Blue Cross association estimates AI coding played a role in $942 million in extra costs
The Blue Cross Blue Shield Association said on 24 September 2026 that AI-assisted hospital coding was associated with an estimated $942 million in additional member-plan costs in 2024 and 2025, compared with 2023, fueling a dispute over rising bills.

The association's analysis tracked inpatient billing data from hospitals and other facilities, comparing 2024 and 2025 claims with a 2023 baseline, BCBSA said. The share of stays classed as medically complex rose from 37% at the start of 2023 to 40% by the end of 2025. About 70% of the coding-intensity increase came from more than 55,000 additional cases above the 2023 baseline in which a secondary diagnosis moved a claim into a higher-severity, higher-reimbursement diagnosis-related group, or DRG, Fierce Healthcare reported. The secondary-diagnosis portion, $653 million, worked out to about $11,000 per excess complex case, BCBSA said. The association represents 31 independent Blue Cross Blue Shield insurers serving more than 100 million people.
Luke Chalker, BCBSA's senior vice president of product and data science, told CNBC that about 70% of the $942 million, or $653 million, was tied to additional diagnoses that were not accompanied by a change in care. He stopped short of attributing the entire increase to AI. "While multiple factors contribute to coding intensity, the findings suggest AI-enabled coding and documentation tools are playing a role," he said, according to CNBC. He added that more complex coding can raise reimbursement without more care, and those costs can eventually appear in premiums and out-of-pocket spending.
What the analysis can and cannot show
BCBSA acknowledged the analysis is limited because it relies on claims rather than clinical documentation, which would more directly measure whether patients were substantially sicker, Fierce Healthcare reported. Chalker said he believed a similar analysis with clinical data would make a stronger case for the association's position.
Hospitals push back
The American Hospital Association disputed the premise. A spokesperson told CNBC that patients today are older and more clinically complex, and that AI tools help providers capture their conditions for care planning. The AHA said the analysis lacks the context needed to assess how the tools affect quality, access or spending, and criticized insurers for relying on automated downcoding and denial practices. Chalker said Blue Cross Blue Shield companies also use AI in claims review, but that "any clinical denial is always reviewed by a qualified human clinician." Hospitals say the software helps them code claims accurately and efficiently under heavier scrutiny from payers, Fierce Healthcare reported.
Not every added diagnosis is wrong. Christopher Whaley, a health economist at Brown University, told CNBC that many additional diagnoses were legitimate and had not been captured before. But he said some conditions clinically do not matter and do not influence care while still allowing another billing code and higher payment.
Examples from the analysis
Within major bowel procedures, claims at the highest level of complexity rose from 10.2% to 22.7%, while non-complex cases fell from 36.6% to 32.8%. Together they accounted for nearly $61 million of the incremental claims costs, Fierce Healthcare reported.
BCBSA also flagged posthemorrhagic anemia, which it called a common bump code. The top quartile of hospitals diagnosed it in 13.7% of cases against 9.9% elsewhere, while transfusing a smaller share of those patients, 16.9% versus 19.3%, according to Fierce Healthcare. Among the top quartile of hospitals by complex DRG cases, ICU use, transfusions, reoperations and median length of stay were similar to or lower than peers, BCBSA said.
Some of the tools are ambient scribes, which listen to patient conversations and draft medical notes. Providers used them to spot secondary conditions, BCBSA said, according to PYMNTS. BCBSA cited a June survey in which more than 63% of healthcare organizations reported using AI in revenue cycle workflows, Fierce Healthcare reported. The association said the growth in complex coding came during a period when 60% of hospital systems began using AI coding tools.
Premium pressure
Benefits consulting firm Marsh expects the cost per employee for health coverage to rise 8.2% on average in 2027, which would be the highest increase since 2003, CNBC reported.
Sources
18- 01Health insurer points finger at AI as nearly $1 billion in questionable hospital charges appearEN
- 02Hospitals' use of AI coding tools cost BCBSA plans $942M more for similar care: analysisEN
- 03AI-Generated Medical Coding Adds Nearly $1 Billion to Blue Cross CostsEN
- 04Insurers claim AI is already increasing healthcare costsEN
- 05Hospital AI use is causing medical bills to soar, with Blue Cross estimating $1 billion in extra insurance costsEN
- 06Insurers say they’ll deploy more AI to combat ‘aggressive’ coding by hospitals - STATEN
- 07Eight Trends Shaping 2026 Healthcare Costs - Peterson-KFF Health System TrackerEN
- 08Are bigger hospitals and insurers better for consumers? - Association of Health Care JournalistsEN
- 09Amir Hamzah to discuss insurers’ restrictions on specialists with BNM, warns against practices driving up medical costs - The Edge MalaysiaEN
- 10STAT+: States sue Trump administration again to delay Medicaid work requirementsEN
- 11STAT+: After Medicare punts RPM overhaul to 2029, advocates see window to fight backEN
- 12STAT+: Democrats’ midterm landslide spells relief for hospitalsEN
- 13Opinion: AI depends on medical records that don’t always tell the full truth about womenEN
- 14Insurers After $6.3 Billion Crypto Heist: D’oh!EN
- 15Secure I.T. Environments to deliver modular data centers to hospitals in WalesEN
- 16Gene editing could replace a lifetime of medication, but there's a catchEN
- 17VMO2 to axe 30% staff costs, halt fiber build in £800M savings planEN
- 18Here's how much Trump's Iran war has added to Americans' fuel costsEN
All figures and quotations in this text come from the sources listed below.
Content prepared by the editorial team with AI assistance.
Comments
0- No comments yet — be the first.