ProPublica: How UnitedHealth’s Playbook for Limiting Mental Health Coverage Puts Countless Americans’ Treatment at Risk

By: Annie Waldman | November 19, 2024

For years, it was a mystery: Seemingly out of the blue, therapists would feel like they’d tripped some invisible wire and become a target of UnitedHealth Group.

A company representative with the Orwellian title “care advocate” would call and grill them about why they’d seen a patient twice a week or weekly for six months.

In case after case, United would refuse to cover care, leaving patients to pay out-of-pocket or go without it. The severity of their issues seemed not to matter.

Around 2016, government officials began to pry open United’s black box. They found that the nation’s largest health insurance conglomerate had been using algorithms to identify providers it determined were giving too much therapy and patients it believed were receiving too much; then, the company scrutinized their cases and cut off reimbursements.

By the end of 2021, United’s algorithm program had been deemed illegal in three states.

But that has not stopped the company from continuing to police mental health care with arbitrary thresholds and cost-driven targets, ProPublica found, after reviewing what is effectively the company’s internal playbook for limiting and cutting therapy expenses. The insurer’s strategies are still very much alive, putting countless patients at risk of losing mental health care.

Optum, its subsidiary that manages its mental health coverage, is taking aim at those who give or get “unwarranted” treatment, flagging patients who receive more than 30 sessions in eight months. The insurer estimates its “outlier management” strategy will contribute to savings of up to $52 million, according to company documents.

The company’s ability to continue deploying its playbook lays bare a glaring flaw in the way American health insurance companies are overseen.

While the massive insurer — one of the 10 most profitable companies in the world — offers plans to people in every state, it answers to no single regulator.

The federal government oversees the biggest pool: most of the plans that employers sponsor for their workers.

States are responsible for plans that residents buy on the marketplace; they also regulate those funded by the government through Medicaid but run through private insurers.

In essence, more than 50 different state and federal regulatory entities each oversee a slice of United’s vast network.

So when a California regulator cited United for its algorithm-driven practice in 2018, its corrective plan applied only to market plans based in California.

When Massachusetts’ attorney general forced it to restrict the system in 2020 for one of the largest health plans there, the prosecutor’s power ended at the state line.

And when New York’s attorney general teamed up with the U.S. Department of Labor on one of the most expansive investigations in history of an insurer’s efforts to limit mental health care coverage — one in which they scored a landmark, multimillion-dollar victory against United — none of it made an ounce of difference to the millions whose plans fell outside their purview.

It didn’t matter that they were all scrutinizing the insurer for violating the same federal law, one that forbade companies from putting up barriers to mental health coverage that did not exist for physical health coverage.

For United’s practices to be curbed, mental health advocates told ProPublica, every single jurisdiction in which it operates would have to successfully bring a case against it.

“It’s like playing Whac-A-Mole all the time for regulators,” said Lauren Finke, senior director of policy at the mental health advocacy group The Kennedy Forum. The regulatory patchwork benefits insurance companies, she said, “because they can just move their scrutinized practices to other products in different locations.”

Now internal documents show that United, through its subsidiary Optum, is targeting plans in other jurisdictions, where its practices have not been curbed. The company is focused on reducing “overutilization” of services for patients covered through its privately contracted Medicaid plans that are overseen by states, according to the internal company records reviewed by ProPublica. These plans cover some of the nation’s poorest and most vulnerable patients.

Internal company documents obtained by ProPublica reveal the strategy by Optum, a UnitedHealth Group subsidiary, to scrutinize and reduce outpatient mental health care. Credit:Obtained by ProPublica

United administers Medicaid plans or benefits in about two dozen states, and for more than 6 million people, according to the most recent federal data from 2022. The division responsible for the company’s Medicaid coverage took in $75 billion in revenue last year, a quarter of the total revenue of its health benefits business, UnitedHealthcare.

UnitedHealthcare told ProPublica that the company remains compliant with the terms of its settlement with the New York attorney general and federal regulators. Christine Hauser, a spokesperson for Optum Behavioral Health, said its process for managing health care claims is “an important part of making sure patients get access to safe, effective and affordable treatment.” Its programs are compliant with federal laws and ensure “people receive the care they need,” she said. One category of reviews is voluntary, she added; it allows providers to opt out and does not result in coverage denials.

ProPublica has spent months tracking the company’s efforts to limit mental health costs, reviewing hundreds of pages of internal documents and court records, and interviewing dozens of current and former employees as well as scores of providers in the company’s insurance networks.

One therapist in Virginia said she is reeling from the costly repercussions of her review by a care advocate. Another in Oklahoma said she faces ongoing pressure from United for seeing her high-risk patients twice a week.

“There’s no real clinical rationale behind this,” said Tim Clement, the vice president of federal government affairs at the nonprofit group Mental Health America. “This is pretty much a financial decision.”

Former care advocates for the company told ProPublica the same as they described steamrolling providers to boost cost savings.

