12
Subsidiary plans
3
Baselines
6
Change authority tiers
5
Conflicts resolved
Contents
1. Integration, Not Summary
There are 12 subsidiary plans in this program, each owned by somebody with the expertise to own it. This document does not summarize them. If it did, it would be a table of contents with a cover page, it would go stale the first time one of them changed, and a reader would have two versions of the same fact with no way to tell which was current.
| This plan holds | Because |
|---|---|
| The three baselines | A baseline is not a plan. It is the thing variance is measured against, and it has to live somewhere that is not owned by any single function. |
| The change authority | Every subsidiary plan can be changed. Who may change which, at what threshold, and by what instrument is an integration question by definition. |
| The resolution of conflicts BETWEEN plans | Two well-written plans can give contradictory guidance, and usually do. Nothing inside either one can settle it. That is the actual work of integration, and §5 is the only place in the suite where it is written down. |
| What was deliberately tailored out | A methodology applied without subtraction is a methodology nobody thought about. |
The test for whether a consolidated plan is doing any work: delete it, and see what becomes
unanswerable.
Delete this one and the program can still say what its scope is, when things happen, what they cost and who does them — all of that lives elsewhere and is linked below. What it can no longer say is who is allowed to change any of it, and what happens when the risk plan and the budget disagree. Those two questions are the reason this document exists.
Delete this one and the program can still say what its scope is, when things happen, what they cost and who does them — all of that lives elsewhere and is linked below. What it can no longer say is who is allowed to change any of it, and what happens when the risk plan and the budget disagree. Those two questions are the reason this document exists.
2. The Subsidiary Plans
Each is linked rather than restated. The two columns that matter are the last two: what each plan decides, and — more usefully — what it explicitly does not.
| Area | Plan | Owner | What it decides | What it explicitly does not decide |
|---|---|---|---|---|
| Scope | Work Breakdown Structure | C. Tyrrell | What is in the program, decomposed to work packages under the 100% rule | How any of it is done — a WBS node is a deliverable, not an activity |
| Schedule | Integrated Development Schedule | C. Tyrrell | Sequence, dependency and the critical path across clinical, CMC and regulatory | Effort or staffing — it is a logic network, deliberately not resource-levelled |
| Milestones | Milestone & Deliverable Plan | C. Tyrrell | Which commitments are dated, who owns each, and what evidence closes it | Whether the program is on course — milestones are lagging by construction |
| Cost | Program Budget | M. Thornbury | Ceiling, tranche release, cost pools and the contingency mechanism | Whether money is well spent — that is a gate question, not a budget question |
| Resource | Resource Plan | C. Tyrrell | Headcount, hours and concurrent FTE by function and stage | Who specifically does the work — function heads own that, and should |
| Quality | GxP Compliance Plan | Dr. I. Solberg | The compliance framework, audit program and CAPA process | Product quality decisions — those sit with Quality Assurance independently |
| Risk | RAID Log | C. Tyrrell | Risks, assumptions, issues and dependencies with owners and responses | Risk appetite — that is set at the gate, not in the register |
| Procurement | CRO Oversight Plan | G. Petrossian | How transferred obligations are specified, monitored and audited | The vendor's own plan — the program contracts deliverables, not activities |
| Governance | Governance Model | C. Tyrrell | Who decides what, the gate mechanism, and the escalation path | Anything the DMC decides — four matters sit outside Committee authority |
| Safety | Safety Reporting Plan | Dr. N. Halloran | Expedited reporting obligations and the two regulatory clocks | Clinical safety judgment — that is the DMC's and the medical monitor's |
| Data | Data Management Plan | Dr. F. Achterberg | Capture, cleaning, coding, reconciliation and the lock sequence | The analysis itself — that is the SAP, approved before the lock |
| Regulatory | Regulatory Strategy | Dr. P. Raghunathan | Pathway, agency interaction strategy and submission content | Agency decisions — the sponsor controls readiness, never outcome |
The right-hand column is the one to read.
Most integration failures are not caused by a gap in the plans. They are caused by two plans both believing they own the same decision, or by a decision that every plan assumes another one covers. Writing down what each plan does not decide is cheaper than discovering the overlap during an audit, and it is the only part of this table that could not be reconstructed from the plans themselves.
