83%
Delivered externally
4/4
Make / buy
8
Agreements
15%
Weight given to price
Contents
1. A Program That Is Mostly Bought
83.4% of this program is delivered by organizations that do not report to it. That single figure determines the shape of everything in this document, and it makes procurement a program-management discipline here rather than a support function.
| Value | What follows from it | |
|---|---|---|
| External | $180,992,000 — 83.4% | The program is directed through contracts, service levels and audit rights rather than through line management. |
| Internal | $36,008,000 — 16.6% | Buys the capability to specify, contract, oversee, verify and accept. It does not buy execution. |
| Largest single contract | $93,900,000 — Meridian | 43% of the base program in one counterparty relationship. |
| Agreements in the register | 8, 7 fully executed | The outstanding one is not a document problem — it is 29 site contracts still in negotiation, which is the program's central operational constraint. |
This is not a document about buying things. It is a document about which parts of a program you
are willing to not do yourself.
Every line in the make-versus-buy analysis is a decision about capability rather than cost. The question is never “can somebody do this cheaper” — somebody almost always can. The question is whether the company can afford to lose the capability, and whether the thing being bought can be specified well enough that buying it is meaningful.
A capability you buy is a capability you stop having. That is acceptable for site monitoring, which the company would not keep employed between programs, and unacceptable for the argument in the submission, which the company has to defend.
Every line in the make-versus-buy analysis is a decision about capability rather than cost. The question is never “can somebody do this cheaper” — somebody almost always can. The question is whether the company can afford to lose the capability, and whether the thing being bought can be specified well enough that buying it is meaningful.
A capability you buy is a capability you stop having. That is acceptable for site monitoring, which the company would not keep employed between programs, and unacceptable for the argument in the submission, which the company has to defend.
2. Make Versus Buy
4 kept, 4 bought. The pattern is consistent, and it is not about cost.
| Capability | Decision | Why | What it costs us |
|---|---|---|---|
| Protocol design and medical monitoring | MAKE | The scientific argument in the submission is the sponsor's to make and defend. Outsourcing it would mean defending someone else's reasoning at an advisory committee. | 12 people in Clinical Development, at the highest blended rate in the program. |
| Trial execution — sites, monitoring, data capture | BUY | Requires a standing global site network and monitoring capacity that a company running one pivotal program cannot keep employed between trials. | The single largest contract in the program. |
| Drug substance and drug product manufacture | BUY | A commercial-scale peptide facility is a nine-figure capital decision. The program does not have the volume to justify one, and would not have it in time. | Single-source dependency on Aldergate for the life of the program. |
| Bioanalytical and safety laboratory work | BUY | Results must be comparable across 260 sites, which means one laboratory, one method, one set of reference ranges — not 260 local labs. | Central lab contracted; site-local labs used only for safety screening. |
| Regulatory strategy and agency interaction | MAKE | The relationship with the review division is an asset the company keeps. Commitments made in a meeting bind the sponsor, not the advisor. | 10 people carrying every commitment made to the agency. |
| Submission publishing and eCTD assembly | BUY | Specialist tooling and a skill used intensively for four months every few years. | Contracted to a publishing vendor under the regulatory counsel agreement. |
| Pharmacovigilance case processing | MAKE | ⚠ Considered for outsourcing and declined. The expedited reporting clock has no tolerance band, and a handoff adds a step to a process measured in calendar days. | 8 people, and the on-time figure has stayed at 100%. |
| Quality assurance and GxP audit | MAKE | Independence is the control. An audit function contracted by the party it audits is not independent, whatever the contract says. | Reports to the Chief Quality Officer, not to this program. |
The rule that produced this table: buy capacity, keep judgment.
Everything in the BUY column is capacity — a site network, a manufacturing plant, a laboratory, a publishing toolchain. They are expensive to own, they sit idle between programs, and their output can be specified precisely enough to be contracted for.
