BCBS’s Coding Intensity Complaint Is a Map of Upcoming DRG Downgrades

DRG downgrades are coming… On September 24, Blue Cross Blue Shield released a claims analysis that blamed hospital AI coding tools for $942M in added costs to Blue plans over two years. The New York Times ran a similar article that day, about hospital AI and insurer AI driving medical costs higher, and most of the framing focused on hospitals gaming the system. So much context was left out (of both).

1. The BCBS total comes from a payer-authored analysis built on payer claims data. BCBS acknowledged it never looked at clinical documentation. It measured what hospitals billed, not what the chart showed.
2. A BCBS executive told the New York Times the situation was “a completely one-sided bloodbath.” The financials show two sides bleeding. Kaufman Hall’s adjusted year-to-date median hospital operating margin closed out March at 1.7%, and national hospital operating margins through May were down 5% from a year earlier. Blue plans had a hard 2025 too, with combined margins near 1.4%. But Elevance Health, the largest Blue licensee, still earned $1.45B in Q2 and returned more than $600M to shareholders through buybacks and dividends. Health systems don’t have a buyback line.
3. Payers’ use of AI to deny both health systems and patients is far more prevalent and advanced. Hospitals didn’t start the chart-scanning. McLaren Health Care’s CFO told the NYT that payers already use AI to comb charts for claims to deny, and hospitals are reading the same charts in response.
4. BCBS acknowledged the analysis relied on claims rather than clinical documentation. It measured what hospitals billed. It never looked at what the chart showed.

The misleading framing aside, the report is useful for revenue cycle leaders who are in the trenches. The report lists the exact diagnoses, DRG families and statistical tests payers will use to drive the next round of DRG downgrades.

What BCBS claims

BCBS found the share of inpatient claims billed as medically complex rose from 37% in 2023 to 40% by the end of 2025. About 70% of that increase came from secondary diagnoses that moved a claim into a higher-severity DRG, which BCBS prices at $653M (~$11k per case). Its core argument is “clinical discordance.” The hospitals coding the most complexity showed similar or lower ICU use, transfusion, reoperation and length of stay than their peers.

Five DRG downgrade targets the report just announced

1. Posthemorrhagic anemia and other bump codes. BCBS named anemia specifically. Top-quartile hospitals diagnose it in 13.7% of cases vs 9.9% for everyone else, yet transfuse diagnosed patients less often (16.9% vs 19.3%). Expect clinical validation reviews asking for hemoglobin trends, transfusion or iron orders, and documented clinical significance on any claim where anemia is the CC or MCC setting the DRG. Hyponatremia appeared alongside it in press coverage.
2. Major bowel procedures. Highest complexity claims in this family climbed from 10.2% to 22.7% of cases, about $61M of the total. Surgical DRGs with steep severity ladders move the most money per secondary diagnosis, so they’ll get reviewed first.
3. Treatment intensity as an acuity test. BCBSA compared coded severity against ICU days, transfusions, reoperations and length of stay. Payers can run that test on your claims without requesting a single record. A complex DRG attached to a short, uneventful stay will be a target.
4. Peer divergence. BCBS’s VP of Clinical Affairs asked why one of two similar hospitals treating similar patients would diverge. That’s a benchmarking screen. If your CC/MCC capture rate for a DRG family sits in your market’s top quartile, your claims will land in the review sample regardless of your documentation.
5. Appeal volume. Payer-side analytics vendors told reporters they’re seeing LLM fingerprints on incoming appeals. High-volume appealers should expect tighter scrutiny and more templated rejections.

Capturing more codes doesn’t close the gap

Better documentation is important, and some of the “coding intensity” BCBS measured reflects real acuity that was undercoded for years. But payers adjust at bulletin speed and providers adjust at contract-amendment speed. Coding pressure won’t disappear, it’ll move into clinical validation audits, post-payment reviews and DRG downgrades.

What the report can’t see, and you can

BCBS argued from claims alone, but health systems can access both the chart and the remittance. Connecting the two by payer, DRG and diagnosis shows where you’re exposed before audits and enables you to track/trend payer behavior and identify root causes of adverse outcomes.

Sift’s RevProtect platform ties clinical inputs to financial outcomes to surface payer-specific downgrade patterns, so teams know which reviews they can win and which documentation gaps to fix first.