Table of Contents
AI in Healthcare: At his rehabilitation medicine practice in Illinois, Dr. Azlan Tariq was locked in a relentless struggle, dedicating a grueling seven hours weekly to persuading insurance companies to cover his patients’ treatments. Too often, these battles ended in defeat. There was the 45-year-old man who languished in a wheelchair for five months, repeatedly denied a prosthetic leg.
Or the stroke survivor rehospitalized after a fall because his insurer deemed his rehab could be done at home. Such stories became all too common in Dr. Tariq’s 12-year career, with an ever-growing list of treatments requiring pre-approval and an escalating number of denials.
The Unexpected Ally AI in Healthcare for Treatment Approvals
In his quest to shield his patients from subpar care and himself from a deluge of paperwork, Dr. Tariq turned to an unexpected ally: generative AI. For a growing number of physicians, AI chatbots, capable of drafting insurer letters in mere seconds, are revolutionizing the fight for claim approvals, achieving in minutes what years of advocacy and AI in healthcare reforms have not.
“We lacked legislative and policy tools to fight back,” Dr. Tariq said. “Now, we finally have a weapon.” This technological shift comes as some of the nation’s largest insurers face class-action lawsuits for allegedly using AI in healthcare to swiftly deny vast numbers of claims, depriving critically ill patients of necessary rehabilitation treatments.
Experts fear an impending AI “arms race” in prior authorizations, pitting bot against bot. Among doctors, the process is almost universally despised. Dr. Robert Wachter, chair of the medicine department at the University of California, San Francisco, humorously remarked that mentioning prior authorizations at a cocktail party is a surefire way to see a physician lose their composure.
Insurance companies designed the prior authorization process to control AI in healthcare costs by curbing doctors’ use of unnecessary and expensive treatments. However, Dr. Jonathan Tward, a Utah-based radiation oncologist, said insurers often demanded exhaustive documentation, even for standard cancer treatments, making him feel like a soldier in a never-ending “war of attrition.”
AI in Healthcare Authorization, Easing the Burden on Physicians
Physicians and their staff spend an average of 12 hours a week on prior authorization requests, a burdensome process widely regarded as detrimental to patient health, according to an American Medical Association survey.
Enter ChatGPT. Dr. Tward now types a few sentences describing the purpose of a letter and the scientific studies he wants referencing, and a draft is produced in seconds. He can then instruct the chatbot to expand it fourfold. “If you’re going to put up barriers for my patients, I’m going to make it very time-consuming when I push back,” he declared.
Dr. Tariq has also embraced Doximity GPT, a HIPAA-compliant chatbot version, which has halved his prior authorization time. More crucially, this tool—drawing from his patients’ medical records and insurers’ coverage requirements—has significantly increased the success rate of his letters. Since adopting AI for prior authorization requests, about 90% of Dr. Tariq’s coverage requests have been approved, compared to a meager 10% before.
Generative AI is particularly beneficial for doctors at small practices, who might not have the time to appeal insurer decisions even when patient care is at stake. Nearly half of doctors surveyed by the AMA cited lack of time or resources as reasons for not appealing claim denials. Dr. Michael Albert, an obesity medicine specialist in Oklahoma, said AI-enabled his small telehealth practice to escalate from almost never appealing insurance denials to filing 10 to 20 appeals weekly.
Now, Dr. Albert feels his practice can “compete with companies that have virtually infinite resources.”
Mainstreaming AI in Healthcare: The Battle Between Physicians and Insurers
While AI use remains primarily among tech-savvy physicians, companies are working to mainstream the technology. Epic, one of the nation’s largest electronic health record companies, has introduced an AI-powered prior authorization tool to a select group of physicians, according to developer Derek De Young. Several major health systems are piloting Doximity GPT, aimed at assisting with various administrative tasks, including prior authorizations.
Insurers, however, are not standing still. Chris Bond, a spokesman for America’s Health Insurance Plans, stated that insurers welcome efforts to streamline the process, including appropriate AI use. Dr. Jeff Levin-Scherz, a health policy expert at the Harvard T.H. Chan School of Public Health, noted that most health plans are at least evaluating AI for claims reviews if not already employing it. Nonetheless, these AI tools could also facilitate billing for unnecessary treatments, Dr. Levin-Scherz warned.
As physicians accelerate their use of AI to craft prior authorization letters, Dr. Wachter is confident that insurers will likewise enhance their AI capabilities to counter them. “It’s an automatic conflict,” said Dr. Wachter, who authored a book on digital technology in medicine. “Their AI will deny our AI, and we’ll go back and forth.”
Dr. Wachter envisions a future where advanced AI technologies might help create a system that relies less on sending letters and more on automation. Algorithms could automatically approve insurance coverage based on the latest scientific literature.
An AI tool might notify a doctor that a $100,000 drug can be substituted with a similarly effective $250 pill. Until then, many doctors relish a newfound, albeit temporary, advantage in the insurance claims battle. “It’s definitely an AI race,” Dr. Tariq said, “but it wouldn’t be fair if the physicians didn’t have AI.”
