depends on. If the criteria could be adjusted as the gate approached, a readiness assessment would measure nothing — the target would move to meet the position, and the tracker would report green throughout.
The same discipline the Committee applies to gate conditions applies here: the standard is set by the body that will judge it, at a time when nobody knows who will fall short.
That is the rule this whole artifact depends on. If the criteria could be adjusted as the gate approached, a readiness assessment would measure nothing — the target would move to meet the position, and the tracker would report green throughout.
The same discipline the Committee applies to gate conditions applies here: the standard is set by the body that will judge it, at a time when nobody knows who will fall short.
2. The Criteria
12 criteria: 10 must-meet and 2 should-meet. A must-meet that is unsatisfied produces a RECYCLE rather than a condition; a should-meet that is short produces a condition.
| Ref | Domain | Criterion | Kind | Target at Gate 5 | Position at 15 Oct 2026 | Assessable? |
|---|---|---|---|---|---|---|
| G5-01 | Clinical | Both pivotal trials complete with database lock | Must | Locked, with the clinical study reports signed | Enrolment in progress | Not yet |
| G5-02 | Clinical | Primary endpoint meets the minimum acceptable threshold | Must | ≥10% mean weight reduction at week 68 | Blinded | Not yet |
| G5-03 | Clinical | Safety profile supports the sought label | Must | No unexpected serious risk; DMC has not recommended stopping | No DMC concern raised | Partly |
| G5-04 | CMC | Registration batches manufactured and released | Must | 3 of 3 | 2 of 3 | Assessable |
| G5-05 | CMC | Analytical methods transferred and re-qualified | Must | 3 of 3 | 2 of 3 | Assessable |
| G5-06 | CMC | Process validation (PPQ) complete | Must | 3 consecutive conforming batches | Scheduled Q1–Q2 2027 | Not yet |
| G5-07 | CMC | Stability data sufficient to support the proposed shelf life | Must | 24 months | 18 months elapsed | Partly |
| G5-08 | Regulatory | Submission content agreed with the agency | Must | Pre-NDA meeting held, minutes received | Scheduled Aug 2028 | Not yet |
| G5-09 | Regulatory | No major amendment outstanding | Must | Zero | Zero | Assessable |
| G5-10 | Quality | Inspection readiness at all named manufacturing sites | Should | Two internal audits closed | First audit scheduled 2027 | Not yet |
| G5-11 | Commercial | Market access evidence package complete | Should | Payer dossier and advisory board program complete | 2 of 5 boards unscheduled | Assessable |
| G5-12 | Program | Budget within authorized ceiling | Must | Within ceiling | Within ceiling | Assessable |
3 criteria are assessable today and currently short — the third registration batch, the third analytical method, and the market access evidence program. All three are already tracked elsewhere: two as CMC readiness in the status report, and one as gate condition GC-03 from Gate 4.
3. The Column That Matters: Assessable Yet?
Two years out, most of these criteria cannot be assessed at all, and saying so is more useful than scoring them anyway.
| Assessability | Count | What it means |
|---|---|---|
| Assessable | 5 | A real position exists today. These are the only criteria where an early warning is possible, and all three current shortfalls are here. |
| Partly | 2 | Assessable in the negative, or accruing on a clock. Stability at 18 of 24 months is a real position; the absence of a DMC stop is real evidence. |
| Not yet | 5 | ⚠ Structurally unassessable. Not neglected — there is nothing to see. |
G5-02 cannot be assessed early without destroying the thing being assessed.
The primary endpoint criterion asks whether mean weight reduction clears the minimum acceptable threshold. Assessing that before the database is locked would require breaking the blind, and a blind broken early does not produce an early answer — it produces a trial that can no longer answer the question at all.
This is the sharpest example of a general point. A readiness tracker that reports a status against every criterion, every month, has quietly invented positions for the ones it cannot see — and those invented positions are indistinguishable, in a report, from the real ones.
The primary endpoint criterion asks whether mean weight reduction clears the minimum acceptable threshold. Assessing that before the database is locked would require breaking the blind, and a blind broken early does not produce an early answer — it produces a trial that can no longer answer the question at all.
This is the sharpest example of a general point. A readiness tracker that reports a status against every criterion, every month, has quietly invented positions for the ones it cannot see — and those invented positions are indistinguishable, in a report, from the real ones.
The honest shape of a readiness assessment at this distance is therefore mostly blank, and the fact base asserts it: if more criteria were assessable than not, two years before the gate, the build fails. That check exists because the temptation to fill the column is strong and the resulting document looks better than it is.
4. What Each Assessment Rests On
| Ref | What the assessment actually rests on |
|---|---|
| G5-01 | Cannot be assessed until the last participant completes. The only meaningful proxy is the enrolment curve, which is a schedule measure rather than a readiness one. |
| G5-02 | ⚠ Structurally unassessable before unblinding. Assessing it early would require breaking the blind, which would destroy the thing being assessed. |
| G5-03 | Assessable continuously in the negative — the absence of a DMC stop is real evidence, and it is the only safety evidence the sponsor is entitled to see. |
| G5-04 | Third batch is scheduled within thirty days. Assessable today, currently short, and the item every other CMC criterion waits on. |
| G5-05 | ⚠ Already failed once. The re-run of method 3 is the single most watched item in the CMC track. |
| G5-06 | Cannot start until registration batches exist, so its readiness is entirely a function of G5-04. |
| G5-07 | Accrues at one month per month and cannot be accelerated by any action. The only risk is an out-of-trend result, which would be visible when it happened. |
| G5-08 | The meeting is two years out. What is assessable now is whether the questions that will be asked at it are being accumulated. |
| G5-09 | Assessable continuously and currently met. Zero major amendments is a credit that belongs upstream, to the people who agreed the content early. |
| G5-10 | A should-meet: a shortfall here produces a condition rather than a stop. |
| G5-11 | ⚠ Gate condition GC-03 from Gate 4. Assessable today, and currently behind. |
| G5-12 | Assessable continuously. The forecast internal-labor overrun is absorbed within contingency, so the ceiling itself is not in question at this date. |
Three of these read as dependencies rather than as criteria, and that is the useful
observation.
