On trial geography

Where a First-in-Human Trial Begins Is Also a Decision That Is Not Yours

Published 2026-09-07 · Regulatory references are identified below as statute, scheme documentation, or proposal. Guidance is non-binding and does not establish legally enforceable responsibilities. Not legal or regulatory advice.

The argument that early-phase clinical research is leaving the United States is now mainstream. The Institute for Progress has made the case and mapped the mechanism — a notification pathway instead of a full submission, ethics-committee review instead of agency review, and Phase I trials formally exempt from the commercial-scale manufacturing standard that can multiply early costs.

In June 2026, HHS launched Operation TrialBlazer to try to close the gap. Its FDA component is a pilot FDA has proposed and not yet opened — and it is not a notification route. What it proposes is rolling submission, and a network of qualified research institutions that would advise sponsors before the agency sees the package: the IND gate and its review clock stay where they are. Handed the speed argument, the agency's answer so far has been to start its involvement earlier, not to step back from it.

The debate is real, the data behind it are real, and a sponsor weighing where to run a first-in-human oncology trial should take it seriously. But it is conducted almost entirely in one currency: how fast and how cheaply you can start. That is an operational question, and it has an operational answer. It is not the only question the geography decision asks.

Because where a trial begins is not only a matter of speed. It changes what the trial is for.

The frame, briefly

An earlier essay argued that a first-in-human oncology trial is worth exactly the decisions it lets three people resolve: the patient who said yes to it, the person spending the next development dollar, and the person spending the next investment dollar. When all three reach the same answer from the same data, the program is honest. When any of them would reach a different answer, the divergence is the finding.

The geography decision moves all three chairs at once — and not in the same direction. That is the part the speed debate does not price. Alison Schram, an academic oncologist at Memorial Sloan Kettering Cancer Center, framed the systemic version of this to the New York Times: “America is still leading in terms of basic science. But there must be a dialogue between what is discovered in the lab and what is tested in patients.” When a program's geography breaks that dialogue, the cost is not only slower enrollment — it is a data package built to answer the wrong room's questions.

The patient chair

A trial that begins under a notification pathway and a trial that begins under a full investigational-new-drug submission are not the same consent.

Under Australia's Clinical Trial Notification scheme, the sponsor notifies the regulator and the study is reviewed by a human research ethics committee; the regulator does not review the trial's data package before it begins. Under a US IND, the agency does. Australia runs this at scale: 18,453 trials registered as recruiting there between 2006 and 2020. But the comparison the debate assumes is not easy to find: neither the Australian registry analysis nor the TGA's own materials measure whether the faster route produces different safety outcomes, and a search for a published comparison of the two routes found none. That is not evidence of harm, nor proof that no such analysis exists — it is a gap where the settling evidence should be. A program choosing its geography on safety grounds, in either direction, is choosing on something it has not seen measured.

But a patient's yes is a wager on more than the molecule. It is a wager on the system that vetted the protocol before anyone was asked to enroll. Whether that vetting sits with an ethics committee close to the site or with a regulator holding a national standard is not a difference in speed. It is a difference in what the patient is trusting. A sponsor is entitled to choose either — many choose the faster route for good reasons — but a sponsor who has not noticed that the choice reshapes the consent has decided something without fully deciding it. The question worth asking out loud is the one the trial asks the patient to answer first: is this the right way to resolve what this experiment needs to resolve, and who stood behind that judgment before the patient did?

The developer's dollar

The Institute for Progress makes a sharp operational point, and it is correct: Phase I trials are formally exempt in Australia from the full manufacturing standard that applies to commercial product, and in the US most sponsors end up building to something close to commercial scale anyway. In statute, the US standard is the current-good-manufacturing-practice requirement at section 501(a)(2)(B) of the Food, Drug, and Cosmetic Act — the provision reform advocates want amended to carve out early-phase trials. Starting outside the US can buy six to twelve months of runway and materially cheaper chemistry and manufacturing.

That runway is worth real money. As the New York Times put it, for a biotechnology company spending $2 million per month, the gap between the two routes “can be the difference between surviving and folding.” But the earlier essay already named what the faster start spends. The US pre-IND channel that the notification route skips is the one interaction designed to surface, before the trial runs, the objections that are expensive to discover afterward — the population that cannot be interpreted, the comparator that has stopped being standard of care, the endpoint that will not carry the claim. The developer's dollar buys speed in one currency and forfeits alignment in another. Neither choice is wrong. But the trade is being made whether or not anyone at the table has priced it, and the price is not in the manufacturing line item. It is in the design questions that never got asked in a room where the agency was present.

The investor's tranche

Encouraging human data unlocks capital. The Institute for Progress frames this as survival for small biotech, and that is not an overstatement: a first efficacy signal, or even a clean safety and exposure profile, is often what moves the next tranche.

But an investor reading an asset that started under a notification pathway is reading a data package assembled against a different adjudicator's prior questions. If the development plan points at a US registration, the evidence generated to satisfy an ethics committee and a leaner regulator is not automatically the evidence a US review division will ask for — and the gap does not show up as a failure in the data room. It shows up later, as a question the package cannot answer. This is where the three chairs can quietly diverge: the developer optimized for speed to first dose, the investor is underwriting a path to approval, and the same data support two different answers because they were built for two different rooms. The divergence was not created at the financing. It was seeded at the geography decision, months or years earlier, when nobody wrote down which adjudicator the evidence was for.