Most integration failures are not caused by a gap in the plans. They are caused by two plans both believing they own the same decision, or by a decision that every plan assumes another one covers. Writing down what each plan does not decide is cheaper than discovering the overlap during an audit, and it is the only part of this table that could not be reconstructed from the plans themselves.
3. The Three Baselines
Scope, schedule and cost. Everything else in this program is managed; only these three are baselined, and the distinction is not decorative.
| Baseline | What it fixes | Set at | Movable by | Change since Gate 4 |
|---|---|---|---|---|
| Scope | The 64 work packages of the WBS | Gate 4 | Development Committee | 2 changes — CR-01 and CR-02, both discrete packages so their cost stays visible |
| Schedule | The gate dates and the submission date | Gate 4 | Development Committee | No change. The enrolment shortfall is carried as exposure, not re-baselined |
| Cost | $217,000,000 base plus $26,040,000 contingency | Gate 4 | Board (ceiling) / Committee (release) | No change to the ceiling; $9,050,000 of contingency drawn |
A baseline you can quietly move is not a baseline. It is a forecast with a formal-sounding
name.
The schedule baseline is the clearest case here. At the status date the program is 76 participants behind the enrolment curve and the submission date has not moved. There is a permanent temptation in that situation to re-baseline the schedule so that variance reads as zero — and it would be defensible, because the original curve was an estimate and the new one is better informed.
The program has not done it, and the reason is that variance against a baseline is the only early warning the governance system gets. Re-baselining to the current forecast converts a visible problem into an invisible one and resets the clock on everyone's memory of the commitment. The shortfall is carried as an exposure in the RAID log instead, where it stays uncomfortable.
The schedule baseline is the clearest case here. At the status date the program is 76 participants behind the enrolment curve and the submission date has not moved. There is a permanent temptation in that situation to re-baseline the schedule so that variance reads as zero — and it would be defensible, because the original curve was an estimate and the new one is better informed.
The program has not done it, and the reason is that variance against a baseline is the only early warning the governance system gets. Re-baselining to the current forecast converts a visible problem into an invisible one and resets the clock on everyone's memory of the commitment. The shortfall is carried as an exposure in the RAID log instead, where it stays uncomfortable.
Baselines are set at Gate 4, not at Gate 0. Before Phase 2 survives, the program is a funded option rather than a commitment, and baselining an option produces variance reporting against a number nobody intended to be held to. This is the same reasoning that anchors the charter at Gate 4.
4. Change Authority
Six tiers. The tiering is the substance — without it, “change control” means everything reaches the Development Committee, which means the Committee stops reading and nothing is controlled at all.
| Change | Authority | Instrument |
|---|---|---|
| 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 |
Two rows here are deliberately asymmetric.
The ceiling is not a Committee matter at any size. Not a large change, not any change. The Committee releases money against a ceiling the Board authorized; a body that can raise its own ceiling is not operating under one.
Safety requires no approval at all. An action taken to protect a participant is taken and then reported, never proposed and then approved. Every escalation path in this program is a commercial path, and the safety route bypasses all of it — which is why an urgent safety measure may be implemented before it is submitted.
The ceiling is not a Committee matter at any size. Not a large change, not any change. The Committee releases money against a ceiling the Board authorized; a body that can raise its own ceiling is not operating under one.
Safety requires no approval at all. An action taken to protect a participant is taken and then reported, never proposed and then approved. Every escalation path in this program is a commercial path, and the safety route bypasses all of it — which is why an urgent safety measure may be implemented before it is submitted.
Two change requests have been raised against the Gate 4 baseline in the life of the program: CR-01 and CR-02. Both were held as discrete work packages in the WBS rather than absorbed into existing ones, so that their cost stays visible for the rest of the program instead of disappearing into a larger number that nobody can decompose two years later.