Process validation cannot start until registration batches exist, so G5-06's readiness is entirely a function of G5-04. Stability accrues at one month per month and cannot be accelerated by any action. And the clinical criteria wait on the last participant, which waits on enrolment.
A readiness assessment that lists twelve independent criteria has misrepresented the structure. Several of these are one chain, and the earliest link in it — the third registration batch — is the item worth watching, because everything behind it moves when it moves.
Process validation cannot start until registration batches exist, so G5-06's readiness is entirely a function of G5-04. Stability accrues at one month per month and cannot be accelerated by any action. And the clinical criteria wait on the last participant, which waits on enrolment.
A readiness assessment that lists twelve independent criteria has misrepresented the structure. Several of these are one chain, and the earliest link in it — the third registration batch — is the item worth watching, because everything behind it moves when it moves.
5. What Happens If a Criterion Is Still Short
The tracker has no authority, so its only job is to make the gate's options visible early. Those options are fixed by the gate framework and they are narrower than people expect.
| Situation at the gate | Outcome available | Why |
|---|---|---|
| A must-meet is unsatisfied | RECYCLE — not a condition | ⚠ Must-meets are not outvotable and cannot be conditioned. An unsatisfied must-meet means the program returns to the stage and comes back; it does not proceed with a promise. |
| A should-meet is short | GO WITH CONDITIONS | A condition with a named owner, a due date, and a verification method fixed at issue. This is what happened at Gate 4 with market access evidence. |
| A criterion is still unassessable | Neither — the gate cannot be held | ⚠ If the database is not locked, there is no submission to authorize. The gate moves; it does not proceed on a forecast. |
| Everything is met | GO | Which, on the evidence of the four gates behind us, is the least likely outcome. |
The most useful thing this tracker can produce is not a green column. It is twelve months of
warning that a must-meet is going to be short.
Because the response to a must-meet shortfall is RECYCLE, and RECYCLE at Gate 5 means the submission does not go in. Every month of early warning is a month in which the shortfall might be closed instead — and for the three currently short, the closing actions are all operational and all already running.
A shortfall found at the gate is a decision about the program. The same shortfall found two years out is a piece of work. That conversion, from decision back into work, is the entire return on maintaining this document.
Because the response to a must-meet shortfall is RECYCLE, and RECYCLE at Gate 5 means the submission does not go in. Every month of early warning is a month in which the shortfall might be closed instead — and for the three currently short, the closing actions are all operational and all already running.
A shortfall found at the gate is a decision about the program. The same shortfall found two years out is a piece of work. That conversion, from decision back into work, is the entire return on maintaining this document.
Which is also why the tracker never edits its criteria. If Gate 4's criteria turn out to have been wrong — too strict, or measuring the wrong thing — that is a legitimate matter for the Committee to consider at the gate, on the record, as a decision. It is not fixed by quietly amending the standard in the document that tracks it.
6. How This Is Maintained
| Rule | Why |
|---|---|
| Updated monthly, tabled quarterly | Alongside the status report. A readiness tracker reviewed only near the gate is a retrospective document. |
| A criterion moves to assessable only when evidence exists | Not when somebody has an opinion about how it will turn out. The distinction between position and forecast is the whole value of the artifact. |
| Shortfalls are reported as shortfalls, not as plans | ⚠ “2 of 3, third scheduled” is a shortfall with a plan attached. “On track” is a plan presented as a position. |
| The criteria are never edited | If Gate 4's criteria turn out to be wrong, that is a matter for the Committee at the gate, recorded as such. It is not fixed by amending the tracker. |
The purpose of this document is to make Gate 5 boring.
A gate where the Committee learns something material for the first time has already failed, whatever it decides — because the decision is then being taken on a briefing rather than on two years of accumulated evidence, and there is no time to act on anything it reveals.
Every shortfall in this tracker at the status date — the third batch, the third method, the access evidence — will have been visible for twenty-four months by the time the gate is held. If they are still short then, the Committee will be deciding about a known problem rather than discovering one, and that is the difference between governance and theatre.
A gate where the Committee learns something material for the first time has already failed, whatever it decides — because the decision is then being taken on a briefing rather than on two years of accumulated evidence, and there is no time to act on anything it reveals.
Every shortfall in this tracker at the status date — the third batch, the third method, the access evidence — will have been visible for twenty-four months by the time the gate is held. If they are still short then, the Committee will be deciding about a known problem rather than discovering one, and that is the difference between governance and theatre.
The counterpart risk is worth naming. A tracker that is green for two years and then reports a failure at the gate has not been maintained; it has been filled in. That is why the assessability column is as important as the position column, and why the criteria that cannot yet be assessed are shown as blank rather than as green.