AI Linked to $942M Surge in Hospital Costs as Patients Are Coded 'Sicker' Without More Treatment

Artificial intelligence-assisted hospital billing may have contributed to an estimated $942 million (£712.29 million) increase in healthcare spending as more patients were classified as medically complex without a corresponding rise in treatment, according to a new analysis from the Blue Cross Blue Shield Association.

The research found that hospitals increasingly placed patients into higher-reimbursement categories in 2024 and 2025 than in 2023, often after documenting additional or 'secondary' diagnoses.

Around 70% of the additional spending, approximately $653 million (£493.76 million), was tied to those secondary conditions, BCBSA said.

The findings come as AI use spreads across healthcare revenue-cycle operations, with one survey cited by BCBSA finding that more than 63% of healthcare organisations use AI in those workflows.

More Diagnoses, But Not More Treatment

BCBSA said one of the most striking findings was the gap between what appeared in patients' claims and the care they actually received.

Researchers examined conditions including anaemia following major bowel surgery. Hospitals recorded significantly more anaemia diagnoses, but researchers did not find a comparable increase in blood transfusions, which might be expected if patients had become substantially sicker.

'If patients are truly sicker, we'd expect to see more treatment,' said Luke Chalker, BCBSA's senior vice-president of product and data science.

'The disconnect between diagnoses and treatment suggests that AI is identifying more billable conditions, not sicker patients,' he added.

The analysis used de-identified claims from Blue Cross and Blue Shield companies, which collectively cover roughly one in three Americans.

BCBSA warned that higher reimbursements ultimately put pressure on insurance premiums and out-of-pocket expenses for households, employers and taxpayers.

AI Can Turn a Lab Result Into a Higher-Paying Claim

AI coding systems are designed partly to identify conditions that human coders or doctors might otherwise overlook.

A secondary diagnosis can sometimes be identified from a single laboratory result. Once added to a patient's record, however, that condition can move the hospital stay into a more severe billing category carrying a higher reimbursement.

The research does not establish that AI alone caused the entire $942 million increase, and BCBSA describes the technology as a potential contributor rather than proof that hospitals deliberately inflated claims.

BCBSA also acknowledged that its analysis relied on insurance claims rather than patients' underlying clinical records, limiting its ability to directly determine whether patients had become medically more complex.

Reuters reported that insurers including Centene have raised concerns that health systems' use of AI could contribute to aggressive or inappropriate reimbursement claims even when the treatment delivered has not materially changed.

AI Coding Companies Push Back

Companies behind the technology dispute the suggestion that higher coding necessarily amounts to inappropriate 'upcoding'.

Hamid Tabatabaie, president and CEO of AI coding company CodaMetrix, told Healthcare Dive that automated systems are often identifying legitimate conditions that hospitals previously failed to document.

'When last year they submitted their claims, if they weren't using a valid system, they were missing it,' he said.

Dr Travis Bias, deputy chief medical officer for health information systems at Solventum, similarly argued that the underlying issue is America's fee-for-service reimbursement model, which pays providers partly according to the complexity of patients' conditions.

Both sides therefore agree on one point: AI is changing how much information hospitals can capture from a patient's medical record.

The billion-dollar question is whether that produces more accurate bills, or simply more expensive ones.