One said he felt like “a cog in the wheel of insurance greed.”

Under ALERT

The year 2008 was supposed to mark a revolution in access to mental health care.

For decades, United and other insurers had been allowed to place hard caps on treatment, like the number of therapy sessions. But after Congress passed the ​​Mental Health Parity and Addiction Equity Act, insurers could no longer set higher copays for behavioral services or more strictly limit how often patients could get them; insurers needed to offer the same access to mental health care as to physical care. The law applied to most plans, regardless of whether federal or state regulators enforced it.

As access to services increased, so did insurers’ costs. Company documents show United was keenly aware of this threat to its bottom line.

But there was a loophole: Insurers could still determine what care was medically necessary and appropriate.

Doing so case by case would be expensive and time-consuming. But United already had a tool that could make it easier to spot outliers.

Called ALERT, the algorithmic system was created years earlier to identify patients at risk of suicide or substance use. The company redeployed it to identify therapy overuse.

Company and court filings reveal that ALERT comprised a suite of algorithms — totaling more than 50 at one point — that analyzed clinical and claims data to catch what it considered unusual mental health treatment patterns, flagging up to 15% of the patients receiving outpatient care.

The algorithms could be triggered when care met the company's definition of overly frequent, such as when patients had therapy sessions twice a week for six weeks or more than 20 sessions in six months. Therapists drew scrutiny if they provided services for more than eight hours a day, used the same diagnosis code with most clients or worked on weekends or holidays — even though such work is often necessary with patients in crisis.

The system was originally designed to save lives, said Ed Jones, who co-developed the algorithm program when he worked as an executive at PacifiCare Behavioral Health, which later merged with United. Using ALERT to limit or deny care was “perverting a process that was really pretty good,” he told ProPublica.

Once patients or therapists were flagged, care advocates, who were licensed practitioners, would “alert” providers, using intervention scripts to assess whether care was medically necessary. The calls felt like interrogations, therapists told ProPublica, with the predetermined conclusion that their therapy was unnecessary or excessive.

ProPublica spoke with seven former employees from Optum who worked with the ALERT system from 2006 through 2021. They requested not to be named in order to speak freely, some citing fears of retaliation.

Even though the reviews were purportedly intended to identify cases where care was inappropriate or violated clinical standards, several former care advocates said these instances were rare. Instead, they questioned care if it passed an allotted number of sessions.

“It had to be really extreme to help the client be able to continue with the care,” said one former care advocate, who was troubled by the practice. “Not everyone with depression is going to be suicidal, but they still need therapy to support them.”

The advocates often overruled a provider’s expertise, a former team manager said. “There was always this feeling, ‘Why are we telling clinicians what to do?’” he said. “I didn’t think it was OK that we were making decisions like that for people.”

If the advocates found fault with therapists’ explanations — or couldn’t persuade them to cut back on care — they elevated the case to a peer-to-peer review, where a psychologist could decide to stop covering treatment.

According to court records, regulators alleged United doled out bonuses to care advocates based on productivity, such as the number of cases handled, and pushed workers to reduce care by modifying a therapist’s treatment or referring therapists to peer review in 20% of assigned cases.

At one point, care advocates were referring 40%, regulators alleged in court filings. Each peer review tended to last less than 12 minutes, offering providers little time to prove they had a “clear and compelling” reason to continue treatment.

Former advocates described feeling like parts of a machine that couldn’t stop churning. “Literally, we had to tell the company when we were going to the restroom,” one advocate said, “and so you would do that and come back and your manager would say, ‘Well, that was a little long.’”

The former workers told ProPublica they were pressured to keep calls brief; the rush added to the tension as therapists pushed back in anger.

“There was an expiration date on those jobs because there was such a pull on you emotionally,” one former care advocate said.

Three of them quit, they told ProPublica, citing damage to their own mental health.

New York and federal regulators started looking into the practice around 2016. A California regulator and the Massachusetts attorney general’s office soon followed.

All concluded that while United may not have set official caps on coverage, it had done so in practice by limiting mental health services more stringently than medical care. Therefore, it was breaking the federal parity law.

While California and Massachusetts got United to scale back its use of ALERT within their jurisdictions, New York was able to stretch its reach by teaming up with the U.S. Department of Labor to investigate and sue the insurer. Together, they found that from 2013 through 2020, United had denied claims for more than 34,000 therapy sessions in New York alone, amounting to $8 million in denied care.

By using ALERT to ration care, United calculated that it saved the company about $330 per member each time the program was used, the regulators said in court records. Cut off from therapy, some patients were hospitalized. The regulators did not specifically address in court filings whether the treatment denials met medical guidelines.

The company, which denied the allegations and did not have to admit liability or wrongdoing, agreed to pay more than $4 million in restitution and penalties in 2021. Notably, it also agreed to not use ALERT to limit or deny care.

The final terms of the settlement, however, only applied to plans under New York and federal regulators’ jurisdiction.

This article was originally published on ProPublica.