5. Where the Plans Conflict
This is the section that justifies the document. Each row is a real contradiction between two well-written subsidiary plans, neither of which is wrong, and neither of which can resolve it from the inside.
| Where they conflict | The conflict | How this plan resolves it |
|---|---|---|
| Risk response vs cost control | The RAID log's response to the enrolment risk is more sites; the budget has no uncommitted line for them. | Reserve sites are funded from contingency, not from the base, and every draw records the rejected alternative. The risk register does not get its own budget. |
| Schedule float vs resource levelling | The schedule is a logic network with float in CMC; the resource plan shows Stage 3 at 7.2 FTE, which reads as slack. | The two are deliberately NOT reconciled. Levelling a logic network hides which constraint binds. Stage 3 is long because participants are being treated, not because it is under-loaded. |
| Quality independence vs program accountability | The Program Director is accountable for delivery but cannot direct Quality Assurance, whose Chief Quality Officer reports to the CEO. | Unresolved by design. The independence line is the control. The PD escalates to the Committee; the PD does not instruct. |
| Vendor oversight vs contracted deliverables | The oversight plan requires visibility the contract does not grant — the CRO owes deliverables and service levels, not a staffing plan. | Oversight is exercised through audit rights and performance data, never through directing vendor staff. 21 CFR 312.52 governs what was transferred; silence means retained. |
| Data lock sequence vs filing date | Compressing the SAP and the database lock would recover schedule. | Refused as a constraint rather than a commitment. The SAP is signed before the lock and the lock precedes unblinding, or the result is unfalsifiable. |
Notice that one of these is resolved by leaving it unresolved.
The Program Director is accountable for $243.0M of delivery and cannot direct Quality Assurance, because the Chief Quality Officer reports to the Chief Executive. That is not a gap in the organization design; it is the organization design. The independence line exists precisely so that the person under delivery pressure cannot instruct the person whose job is to stop delivery when it is non-compliant.
A consolidated plan that “fixed” this by granting the PD authority over Quality would be a better-integrated plan and a worse-governed program. Some conflicts are controls.
The Program Director is accountable for $243.0M of delivery and cannot direct Quality Assurance, because the Chief Quality Officer reports to the Chief Executive. That is not a gap in the organization design; it is the organization design. The independence line exists precisely so that the person under delivery pressure cannot instruct the person whose job is to stop delivery when it is non-compliant.
A consolidated plan that “fixed” this by granting the PD authority over Quality would be a better-integrated plan and a worse-governed program. Some conflicts are controls.
The other four resolve toward the same principle in different clothing: the thing that cannot be recovered later wins. Contingency protects the ceiling, the logic network protects the constraint that actually binds, the contract protects what was legally transferred, and the lock sequence protects the falsifiability of the trial. In every case the alternative would have bought schedule or money that could be recovered, at the cost of something that could not.
6. What Was Tailored Out
A methodology applied without subtraction is a methodology nobody thought about. Four standard elements were deliberately not used, and recording the omissions is part of the plan.
| Not used | Why not |
|---|---|
| Earned value management as the primary control | 83% of the program is external, and the dominant cost driver is participants dosed rather than time elapsed. Earned value on a portfolio of per-participant contracts measures how many people have been enrolled, which the enrolment dashboard already reports directly and more usefully. A performance measurement baseline is maintained for the internal labor envelope only. |
| A single integrated master schedule including vendor activities | The program contracts deliverables and service levels from Meridian, not activities. A sponsor schedule containing the CRO's internal tasks would assert an oversight the contract does not grant and the sponsor cannot verify. |
| Quantitative risk analysis (Monte Carlo on the schedule) | The critical path is 29 of 39 activities and the dominant uncertainties are binary rather than distributional — a pivotal reads out or it does not. Simulating a near-serial network with binary risks produces a confident-looking distribution around an assumption nobody tested. |
| A separate stakeholder engagement plan | Folded into the governance model. The stakeholders who matter to this program either sit on the Committee, hold a contract, or are a regulator — and each of those relationships is already governed by an instrument stronger than an engagement plan. |
Every one of these is defensible and every one is a judgment that could have gone the other
way.
That is why they are written down rather than simply not done. An auditor, a new program director, or a hiring manager reading this suite can disagree with any of the four — but they can see that the choice was made deliberately, on a stated rationale, rather than by a methodology being applied only as far as somebody's patience lasted.
That is why they are written down rather than simply not done. An auditor, a new program director, or a hiring manager reading this suite can disagree with any of the four — but they can see that the choice was made deliberately, on a stated rationale, rather than by a methodology being applied only as far as somebody's patience lasted.