A recent investigation by STAT News found that AI algorithms have influenced how Medicare insurers deny insurance to patients. In some cases, insurers cut off benefits for elderly patients because the AI says they should be better, ignoring what human doctors have to say about the patient’s condition.
STAT News reports that as AI continues to become integrated into various industries, its impact on the healthcare sector is starting to become apparent. A recent investigation by STAT has revealed that AI’s influence on Medicare Advantage insurers may be driving denials to unprecedented levels, affecting millions of older Americans who rely on the taxpayer-funded program.
In one striking example, Frances Walter, an 85-year-old Wisconsin woman with a broken left shoulder and an allergy to painkillers, was expected to make a quick recovery by the algorithm, which didn’t factor in the opinion of her human doctor. Her Medicare Advantage insurer, Security Health Plan, followed the algorithm’s estimate and cut off payment for her care after 17 days, even though she still required assistance. As a result, Walter had to spend her life savings and enroll in Medicaid to continue her treatment.
The STAT investigation found that health insurance companies are increasingly using unregulated predictive algorithms to determine when to stop payments for older patients’ treatments. While the insurers claim that these tools are merely suggestive, in practice, they often serve as hard-and-fast rules that don’t account for individual circumstances or changes in a patient’s conditions.
As more Americans over 65, and those with disabilities, choose plans with lower premiums and prescription drug coverage, Medicare Advantage has grown significantly more profitable for insurers. These plans, however, give insurers more discretion to limit and refuse services. The last ten years have seen the emergence of a new industry focused on using AI to predict patient discharge dates from hospitals, doctor types, and therapy hours. Insurers have even acquired companies specializing in these predictive tools.
The increasing reliance on AI algorithms to make crucial decisions about patient care is raising concerns among medical professionals and patient advocates. In many cases, the algorithms’ recommendations conflict with basic rules on what Medicare plans must cover, creating heated disputes between doctors and insurers and often delaying treatment for seriously ill patients.
The FDA assesses the AI models used by doctors to identify diseases like cancer or recommend the best treatment. In contrast, the tools used by insurers to decide whether to pay for those treatments are not subjected to the same scrutiny, despite their influence on the care of the nation’s sickest patients.
Doctors and medical administrators report that Medicare Advantage payment denials for services that are regularly covered by traditional Medicare are happening more frequently. Insurers like UnitedHealthcare and others claim they discuss a patient’s care with providers before denial, however, many service providers claim that when they ask for explanations, they are met with blank looks and denials of their requests for more details.
“They say, ‘That’s proprietary,’” said Amanda Ford, who facilitates access to rehabilitation services for patients following inpatient stays at Lowell General Hospital in Massachusetts. “It’s always that canned response: ‘The patient can be managed in a lower level of care.’”
The lack of regulation and oversight of these predictive algorithms raises concerns about the impact on patient care and treatment access. As the influence of these tools continues to grow, the precise role they play in insurers’ decisions remains opaque. This raises important questions about the ethical use of AI in healthcare and the potential consequences for vulnerable patients who depend on Medicare for their medical needs.
Read more at STAT News here.
Lucas Nolan is a reporter for Breitbart News covering issues of free speech and online censorship. Follow him on Twitter @LucasNolan
Read the full article here
Discussion about this post