Blue Cross estimates $942 million in added costs from complex hospital coding

The Blue Cross Blue Shield Association estimates that a rise in complex hospital billing added $942 million to its member plans' inpatient costs in 2024 and 2025, compared with 2023 coding rates. Its September 24 analysis found more secondary diagnoses moving claims into higher-paying categories. BCBSA suspects the spread of AI coding tools is contributing to the change. The claims study does not measure what any particular tool did.

The big change

• What changed: More inpatient stays were billed in complex categories. BCBSA estimates the shift raised payments across Blue plans by $942 million over two years, with $653 million attributed to additional secondary diagnoses.
• Why it matters: Software that searches records and lab results can help hospitals document conditions. In a payment system where some diagnoses change the billing category, documentation can also change the price of a stay. Patients, employers and insurers have an interest in whether the added codes reflect patients' clinical needs.
• What to watch: The insurer compared diagnoses with treatments recorded on claims, but did not review each patient's chart or identify which hospitals used which AI tools. Clinical validation would help resolve the dispute over appropriate coding and patient severity.

Where the estimate comes from

BCBSA's three-page paper examines Blue plan inpatient claims from the first quarter of 2023 through the fourth quarter of 2025. The share coded with a complication or comorbidity, or a major complication or comorbidity, rose from roughly 37% in early 2023 to about 40% in late 2025. Against 2023 rates, the authors count 55,158 additional complex cases and estimate $653 million in higher payments from those cases, averaging $11,800 each. That is about 70% of their $942 million estimate for increased coding intensity.

The mechanism is a diagnosis-related group, or DRG. A hospital claim records the main diagnosis, other conditions and procedures. In DRG systems, some qualifying secondary diagnoses move a stay into a group with a higher payment. Medicare's explanation of its inpatient system describes that grouping rule; BCBSA's dollar estimate concerns its member plans' claims, not Medicare spending.

For one surgical category, major bowel procedures, BCBSA reports that the share of claims in the highest-complexity group rose from 20.2% in 2023 to 22.7% in 2025. The share without a complexity designation fell from 36.6% to 32.8%. It estimates $60.8 million in additional payments in that category alone. The paper does not publish a full patient-level dataset or its detailed cost calculation, limiting outside checks of the aggregate estimate.

The treatment comparison, and its limit

BCBSA compared hospitals in the top quarter for growth in complex coding with other hospitals treating major bowel cases. In 2025, 75.6% of claims at the top-growth hospitals were coded as complex, versus 65.0% elsewhere. Intensive-care use was 11.5% versus 13.2%, and median stays were four days in both groups. The authors argue that more severe diagnoses without more intensive treatment point toward a documentation change.

Their anemia example sharpens the question. Top-growth hospitals coded posthemorrhagic anemia in 13.7% of bowel-procedure claims, compared with 9.9% elsewhere. Among patients with that diagnosis, 16.9% received a transfusion at top-growth hospitals, compared with 19.3% at the others. These are treatment indicators visible in claims. A patient may have a valid diagnosis without needing a transfusion; the figures alone cannot decide whether an individual code was justified.

The American Hospital Association's August fact sheet disputes insurers' broader AI upcoding argument. It says hospitals are treating a sicker inpatient population as less complex care moves to outpatient settings, and that AI can improve the accuracy of documentation under existing coding rules. AHA also criticizes insurers' automated reductions of hospital claims. Its fact sheet predates this September analysis and is the hospital trade group's position, not a patient-level rebuttal to BCBSA's new sample.

The missing link is direct evidence connecting the billing shift to specific AI tools. BCBSA points to timing, diagnoses that software could extract from lab data, and differences between coding and treatment. Fierce Healthcare's report on the briefing says BCBSA acknowledged that clinical records would more directly show whether patients were sicker. Until that comparison and tool-use data are available, the $942 million is BCBSA's estimate of extra payments associated with coding changes, rather than a measured cost caused by AI alone. It also does not show how much any one patient's premium or bill changed.

Sources & further reading

• BCBSA's September 2026 Hospital Coding Intensity Analysis: Major Bowel Procedures is the original three-page claims analysis. It provides the sample period, coding and treatment comparisons, and the insurer's $942 million estimate. Its published methods do not identify AI tools on individual claims or disclose a patient-level dataset.
• BCBSA's September 24 announcement explains the association's interpretation and its concern about AI-assisted coding. The March 2026 maternity analysis cited there is a separate study with a different population and estimate.
• The American Hospital Association's August 2026 fact sheet sets out the provider trade group's position on patient acuity, coding accuracy and insurer downcoding. It predates BCBSA's September paper.
• CMS's January 2026 Medicare payment systems guide explains how principal and secondary diagnoses affect severity groups in Medicare inpatient payment. It does not establish payment rules for every Blue plan contract.
• Fierce Healthcare's September 24 briefing report records BCBSA officials' account and their acknowledgment that claims alone cannot directly establish patient acuity.