When AI finds more diagnoses, hospital bills rise: what the $942 million claim shows
secondary diagnosis
An additional health condition listed besides the main reason for a hospital stay.
de-identified inpatient claims
Hospital insurance records with names and direct identity details removed.
ambient scribes
Tools that listen to visits and draft medical notes.
What happened
The Blue Cross Blue Shield Association (BCBSA), a federation of U.S. health insurers, released a claims analysis on September 24. It examined de-identified inpatient claims from Blue Cross and Blue Shield plans. Compared with 2023, it estimates $942 million in additional spending during 2024 and 2025. About $653 million came from secondary diagnoses. These added conditions moved claims into higher-paying categories, BCBSA says. The number describes extra spending by member plans. It does not mean hospitals delivered $942 million in extra treatment. Read the association’s analysis
The background
Medical billing turns a patient’s conditions and care into standard codes. A secondary diagnosis is an additional condition beyond the main reason for a hospital stay. Adding one can make a case look more complex to the payment system. AI coding tools scan lab results, electronic records, and doctors’ notes for such details. Some ambient scribes listen to visits and draft notes. BCBSA says more than 60% of hospital systems now use AI-enabled tools. Hospitals say these systems improve accuracy and reduce administrative work. They also say the tools help answer close scrutiny from insurers.
Why it matters
That creates a difficult question. If coding becomes more detailed while care stays similar, the system may pay more for paperwork rather than treatment. BCBSA warns that higher plan spending can eventually pressure premiums and out-of-pocket costs for families, employers, and taxpayers.
The $942 million estimate applies to BCBS member plans. It is not a measure of all U.S. healthcare spending. The story also shows why AI may not lower costs automatically. Hospitals and insurers already argue over claims. Faster AI submissions and faster AI reviews could intensify that conflict. TechCrunch reported on the wider dispute.
What is documented
Several findings support the association’s concern. More than 55,000 cases above the 2023 baseline moved into higher-severity, higher-payment groups after secondary diagnoses were recorded. In major bowel procedures, the highest-complexity category rose from 10.2% to 22.7%. The non-complex category fell from 36.6% to 32.8%. BCBSA says those shifts represented almost $61 million in incremental claims costs.
It also found similar or lower intensive-care use, transfusion, reoperation, and length of stay among hospitals with more complex coding. Anemia diagnoses were more common in one group, but transfusions were less common. Fierce Healthcare detailed the findings.
What remains unknown
The evidence has limits. Claims data are not the same as complete medical charts. They cannot show whether each additional diagnosis was clinically correct. They cannot show every treatment decision. They also cannot prove that AI caused every dollar of the increase.
The analysis links rising AI adoption with more complex coding. That is an important signal, not a final causal verdict. Hospitals may be documenting real conditions more completely. Insurers may also interpret the numbers through their own payment concerns. The sides need independent review.
What to watch next
Researchers should compare billing codes with clinical records. Auditors should test whether added diagnoses changed treatment, resources, or outcomes. Hospitals may need to disclose how coding tools are used. Insurers may need to explain their own review systems. Payment rules may also need to reward care that patients receive, not simply more labels.
BCBSA says it plans analyses of outpatient care and other diagnosis groups. The central question is simple: can AI make records clearer without making the entire system more expensive? See the wider reporting.
Did AI make hospital bills more expensive?
📰 Full story: When AI finds more diagnoses, hospital bills rise: what the $942 million claim shows
A health insurer group says AI helped change how hospitals describe patient care. That change may have raised payments.
secondary diagnosis
An extra health problem listed besides the main reason for a hospital stay.
payment group
A billing category that helps decide how much an insurer pays.
inpatient claim
A bill sent after someone stays in a hospital.
💡 The gist
- Artificial intelligence helped hospitals find more health problems.
- Extra labels moved some bills into higher payment groups.
- Blue Cross says costs rose by $942 million over two years.
The Blue Cross Blue Shield Association is a group of U.S. health insurers. It studied hospital bills from 2023 through 2025. The group says complex cases became more common. They rose from 37% to 40% of covered inpatient claims. The estimate compares the later years with 2023.
Hospitals send insurers bills after treating patients. Each bill uses codes for illnesses and care. A secondary diagnosis is an extra health problem. It is not the main reason for the hospital stay. Finding one can move a bill into a higher payment group.
AI tools can search lab results and electronic records. Some tools also draft notes from doctor-patient conversations. This can save time. It can also find more conditions to list. BCBSA says more than 60% of hospital systems use such tools.
BCBSA links the coding change to $942 million in extra spending. About $653 million came from secondary diagnoses. The group says treatment did not change in the same way. That could mean the payment system rewards more detailed paperwork. It could also push up premiums or out-of-pocket costs. The estimate covers Blue Cross plans. It is not the total cost of healthcare in America.
The number needs careful reading. The study used insurance claims, not every medical chart. Claims cannot show the full condition of each patient. They cannot prove that every code was wrong. They also cannot prove that AI caused every dollar. Hospitals say AI helps them code more accurately. They also say insurers often delay or reject payments. The two sides disagree about what the change means.
Independent checks matter next. Researchers should compare claims with clinical records. They should ask whether added diagnoses changed care. Auditors may also need to examine hospital and insurer tools. BCBSA says it plans more work on other services and billing groups. Read the association’s analysis. See the reporting from Fierce Healthcare.
A computer helper found more names on hospital bills
📰 Full story: When AI finds more diagnoses, hospital bills rise: what the $942 million claim shows
Insurers say the bills grew, even when care did not grow the same way.
Blue Cross Blue Shield Association
A group that connects many U.S. health insurers.
AI
A computer tool that can find patterns in information.
A hospital writes down what happened to a patient. An AI helper can read those notes. It may find another health problem to add. More health problems on a bill can mean more payment. Blue Cross Blue Shield Association is a group of U.S. insurers. It studied hospital bills from 2023 to 2025. It says costs grew by $942 million over two years. It also says the care did not grow the same way. But the answer is not settled. The study mostly looked at bills. It did not check every doctor’s note. So we do not know if the patients were sicker. We also do not know how much AI caused. People need more checking. Read the association’s analysis.