Everything in the MAKE column is judgment: what the protocol should test, what to tell the agency, whether a case is expedited, whether the quality system is adequate. Judgment cannot be specified in advance, which is exactly why it cannot be contracted for — a contract is a specification, and if you could write the specification you would not need the judgment.
Everything in the BUY column is capacity — a site network, a manufacturing plant, a laboratory, a publishing toolchain. They are expensive to own, they sit idle between programs, and their output can be specified precisely enough to be contracted for.
Everything in the MAKE column is judgment: what the protocol should test, what to tell the agency, whether a case is expedited, whether the quality system is adequate. Judgment cannot be specified in advance, which is exactly why it cannot be contracted for — a contract is a specification, and if you could write the specification you would not need the judgment.
Two rows are worth reading closely. Pharmacovigilance case processing was considered for outsourcing and declined, because the expedited reporting clock has no tolerance band and a handoff adds a step to a process measured in calendar days. And quality assurance is kept because independence is the control — an audit function contracted by the party it audits is not independent, whatever the contract says.
3. Contract Models, and Who Carries What
A contract model is a risk allocation with a payment schedule attached. Choosing one is choosing which party absorbs which uncertainty.
| Contract model | Where it is used | Who carries the risk | When it is the right choice |
|---|---|---|---|
| Fixed price by work package | Meridian — the three CRO packages | Vendor carries execution risk within a defined scope; sponsor carries scope risk. | ⚠ Right where the scope is genuinely definable. It is not a way of avoiding scope management — four change orders have been approved and one is disputed. |
| Unit rate per participant | Site agreements, and the CRO rate card for change orders | Sponsor carries volume risk; vendor carries per-unit cost risk. | Right where the driver is countable and the unit of work is uniform. It is why the cost of a protocol amendment adding one visit is instantly calculable. |
| Cost plus fixed fee | Aldergate — process development only | Sponsor carries cost risk; vendor carries no incentive to be efficient. | Right only where the work cannot be specified in advance. Process development qualifies; commercial manufacture does not, and moved to fixed price at scale-up. |
| Time and materials | Thorne & Vale, and advisory boards | Sponsor carries all risk. | Right for genuinely advisory work bought in small quantities. ⚠ Wrong for anything on the critical path, and the program does not use it there. |
Fixed price does not mean the sponsor has stopped carrying risk. It means the sponsor has
concentrated its risk in scope definition.
Meridian's three work packages are fixed price, and the program has approved 4 change orders worth $3,100,000 against them, with one more disputed. That is not a failure of the contract model — it is the model working. Every one of those change orders is a place where the sponsor's scope moved, and the fixed price made the movement visible and priced.
A time-and-materials contract would have absorbed the same changes silently, and the program would have discovered the cost at the end rather than at the point of decision.
Meridian's three work packages are fixed price, and the program has approved 4 change orders worth $3,100,000 against them, with one more disputed. That is not a failure of the contract model — it is the model working. Every one of those change orders is a place where the sponsor's scope moved, and the fixed price made the movement visible and priced.
A time-and-materials contract would have absorbed the same changes silently, and the program would have discovered the cost at the end rather than at the point of decision.
The rate card annexed to the master agreement is what makes this workable. It prices the units of work in advance, so a change order is an arithmetic exercise rather than a negotiation — and negotiating a price while the trial is running is negotiating from the weaker position.
4. How Vendors Were Selected
| Criterion | Weight | How it is applied |
|---|---|---|
| Regulatory inspection history | Pass/fail gate | ⚠ Assessed BEFORE price is opened. A vendor with an open warning letter is not evaluated further, at any price, because their finding becomes the sponsor's problem at review. |
| Therapeutic-area and phase experience | 30% | Specifically: pivotal weight-management trials, not trials in general. |
| Site network in the required geographies | 25% | The constraint that decided the CRO selection. Meridian had activated sites in the target regions within the previous eighteen months. |
| Quality systems and data integrity | 20% | Audited before award, not after. Part 11 compliance demonstrated rather than asserted. |
| Price | 15% | ⚠ The smallest weighted criterion. A cheaper CRO that activates sites slowly costs more than it saves, and the loss lands on the filing date rather than on the budget. |
| Capacity and key personnel | 10% | Named individuals with committed availability, and a contractual right to approve replacements. |
Price is weighted at 15% and it is the smallest criterion on the list.
