Blue Cross analysis ties $942M in extra hospital costs to AI coding tools
The Blue Cross Blue Shield Association analyzed claims data from 2023 to 2025 and found that as hospitals adopted AI coding tools — ambient scribes and automated chart-scanning software that flag secondary diagnoses — more patients were coded as medically complex without a matching change in the care actually delivered. BCBSA estimates this added $942 million in costs across its plans, with $653 million of that traced to more than 55,000 claims pushed into higher-paying complexity tiers by added secondary diagnoses, averaging about $11,000 per case. BCBSA says the figure excludes cases where documentation reflected genuinely new care. Hospital groups dispute the framing, saying patients are sicker and that AI tools are correcting years of underdocumented complexity.
Why it's notable: Health systems are adopting the same ambient-documentation and coding AI at scale (e.g., the VA's Abridge contract), and a major payer is now using claims data to contest how those tools shift reimbursement — a dispute that will shape audits and contract terms for any practice using AI scribing or coding software.