An analysis of 62 million claims conducted by the Blue Cross Blue Shield Association found that hospitals using AI tools for diagnosis coding saw their costs rise by $2.3 billion over three years. The use of AI technologies, including ambient listening and autonomous coding, pushed the share of admissions coded as complex from 46.8% to 59.8% at the most automated facilities, with no measurable improvement in patient outcomes. In maternity care, the diagnosis of acute posthemorrhagic anemia surged from 4% to over 12% at these hospitals, while transfusion rates remained flat, generating roughly $22 million in questionable charges. Coding intensity alone accounts for approximately 20% of the 9% rise in inpatient costs observed between 2023 and 2024. Insurers are deploying detection tools to identify upcoding practices, but employers are absorbing these additional costs through higher premiums.
Source: Read the original article