One dataset, the wrong room

There is a public illustration of exactly this, and it does not involve trial geography in the notification-versus-IND sense — which is why it is useful, because it isolates the mechanism.

ORIENT-11 was a randomized, double-blind, placebo-controlled phase 3 trial of sintilimab with chemotherapy in non-squamous non-small-cell lung cancer. It enrolled 397 patients, all in China, and it met its primary endpoint decisively. China's regulator approved the drug on that basis. In February 2022, FDA's Oncologic Drugs Advisory Committee took up the US application and voted 14 to 1 that additional trials should be required to show the data applied to US patients and US practice.

Nothing about the dataset changed between those two outcomes. What differed was the question each regulator asked of it. The trial was single-country where US regulation asks foreign data to show applicability to the US population; the comparator was chemotherapy alone, which had stopped being the US standard of care; the endpoint was progression-free survival where recent US approvals in that setting had leaned on overall survival. All three were addressable at the design stage, when the cost would have been a protocol amendment rather than a second trial. They were never raised in time because the trial had been designed without the room that would have raised them. That the same evidence is adjudicated differently by different agencies is a separate argument from this one, and it compounds it.

That is the geography decision's teeth. Not that a trial run in the wrong place fails — it may succeed completely, for the room it was built for. The teeth are that the room a trial is built for is chosen early, often by default, and the cost of the choice arrives at the next adjudicator's door, long after the choice can be unmade.

What this changes for a program

The reform debate treats geography as logistics: where is it fastest and cheapest to start. That question has a real answer and a sponsor should get it. But it is the second question, not the first. The first is evidentiary: whose prior questions is this trial's data being built to answer, and are those the questions that decide the asset's value?

Ask it as three, one per chair:

• The patient chair. Who stood behind the protocol before the patient was asked to enroll, and is that the standard this program wants its consent to rest on?
• The developer's dollar. What alignment does the faster start forfeit, and is the uncertainty it leaves unresolved the uncertainty that most drives this program's value?
• The investor's tranche. Which adjudicator's questions is this data package built for, and is that the same adjudicator whose approval the financing assumes?

When the three answers agree, the geography choice was made deliberately and the program can defend it. When they diverge, the divergence is the finding — and it is far cheaper to find now, on a whiteboard, than later, in a room you cannot re-enter.

---

The mechanics of a fast start are worth having. The judgment — whether this program's geography serves the decisions its next dollar depends on — is the substance of the FIH Diligence Read: ten business days, fixed scope, a gate-by-gate account of what the data support and for whose question, and a board-ready synthesis that keeps fact, inference, and uncertainty separate. Where the evidence does not support a recommendation, the Read says so. If the decision in front of you is an FDA interaction rather than a capital one, that is the Calibration Sprint's terrain — same discipline, different door.

Sources

• R. Teslo and A. Kroetsch, for the Institute for Progress — Why Biotech Companies Go to Australia to Test New Drugs — cited as argument, and for the CTN and manufacturing mechanism.
• HHS — Operation TrialBlazer: HHS Roadmap to Maintaining U.S. Leadership in Early Clinical Research and Development, launched 22 June 2026. The Expedited IND Pilot and its Qualified Research Institutions network were published as a request for information in June 2026, comment period extended July 2026; pilot not launched as of publication.
• FD&C Act §501(a)(2)(B), codified 21 U.S.C. §351(a)(2)(B) — the current-good-manufacturing-practice adulteration provision.
• TGA — Clinical Trial Notification (CTN) Scheme.
• The changing landscape of clinical trials in Australia — Medical Journal of Australia 2023;219(5):192, doi:10.5694/mja2.52059 (ANZCTR and ClinicalTrials.gov). Source for the Australian trial-volume figure: 18,453 trials registered as recruiting in Australia 2006–2020. It reports no comparative safety analysis between regulatory routes, and is not cited for one.
• A. Harding, Make US INDs Great Again, LifeSciVC / Atlas Venture, 2026 — the coupling argument and CTN tradeoffs; cited as argument, not regulatory authority.
• R. Teslo, Doctors Who Discover Miracle Cures Worry About Medicine's Future, The New York Times, 4 September 2026 (print 6 September 2026, Section SR, p. 6) — cited for quoted remarks by named clinicians, not as regulatory authority. Its statement that Australia has run these trials safely for years is not relied on here; see the safety paragraph above.
• FDA Oncologic Drugs Advisory Committee, 10 February 2022, BLA 761222 (sintilimab) — final summary minutes; vote 14 yes / 1 no / 0 abstain.
• Y. Yang et al., Efficacy and Safety of Sintilimab Plus Pemetrexed and Platinum as First-Line Treatment for Locally Advanced or Metastatic Nonsquamous NSCLC (ORIENT-11), J Thorac Oncol 2020;15(10):1636–1646, doi:10.1016/j.jtho.2020.07.014 — 397 patients randomised; conducted in 47 hospitals in the People's Republic of China. Registry: ClinicalTrials.gov NCT03607539.

← Back to Writing  ·  What a first-in-human trial is actually for →  ·  The decision is not yours — and it is not one decision →  ·  The FIH Diligence Read →