By Sriparna Roy and Sneha SK
March 12 (Reuters) – Artificial intelligence has been placed on both sides of a battle between U.S. health care systems that want to pay more for medical procedures and insurers that want proof that services were necessary, and experts are having a hard time predicting a winner.
Centen, an insurer that focuses on the Medicaid program for low-income people, recently raised the issue, saying that hospitals are using revenue software to generate payments either fraudulently or even improperly.
“There are some of these pockets — where people come to the emergency department with a fever, all of a sudden they have sepsis,” Centene CEO Sarah London said at an investor conference in September.
Blue Cross Blue Shield’s analysis of its commercial hospital claims found that about $663 million in inpatient costs and at least $1.67 billion in outpatient costs may be tied to AI-enabled coding practices nationwide.
“We’re seeing more AI tools being used at different points in the care and billing process, and when these tools work independently, they can inadvertently cause friction,” said Razia Hashmi, vice president of clinical affairs at Blue Cross Blue Shield Association.
Aim to reduce costs
In recent months, some health insurers have increased their reliance on AI to help manage treatments and bills they say are unnecessary, even as hospitals use AI tools to document medical services with codes that increase reimbursement, a review of company statements and interviews with dozens of experts and analysts found.
The United States spends more on health care than any other country at about 18% of its gross domestic product.
Both parties hope that AI will reduce costs. Consulting firm McKinsey estimates that for every $10 billion in revenue, AI could save insurers $970 million through claims management, medical prior authorization requests and clinical care guidance.
AI tools are already leading to savings in hospital care and they could amount to $900 billion by 2050, Morgan Stanley said in a September research note.
“The idea of (AI) bot versus bot is inherently a no-win situation,” said Christina Silcox, director of digital health research at the Duke-Margulis Institute for Health Policy.
Corporate Savings Plan
Several analysts, including TD Cowen’s Ryan Langston and Leerank Partners’ Whit Mayo, said insurers and hospitals could save money by using AI, but they did not provide estimates. Companies are also using AI extensively for administrative tasks.
