This is part two of a two-part series.
In part one, “Is CMS’s Latest Effort Enough to Hold Medicare Advantage Companies Accountable?” I described how CMS plans to hold Medicare Advantage (MA) companies to higher standards for America’s seniors. This includes cracking down on misleading marketing tactics for Medicare Advantage plans, as well as instating new requirements so beneficiaries are not misled into signing up for a Medicare Advantage plan that doesn’t meet their needs or whose rules they don’t understand.
Today, I will describe how Medicare Advantage insurance companies have been using artificial intelligence (AI) algorithmic tools — rather than doctors or other medically trained people — to determine whether patients who are enrolled in their Medicare Advantage programs are worthy of care!
—
There is something disturbing happening with Medicare Advantage and Artificial Intelligence that many patients often aren’t aware of until it’s too late! According to a recent report by the medical journal Stat (as reported in “Stat News”), insurance companies have been using AI algorithmic tools, such as nH Predict tool from naviHealth, to pinpoint when they can shut off payment for a patient’s treatment. The denials that have followed are causing serious disputes between doctors and insurers, often delaying treatment of seriously ill patients. These patients are unaware of the algorithms and therefore unable to question their calculations. Since appeals challenging these denials can take weeks, months, or even years, the appeal can sometimes last longer than the patient!
Let’s look at Frances Walter, 85, who fractured her left shoulder. The octogenarian, who lives alone, called the emergency room and was sent to a hospital in Wisconsin. She was not aware of any costs for treatment, as she assumed her Medicare Advantage plan would cover them.
The algorithm used by her insurer estimated that on the 16th day of her postoperative stay, she could go home. Payment for care would then be stopped on the 17th day. The problem is that the doctor recommended a stay of at least three weeks, because Frances is allergic to painkillers and she is recovering less quickly than expected. On the 17th day, Frances still felt intense pain, and she was still unable to dress herself or push a walker.
Due to the algorithm and payment being cut off, Frances was forced to draw on her savings to cover the thousands of dollars necessary to be treated until a complete recovery. In this instance, the amount will fortunately be refunded to her, but that happened over a year later, after many procedures and the decision of a federal judge. This situation is far from being an isolated case, according to the information in Stat News.
It is not uncommon for insurance companies to reject claims for reimbursement of medical expenses. But in the past it has been humans making these decisions, not computers powered by artificial intelligence. And similar to what happened to Ms. Walter in our example, in the event of a disagreement, the patients must pay the bills (or risk collection action and negative credit reporting), appeal, and wait a long time to hope to be reimbursed. Some, who do not have the private funds to pay the bills, either do without care, or if they get the care and are unable to pay the bills, risk financial ruin and bankruptcy.
The nH Predict AI Tool Goes Too Far
Using AI for these purposes began in 2020, when UnitedHealth Group, which owns Optum and UnitedHealthcare, the nation’s largest Medicare Advantage insurer, acquired NaviHealth, a medical data management company whose flagship product is nH Predict. Based on an insured’s age, doctors’ diagnoses, and medical information, the algorithm compares each patient with similar profiles in a database of 6 million patients, renders a report which describes medical needs, estimated length of hospital stay, and target discharge date of a patient.
According to UnitedHealth Group in a press release at the time of the takeover, “(t)his tool is used as a guide” and is designed “as a way to avoid unnecessary care and overcharging.” However, doctors and patients have expressed concerns that it is being used for much more than that today. In fact, the refusals of support started to multiply from the moment that UnitedHealthcare was at the helm of nH Predict. Denials of coverage from UnitedHealthcare and naviHealth have become the norm, even when patients are obviously sicker than the algorithm predicts. As proof, the number of patient appeals between 2020 and 2022 challenging refusals of requests for care increased by 58 percent, according to a federal database. And in the majority of cases, the judge decides in favor of the patient and orders a refund!
Another Problem with AI: the Law Does Not Impose Transparency on Insurers Who Use AI
According to Stat News, when doctors or hospital directors ask for explanations about the short duration of coverage by the insurer, the AI “Black Box” (nH Predict) has become a general excuse to justify refusals of payment. Patients get the same response when they inquire — that the denial is part of the procedure that is being followed. For example, the daughter of Dolores Millam, 89, who was refused support for a stay in a nursing home after a fracture, states: “When you try to call and reason with someone, when you want to get an explanation, you end up talking to a service based in another country. For them, it is simply a question of procedure and has nothing to do with the care.”
Unfortunately, what is happening is not unlawful. Currently, the law does not impose any transparency on insurers who use an AI algorithm. According to Stat News, “(t)hey are neither forced to make public data, studies of effectiveness, nor obliged to respect an ethical code of conduct – at least, not yet.” Last year, The White House published a “Blueprint for an AI Bill of Rights,” which aims to protect American citizens from dangerous or discriminatory uses of AI. But this text is currently not binding. In Europe, the European Parliament is also exploring very strict obligations in the event of the use of AI in health.
It’s unfortunate that if a senior loved one needs medical care, he or she or their loved ones/caregivers will have to battle an algorithm as to whether the loved one deserves the prescribed care. It’s nerve-wracking to think that those in some Medicare Advantage plans could be informed that, according to the judgment of a statistical tool, the treatment that their doctor says is necessary will not be covered!
Consider Switching Back to Original Medicare While You Are Still Healthy
As mentioned in part one of this series, the good news is that you can switch from a Medicare Advantage plan back to Original Medicare during the annual open enrollment period which runs from October 15 to December 7 each year. Refer to part one for more information on this!
To make your life much easier, consider working with an experienced Medicare health insurance agent who can help you shop for a Medicare Supplement policy, and help you get the best possible policy at the least possible cost. At the Farr Law Firm, we work with Retirement & Medicare Together to serve the Medicare needs of our clients. If you are turning 65 and getting ready to enroll in Medicare, click here to schedule your no-cost Medicare consultation today. Also, if you have not done Long-Term Care Planning, Incapacity Planning, or Estate Planning (or if your Planning documents are more than five years old), call us here at the Farr Law Firm to make an appointment for an initial consultation:
Northern Virginia Medicare Planning: 703-691-1888
Fredericksburg, Medicare Planning: 540-479-1435
Rockville, MD Medicare Planning: 301-519-8041
Annapolis, MD Medicare Planning: 410-216-0703
Washington, DC Medicare Planning: 202-587-2797
About Evan H Farr, CELA, CAP