That is not indifference to cost. It is a judgment about where the losses actually land. A CRO that is ten percent cheaper and activates sites four weeks slower costs the program far more than it saves, and the loss arrives as a filing date rather than as a budget variance — where it is harder to see, harder to attribute, and impossible to recover.
The inspection-history criterion is not weighted at all, because it is a gate. A vendor with an open warning letter is not evaluated further at any price. Their finding becomes the sponsor's problem at review, and no commercial term transfers that back.
That is not indifference to cost. It is a judgment about where the losses actually land. A CRO that is ten percent cheaper and activates sites four weeks slower costs the program far more than it saves, and the loss arrives as a filing date rather than as a budget variance — where it is harder to see, harder to attribute, and impossible to recover.
The inspection-history criterion is not weighted at all, because it is a gate. A vendor with an open warning letter is not evaluated further at any price. Their finding becomes the sponsor's problem at review, and no commercial term transfers that back.
The criterion that actually decided the CRO selection was the third one: an activated site network in the target geographies within the previous eighteen months. Sites that a CRO has recently worked with activate faster than sites it has merely contracted with before, and — given what activation duration does to this program's schedule — that was worth more than any price difference on the table.
5. The Agreements Register
| Agreement | Counterparty | Model | Status | Note |
|---|---|---|---|---|
| Master Services Agreement | Meridian Clinical Research | Fixed price by work package | Executed | Covers all three work packages. Rate card annexed for change orders. |
| Transfer of Regulatory Obligations | Meridian Clinical Research | — | Executed and filed | ⚠ Filed with the IND. Obligations not listed are not transferred. |
| Quality Agreement | Aldergate Biologics | — | Executed | Separate from the supply agreement by design — quality terms do not get traded in a commercial negotiation. |
| Supply Agreement | Aldergate Biologics | Cost plus, moving to fixed price at scale-up | Executed | Single-source. Includes a technology-transfer clause the program has not exercised. |
| Central Laboratory Agreement | Kestrel Laboratories | Unit rate per sample | Executed | Method validation and reference ranges fixed at award. |
| Regulatory Counsel Engagement | Thorne & Vale LLP | Time and materials | Executed | Includes eCTD publishing as a subcontracted scope. |
| Clinical Trial Agreements | 260 investigational sites | Unit rate per participant | 231 of 260 executed | ⚠ The bottleneck. Contract execution and IRB review are the two steps in site activation the sponsor influences least. |
| Pharmacovigilance Agreement | Meridian Clinical Research | — | Executed | Defines the 24-hour site-to-sponsor obligation that makes the expedited clock achievable. |
What matters most there looks least like a contract.
The Transfer of Regulatory Obligations is filed with the IND and lists, item by item, which sponsor obligations Meridian has assumed. The governing regulation is unambiguous: any obligation not covered by the written description is deemed not to have been transferred. Silence is not a gap to be argued about later — silence means the sponsor kept it.
That form is the reason the program can answer an inspector's question about who was responsible for something, and “our CRO told us it was fine” is not an answer at inspection.
The Transfer of Regulatory Obligations is filed with the IND and lists, item by item, which sponsor obligations Meridian has assumed. The governing regulation is unambiguous: any obligation not covered by the written description is deemed not to have been transferred. Silence is not a gap to be argued about later — silence means the sponsor kept it.
That form is the reason the program can answer an inspector's question about who was responsible for something, and “our CRO told us it was fine” is not an answer at inspection.
