AI Billing Boosts Hospital Costs by Nearly $1 Billion
According to a recent study by the Blue Cross Blue Shield Association, AI-based billing for more clinically complex inpatient admissions has generated almost $1 billion in extra health-care expenditures over the past two years. The analysis indicates that coding practices have shifted, yet there is no proof that the level of care delivered has changed.
The Role of Secondary Diagnoses
The association, referencing data from BAM.ai, says that over 60% of hospitals and health-care systems now deploy AI tools capable of reviewing laboratory results and patient encounter documentation to flag secondary conditions. The report notes that patients frequently move into a more expensive billing tier after a secondary condition is identified. Because a single lab value can generate such a diagnosis, AI systems are particularly effective at spotting them.
Claims that were re-classified to higher-pay categories because of secondary conditions represented about 70%, approximately $650 million, of the cost growth experienced by Blue Cross and Blue Shield insurers between 2023 and 2025, the analysis shows. Overall, this shift added an estimated $942 million to health-care spending in that period, contributing to larger premiums and out-of-pocket charges for families, employers and taxpayers.
The Disconnect Between Coding and Treatment
In the study, investigators examined secondary diagnoses such as anemia that appeared after major colorectal surgery. The BCBSA highlighted a mismatch between the coding records and the actual treatment delivered.
Luke Chalker, senior vice president of product and data science at BCBSA, remarked, “If patients are truly sicker, we’d expect to see more treatment,” adding, “For example, we’re seeing significantly more anemia diagnoses at these hospitals without a corresponding increase in transfusions. The disconnect between diagnoses and treatment suggests that AI is identifying more billable conditions, not sicker patients.”
These results extend earlier BCBSA work that linked AI-driven coding solutions to rising health-care costs by surfacing extra diagnoses that can boost hospital reimbursements. Hospital operators argue that AI assembles a full medical record, enabling more precise payment calculations. In the revenue-cycle process, AI extracts data from labs, prescriptions, orders, clinician notes and other sources to infer patient care. Professional coders can then locate billing gaps and avoid what providers label as omission errors.
