AI Coding Tools Are Adding Diagnoses — and Nearly $1B in Hospital Costs
A Blue Cross study links AI-assisted medical coding to $942 million in extra inpatient costs in two years, but hospitals say sicker patients explain much of it too.
What happened: The Blue Cross Blue Shield Association (BCBSA) released a study estimating that rising coding intensity, much of it tied to AI tools, added $942 million to member companies' inpatient costs in 2024 and 2025 compared with 2023. About $653 million of that came from hospitals billing more secondary conditions per patient case. It follows an earlier March analysis that linked AI-enabled coding to $663 million in extra inpatient spending and at least $1.67 billion in outpatient spending.
Why it matters: Higher documented severity means higher reimbursement, and insurers say the added cost eventually flows through to premiums for employers and patients. BCBSA points to specific evidence: among bowel surgery patients, diagnoses of partial intestinal blockages rose 55% and excess-acid diagnoses rose 33% between early 2023 and late 2025, without a matching rise in procedures. Anemia diagnoses also climbed without more blood transfusions. 'If patients are truly sicker, we'd expect to see more treatment,' said BCBSA's Luke Chalker.
How it works, plainly: Ambient AI scribes listen during patient visits and draft clinical notes automatically. Separate AI systems scan electronic records, physician notes, and lab results looking for diagnoses that exist but weren't formally logged. When those secondary conditions get added to a chart, a case can shift into a higher-severity billing category and a bigger payout, even when the actual treatment given to the patient stays the same.
The dispute: The American Hospital Association rejects the insurer framing, arguing that more intense coding can reflect patients genuinely getting older and sicker, plus more accurate capture of conditions that were always there. It cites a roughly 5% rise in hospital case-mix severity from 2019 to 2024. Meanwhile, both sides are racing to deploy AI for their own financial benefit: US healthcare AI spending hit $1.4 billion in 2025, and UnitedHealth and HCA Healthcare each project hundreds of millions in AI-driven savings for 2026.