When you hear about AI in healthcare, the focus tends to be on efficiency. That might mean faster turnaround on things like prior authorizations, less paperwork, or better scheduling that leads to shorter wait times. However, new research from the Blue Cross Blue Shield Association is telling a different story.
BCBSA estimates that hospitals' use of AI-assisted billing tools has added roughly $942 million in excess costs to the BCBS system alone over the past two years. Yet patients aren't receiving more care to match. That number comes from how care is being coded, not how care is being delivered.
Coding, not care
According to the analysis, patients aren't having more surgeries or procedures. Instead, additional diagnoses are being attached to claims, which pushes them into higher reimbursement tiers. This is known as upcoding, and BCBSA is alleging that providers are using AI to do it at scale.
What the data shows
The data lends some weight to those allegations. Using maternity admissions as an example, BCBSA found sharp increases in cases coded for acute posthemorrhagic anemia. This condition normally requires medical intervention, such as a blood transfusion. But at the hospitals where this diagnosis grew fastest, that level of treatment wasn't showing up in the claims.
At one hospital that publicly adopted AI coding tools, case complexity rose 6.7%, while other hospitals in the same state saw increases of less than 1%. Hospitals are pushing back, arguing that higher coding reflects better documentation and sicker patients, not manipulation.
What it means for your plan
When the cost of claims increases, premiums follow. Whether you're on a fully insured, level-funded, or self-funded plan, the cost of care will ultimately show up in your employer-sponsored health plan. It's more important than ever to understand your claims and how they're driving your company's costs. If you'd like to take a closer look at what's happening in your plan, let's talk.