A Dark Side of Mixing AI and Medicine

(This originally ran at Empire Financial Research.)

If there's one concern about mixing artificial intelligence ('AI') and medicine, it's that it will over diagnose and lead to under?and?over treatment...

But another concern outside of specialized health care publications isn't getting much attention...

It's that unregulated AI is starting to be used as a tool in insurance pre-authorizations – in some cases denying insurance coverage for people who need it.

Nowhere is that more obvious than with Medicare Advantage's approach to authorizing nursing home stays after surgery or some other acute treatment.

Medicare Advantage is a popular option among Medicare recipients because it's cheaper and more inclusive than traditional Medicare. Think of it as a health maintenance organization ("HMO") that covers 31 million people, or roughly half of all Medicare recipients.

This gets us to our story...

An investigation by health care news site STAT shows how an AI-driven program from a small company called naviHealth – which is owned by UnitedHealth (UNH), the largest provider of Medicare Advantage plans – has been used for nursing home coverage.

Under the headline, "Denied by AI: How Medicare Advantage plans use algorithms to cut off care for seniors in need," the story is a warning of AI run amuck....

Behind the scenes, insurers are using unregulated predictive algorithms, under the guise of scientific rigor, to pinpoint the precise moment when they can plausibly cut off payment for an older patient's treatment. The denials that follow are?setting off heated disputes between doctors and insurers, often delaying treatment of seriously ill patients who are neither aware of the algorithms, nor able to question their calculations.

Older people who spent their lives paying into Medicare, and are now facing amputation, fast-spreading cancers, and other devastating diagnoses, are left to either pay for their care themselves or get by without it. If they disagree, they can file an appeal, and spend months trying to recover their costs, even if they don't recover from their illnesses.

And it's not just STAT that is raising red flags...

Back in October, in conjunction with nonprofit Kaiser Health,?Fortune?raised similar concerns, writing (emphasis added)...

As Medicare Advantage enrollment has spiked in recent years, [Dr. Rajeev] Kumar said, disagreements between insurers and nursing home medical teams have increased.

In addition, he said,?insurers have hired companies, such as Tennessee-based naviHealth, that use data about other patients to help predict how much care an individual needs in a skilled nursing facility based on her health condition. Those calculations can conflict with what medical teams recommend, he said.

Not surprisingly, nursing homes are crying foul.?Fortune?quoted an executive of the nursing home trade association as saying...

Since the advent of these companies, we've seen shorter lengths of stays.

Which ironically is exactly what naviHealth was originally designed to do, since so-called "post-acute care" has been considered by many studies to be – as the?New York Times?put it back in 2013 – "the most financially wasteful."

NaviHealth was founded by Tom Scully, a former head of Centers for Medicare and Medicaid Services under President George H.W. Bush...

While it never mentions AI, in various news releases naviHealth refers to using "predictive and decision support" so "patients can enjoy more days at home, and health care providers and health plans can significantly reduce costs specific to unnecessary care and readmissions."

Speaking on a podcast in 2020, STAT quotes Scully as saying...

Look, I love the nursing home guys, but there were a lot of patients coming out of hospitals spending 20 days in a nursing home in MA... It was just like Pavlov's bell.

Sounded good in principle, and there's no question that the health care system is full of waste...

And plenty of people would rather be at their own homes than a nursing home. But for many, being at home simply isn't realistic...

But in recent years, until being acquired by UnitedHealth, naviHealth has had several owners... so even Scully can't say how its AI has been tweaked or how it's now being used. As STAT reports...

Providers told STAT that as naviHealth was changing hands and enriching its investors, they started noticing an increase in denials under its contracts – that the pendulum had now swung too far in the other direction in an effort to prevent overbilling and make sure patients weren't getting unnecessary services.

And in an ironic twist, as STAT recently reported, even one of naviHealth's original backers is "publicly bashing" the company...

The company told Medicare officials last month that care management firms like naviHealth have pushed "improper" denials of Medicare-covered services, like the inpatient rehab and long-term care Select Medical's hospitals offer. As a result, insurers "are engaging in misleading marketing when they advertise to beneficiaries that they will receive their full Medicare benefits upon enrollment in a MA plan."

The good news is that there has been enough blowback that the government has proposed new rules that say Medicare can't deny coverage "based on internal, proprietary, or external clinical criteria not found in traditional Medicare coverage policies."

NaviHealth didn't respond directly to STAT's questions about the use of its algos...

But according to the publication...

A spokesperson for the company said in a statement that its coverage decisions are based on Medicare criteria and the patient's insurance plan. "The naviHealth predict tool is not used to make coverage determinations," the statement said. "The tool is used as a guide to help us inform providers, families and other caregivers about what sort of assistance and care the patient may need both in the facility and after returning home."

In other words, naviHealth didn't answer the questions.

The bottom line is this... When it comes to saving money, on the nondiagnostic side of health care, the allegations of what has happened here are likely to be a preview of what's to come.

Keep in mind this is all AI that has nothing to do with generative AI like ChatGPT.

You can only imagine what's coming next.

As always, feel free to comment below or email me at [email protected]. I look forward to hearing from you.