Marco's Morning Report

Saturday, August 1, 2026

Normal day · full briefing

Good morning, Marco. Rule season closed yesterday. CMS dropped the FY 2027 inpatient final rule at 3:20 PM Friday, and buried in the quality section is a new readmission measure built on sepsis — which is to say, on your admissions. The post-acute rules landed the same week, so the whole discharge chain repriced at once. Iran is loud again after the pause. Golf tech had nothing new in 48 hours, and the movie slate is genuinely thin, so I'm not going to invent either.

Section A — Top Stories

FY 2027 inpatient final rule is out — and sepsis readmissions are now a payment measure

CMS issued the FY 2027 IPPS final rule yesterday afternoon: a net 2.3% rate update (3.2% market basket minus a 0.9% productivity cut — a bigger cut than proposed), worth about $2.9 billion more in hospital payments, plus $228 million more in DSH/uncompensated care and $779 million more in new-technology add-ons. The piece that touches your shift: CMS adopted a new measure on readmission following sepsis hospitalization into the Hospital Readmissions Reduction Program, with one extra year before it affects payment. Also finalized: every condition-specific IQR measure now includes Medicare Advantage beneficiaries; three measures removed, eight adopted, three modified; CJR-X goes mandatory nationwide Jan. 1, 2028 (three months later than proposed); and the electronic prior authorization measure becomes an optional bonus in CY 2027, mandatory in CY 2028, now extended to drugs under the medical benefit. AHA's Joanna Hiatt Kim called the update "inadequate" and the productivity cut "extremely high."

Why it matters to you: Sepsis is the most common thing you admit at 2 AM and the most common thing that bounces back in 30 days. Until now, sepsis lived in your world as a bundle-compliance problem (SEP-1) and a mortality index. It's about to become a readmissions penalty, which means the hospital's attention will shift from what happens in the first three hours to what happens in the two weeks after discharge — antibiotic completion, volume status, functional decline, the post-sepsis syndrome nobody codes for. The extra year is your runway. Adding MA beneficiaries to the condition-specific measures also quietly enlarges the denominator on your elderly panel, since a large share of your admits are MA.

Talking point: At the next group meeting, ask who owns post-sepsis follow-up today — hospitalist, PCP, or nobody. That's the answer that determines whether this measure costs the hospital money in FY 2029.

The rest of the discharge chain repriced this week too — SNF, IRF, IPF, LTCH

The post-acute final rules all landed in the same seven-day window. Skilled nursing facilities get 2.4% for FY 2027 (July 29). Inpatient psychiatric facilities get 2.3% (July 29). Inpatient rehab facilities got both a payment update and coverage changes (July 30). Long-term care hospitals get 2.3% (July 31). All effective Oct. 1.

Why it matters to you: Every one of these is a place you send someone at 6 AM when the bed board is full. The IRF coverage changes are the ones worth reading rather than skimming — coverage criteria determine whether the rehab liaison says yes to your deconditioned 84-year-old or whether that patient sits another two days waiting on a SNF bed. Payment updates in the 2.3–2.4% range against real cost growth mean post-acute partners will get more selective about who they accept, not less.

Suggested action: Ask case management to flag any change in IRF acceptance patterns starting in October. If the liaison starts declining patients you'd have placed in July, that's the rule, not the patient.

Iran: the pause is over, and the administration is signaling more

After a brief lull in late July, the U.S. resumed a "heavy wave of strikes" overnight Wednesday, and Trump indicated at a Cabinet meeting that they aren't ending soon — "We'll be hitting them very hard and, you know, at some point, they're going to say, 'We just can't take it anymore.'" The White House confirmed further weekend rounds are on the table. Spillover continues: Kuwait's army destroyed drones that entered its airspace, attributing the incursion to Iranian strikes on installations there. This is now five months into the conflict and roughly six weeks past the collapse of the June ceasefire. Al Jazeera has the fullest running account of the regional dimension.

Why it matters to you: Mostly indirectly, but the indirect channel is real: sustained conflict keeps fuel and shipping costs elevated, which keeps drug and supply-chain pricing elevated, which shows up eventually in the same hospital margin conversation that produced the "inadequate update" line above.

Section B — The Debate

Should health plans be allowed to issue prior-authorization denials generated by AI?

Why it's live right now: Reps. Herb Conaway (D-N.J.) and Gregory Murphy (R-N.C.) — both physicians — introduced the bipartisan Protecting Patients from Automated Denials Act on July 16, and the AMA put its weight behind it in its July 31 advocacy update. Meanwhile CMS is running the opposite experiment: the WISeR model has been using AI-assisted prior authorization inside traditional Medicare since Jan. 1 in six states, and yesterday's IPPS rule made electronic prior authorization a mandatory hospital measure starting CY 2028.

The case for AI in the loop

Prior authorization is slow because it is manual, and manual review is where inconsistency lives — the same request gets different answers depending on who opens the file. Automating the first pass is how you get decisions in hours instead of weeks. Abe Sutton, director of the CMS Innovation Center, has argued that the target is low-value care that hurts patients rather than helps them: "Low-value services, such as those of focus in WISeR, offer patients minimal benefit and, in some cases, can result in physical harm and psychological stress. They also increase patient costs while inflating health care spending." He points out that WISeR contractors are not paid to deny — they're instructed to "get the determination right" — and that decisions come within 72 hours, 48 for expedited cases, with human clinical review layered on top. On the commercial side, AHIP reports that plans have already eliminated 11% of prior authorizations — 6.5 million fewer — and are targeting real-time answers on 80% of electronically submitted approvals by Jan. 1, 2027. That target is unreachable without automation.

The case against AI denials

A model can approve quickly without being trusted to deny. Denial is the decision with clinical consequences, and it should carry a physician's name. That's the premise of the Conaway–Murphy bill, which would bar plans from relying solely on automated tools: any AI-generated denial would require documented review by a qualified physician under the medical director's clinical direction, direct access to that reviewing physician, disclosure that AI was used, ten-year record retention, and reports to HHS every 90 days. The AMA's framing is blunt — "Artificial intelligence is no substitute for a highly trained physician," and MA plans "should not be permitted to empower AI to make prior authorization decisions, especially denials of medically necessary treatment." A 2025 AMA survey found 61% of physicians worry AI will increase denials of care they consider necessary. Critics of WISeR, including members of Congress who wrote to CMS, note that technology partners are compensated in part on the volume of care they deny — an incentive structure the model's defenders have to argue around rather than remove.

Where it might land: The Conaway–Murphy bill has no scheduled markup; the nearer-term forcing function is CMS's own CY 2027/2028 electronic prior authorization timeline, which will make hospitals participants in the automated pipeline whether or not Congress sets rules for the denial end of it.

Section C — On My Radar

Section D — Trends to Watch

Section E — Ideas & Opportunities

Section H — Background / Already Covered
📚 Archive · last 7 morning reports