The Quality Agreement with Aldergate is deliberately separate from the Supply Agreement, and that separation is a control rather than an administrative accident. Quality terms negotiated inside a commercial agreement get traded against price and delivery; held separately, they are signed by the quality functions on both sides and are not available as concessions.
The outstanding line is the clinical trial agreements: 231 of 260 executed. Contract execution and institutional review are the two steps in site activation the sponsor influences least and the CRO cannot compel at all, which is why activation is running 22 weeks against a planned 18.
6. Single Source, Accepted Deliberately
Aldergate manufactures both drug substance and drug product, and there is no second source. That is a named, accepted risk rather than an oversight, and the reasoning belongs in a procurement document rather than a risk register.
| Option | What it would cost | Why not |
|---|---|---|
| Second manufacturing site | Technology transfer, a second process validation, a second pre-approval inspection — and roughly two years | Does not fit the authorized ceiling, and the timeline would move the filing date. The scope baseline lists it as explicitly out of scope. |
| Dual sourcing of drug substance only | Partial protection, most of the cost | Comparability between two substance sources is itself a regulatory exercise, and an unresolved comparability question at filing is worse than the single-source risk. |
| Technology-transfer clause, unexercised | Contract term only | ⚠ What the program actually did. The right to transfer is secured; the transfer is not performed. It converts an absolute dependency into a slow remedy. |
The method transfer that failed in March happened at a single-source supplier, which is exactly
when a single-source dependency stops being theoretical.
Two of three analytical methods have since been re-qualified and the third has been re-run. Had it not recovered, the program's options were to wait or to invoke a technology transfer that takes about two years — on a filing date that does not move.
Accepting a risk is not the same as ignoring it, and the difference is whether a remedy was secured while there was still leverage to secure one. The clause was negotiated at award, when the program had alternatives. It could not have been negotiated in March.
Two of three analytical methods have since been re-qualified and the third has been re-run. Had it not recovered, the program's options were to wait or to invoke a technology transfer that takes about two years — on a filing date that does not move.
Accepting a risk is not the same as ignoring it, and the difference is whether a remedy was secured while there was still leverage to secure one. The clause was negotiated at award, when the program had alternatives. It could not have been negotiated in March.
7. Obligations That Stay With the Sponsor
| Not purchasable | Why |
|---|---|
| Accountability | Execution can be contracted; the obligation cannot. The sponsor answers at inspection for work performed by a vendor, and a well-drafted contract changes who pays, not who is answerable. |
| Oversight of what you bought | ⚠ Retaining an obligation and never exercising it is indistinguishable, at inspection, from having delegated it improperly. The oversight capacity has to be staffed — which is what 16.6% of the budget is for. |
| A site's enthusiasm | The 260 investigational sites are independent institutions with their own priorities and their own competing trials. They are neither employees nor vendors, and no contractual term makes one enrol faster. |
| Speed at the agency | No advisor, at any rate, shortens a review. A sponsor can slow one down, which is a different service and one nobody sells. |
This program learned one of those the hard way.
Sites sit in a category that procurement language handles badly. They sign an agreement and receive payment, which makes them look like vendors — but they are independent institutions running their own ethics review on their own calendar, and the trial is one of several competing for the same investigator's attention.
The enrolment shortfall was never solvable by contract management, and the months spent confirming that were months the recovery plan did not have.
Sites sit in a category that procurement language handles badly. They sign an agreement and receive payment, which makes them look like vendors — but they are independent institutions running their own ethics review on their own calendar, and the trial is one of several competing for the same investigator's attention.
The enrolment shortfall was never solvable by contract management, and the months spent confirming that were months the recovery plan did not have.
Which is the closing argument for why procurement in a program like this is not a purchasing activity. 83.4% of the work sits outside the organization, and the program's entire capability to influence it was fixed at the moment the agreements were signed. Every lever available in 2026 was negotiated in 2024, by people who were guessing about the problems the program would eventually have.