Marco's Morning Report

Thursday, August 20, 2026
Normal day · Cape Coral, FL

Good morning, Marco. Quiet news cycle overnight, but the Annals dropped two papers on Tuesday that land squarely in your lane — one on whether AI scribes can actually write a note, one on why a single BNP tells you less than two. The Debate today is the one-hour antibiotic clock, which is a fight you live inside every third admission. Tropics are empty. Mets took the series. Coffee, then sleep.

Section A — Top Stories

Confirmed fresh, last 48 hours.

Documentation · Annals, Aug 18

AI scribes lost to humans on every quality domain — and lost worst in a noisy room

A head-to-head evaluation published Tuesday in Annals of Internal Medicine ran audio from five standardized primary care visits past 11 commercial ambient AI scribe tools and 18 human clinicians, then had 30 blinded raters score every resulting note on the modified PDQI-9. Human notes won across the board — accuracy, thoroughness, usefulness, organization, comprehensiveness. The gap was widest on an acute low back pain case recorded with background noise: clinicians averaged 43.8 out of 50, the AI tools 20.3. Lead author Ashok Reddy's framing is the practical one: these are draft generators, not note authors. (Annals · UW Medicine summary)

Why it matters to you: the worst-case condition in this study — ambient noise, an unfocused complaint — is a fair description of a 3 a.m. ED bay. And a nocturnist's H&P is the document the day team, the consultants, and eventually the utilization reviewer all read. If your group has an ambient scribe in pilot, the failure modes here are thoroughness and organization, which are exactly the domains a tired reader doesn't audit before signing.

Talking point: worth asking at the next group meeting whether anyone is sampling signed AI-drafted notes against the source encounter, and whether the group has a written expectation that the attending edits before signing rather than attests after.

Geriatrics · Annals, Aug 18

Two NT-proBNPs beat one: heart-stress trajectory reclassifies risk in older adults

Also published online Tuesday in Annals: an observational study of 8,454 community-dwelling older adults with no prior cardiovascular disease, with NT-proBNP measured at enrollment and again at year 3. Participants whose levels stayed elevated across both draws carried a substantially higher risk of incident cardiovascular disease and all-cause mortality. The finding that earns the paper its place — those whose elevated level came back down to normal had risk resembling people who never had a high value at all. (Annals · release)

Why it matters to you: this is an outpatient risk-stratification cohort, not an inpatient decision rule, so nothing changes at the bedside tonight. But it's a clean argument for a habit that is inpatient: when you draw a BNP on an 84-year-old with dyspnea, the prior values already sitting in the chart carry information the single number doesn't. A stable elevation and a new elevation are different patients.

Suggested action: a small one — when you pull a BNP on an elderly admission, make a habit of glancing at the trend line and putting the comparison in the assessment sentence rather than the raw value alone. It's free, and it's the thing the cardiology consult asks for anyway.

Section B — The Debate

One live argument, both sides, no verdict.

Should the one-hour antibiotic clock apply to everyone with suspected sepsis, or only to those in shock?

Why it's live right now: CMS has kept the SEP-1 severe sepsis and septic shock bundle in the Hospital Value-Based Purchasing safety domain for fiscal years 2026–2031, so the measure now carries money as well as chart review. Against that, a multicentre cohort of roughly 100,000 adults published August 5 in The Lancet Respiratory Medicine reported that the mortality cost of each additional hour to antibiotics was concentrated in the sickest tier — and that the 2026 Surviving Sepsis Campaign guidelines have already moved toward timing targets graded by sepsis probability and shock status.

Narrow the clock

Chanu Rhee, Harvard Medical School / Harvard Pilgrim Health Care Institute — Lancet Respir Med, Aug 5, 2026

The urgency of antibiotics is not uniform across everyone who trips a sepsis alert, and pretending it is has a cost. In the cohort, among patients with a NEWS2 of 7 or higher, every additional hour before antibiotics tracked with higher mortality; among lower-scoring patients, earlier antibiotics were not associated with better survival. A universal one-hour target therefore buys nothing in the low-risk majority while pushing clinicians toward reflexive broad-spectrum therapy before the diagnosis is clear — and IDSA's position paper on SEP-1 has argued for years, on the same logic, that the measure should be focused on septic shock, where the evidence for immediate treatment is strong. A bedside score that hospitals already collect can do the sorting.

Keep the clock broad

Steven Q. Simpson, University of Kansas; board chair, Sepsis Alliance — Sepsis Alliance; co-author, Surviving Sepsis Campaign 2026

The dominant failure in sepsis has never been overtreatment — it is recognition, and recognition is late far more often than it is early. Shock is a retrospective label: the patient who is in shock at hour four looked like the low-risk tier at hour one, and a measure that tells hospitals to relax on everyone below a threshold will move the whole distribution later, not just the safe part of it. The 2026 guidelines retain an hour-one expectation where sepsis probability is high, and the trial evidence on hour-1 bundle implementation showed antibiotics landing 73 minutes sooner with a point estimate favoring survival. Loosening a national measure is not a neutral act at the hospitals that only meet it because it is measured.

Where it might land: SEP-1's specifications sit inside the Hospital VBP program through FY2031, so any narrowing runs through CMS measure rulemaking rather than the guideline societies — meaning the argument will be had again in next year's IPPS cycle.

Section C — On My Radar

Noteworthy, not urgent.

Section D — Trends to Watch

Section E — Ideas & Opportunities

Section H — Background / Already Covered

Older than 48 hours, or reported recently without new developments
  • Rhee et al., NEWS2 vs shock status (Lancet Respir Med, Aug 5). The ~100,000-patient cohort behind today's Section B. Each hour to antibiotics mattered at NEWS2 ≥7; below that, earlier antibiotics were not associated with better survival.
  • Non-urinary sepsis in nonagenarians (Medscape, Aug 12). Among critically ill patients 90 and older, source of sepsis influenced short-term survival more than previously appreciated — urinary source did better.
  • CMS RAPID coverage pathway (Aug 11). New route to accelerated national Medicare coverage for eligible new technologies. Device-side for now, but the precedent is the same one the AI-device docket is arguing over.
  • CY2027 OPPS proposed rule — site-neutral imaging expansion, 340B to ASP-33.4%. Comments close August 31. Covered August 9 and 12; no new development since.
  • FY2027 IPPS final rule. Published August 4, effective October 1: net 2.3% inpatient increase, outlier threshold $49,346, Medicare DSH uncompensated care pool up $230M to $7.94B.
  • FDA generative-AI device discussion paper (docket FDA-2026-N-7874). Comments open through October 19. Yesterday's debate topic; nothing new filed.
📚 Archive · last 7 morning reports

A machine that can't hear over background noise, a lab value that only means something twice, and an empty tropical map in the third week of August. Sleep well, Marco.