Every day, patients are hit with the news that the treatment their trusted physician ordered has been denied. Now imagine learning that no human, let alone a doctor, even looked at the claim before the denial was issued. It’s not a science fiction plot, but rather the new frontier in health care.
At the center of the issue is prior authorization, a process by which insurers require approval before certain services, prescriptions and treatments can move forward. What began as a cost-cutting measure has morphed into a major barrier for patient care: According to the American Medical Association (AMA), one in three physicians report that prior authorizations are often or always denied, leading to an average of 13 hours a week spent working to cut through prior authorization- related red tape.
Prior authorizations and claim denials slow down patient care, often resulting in worse health outcomes and higher costs in the long run. Now, the rise of artificial intelligence is taking matters out of doctors’ care altogether, and physicians like me are concerned. Will AI become an automated gatekeeper programmed to say “no,” doling out denials left and right with little regard for patient outcomes?
A recent AMA survey found that three in five physicians worry that health insurers’ use of AI is increasing prior authorization denials. We’ve already seen the damage prior authorization can cause when humans are making the decisions; adding AI to the mix will likely magnify the issue.
The role of AI in issuing claim denials is already in place, with expanded applications arising by the day. The Centers for Medicare and Medicaid Services (CMS) recently announced a pilot program coming to Oklahoma and a handful of other states called the Wasteful and Inappropriate Service Reduction (WISeR) Model. The program will implement prior authorization for more than a dozen treatments and procedures within traditional Medicare, ranging from incontinence control devices to steroids for pain management. The model plans to utilize AI systems to determine eligibility for such treatments.
We can’t stop AI from infiltrating health care, and truth be told, there are useful applications including improving patient health through image analysis, helping with personalized treatment plans and predicting disease outbreaks. But, in the hands of insurers, whose primary goal is to cut costs, it has been proven to be a problematic tool. It should never replace a human when it comes to making patient decisions. There are more efficient ways to lower health care costs, such as preventative care.
There’s nothing more personal than medical care: a computer can reject countless claims per minute, but it can’t understand the nuances of pain, compassion or the consequences of delayed treatment. I believe in the primacy of the doctor- patient relationship, which should never be outsourced to a computer.
As we move toward a future in which AI has an even bigger role in medical care, we have to ask ourselves whom we trust more with our health care: Trained doctors or algorithms written to save money, not necessarily vision or lives?
Sumit Nanda, M.D., is a board-certified ophthalmologist with INTEGRIS Health in Oklahoma City. He is president of the Oklahoma State Medical Association.
