In 1986, Congress enacted a law to bar hospitals from turning away patients who are unable to pay. Any hospital with an emergency room that participates in federal health programs must evaluate and stabilize every patient who comes through the door, including those who are uninsured, indigent, addicted to drugs, or mentally ill.
No institution has a similar obligation to ensure that those people have a safe place to sleep. As a society, we’ve effectively decided that people shouldn’t die on the street, but it’s acceptable for them to live there. There are more than half a million homeless in the U.S., about a third of them unsheltered—that is, living on streets, under bridges, or in abandoned properties. When they need medical care or simply a bed and a meal, many go to the emergency room. That’s where America has drawn the line: We’ll pay for a hospital bed but not for a home, even when the home would be cheaper.
Jeffrey Brenner is trying to move that line. He’s a doctor who for more than 25 years has worked largely with the poor, many of them homeless. Recently, his place in the health-care system has shifted. After decades in shoestring clinics and nonprofits, he’s become an executive at UnitedHealth Group Inc., America’s largest health insurer. Brenner is expected to contribute to its bottom line. He plans to do it by giving people places to live.
The research and development lab for this experiment is a pair of apartment complexes in a down-at-the-heels corner of Phoenix called Maryvale. Here, Brenner is using UnitedHealth’s money to pay for housing and support services for roughly 60 formerly homeless recipients of Medicaid, the safety-net insurance program for low-income people. Most states outsource their Medicaid programs to private companies such as UnitedHealth, paying the insurer a per-head monthly fee—typically $500 to $1,000—to manage members’ care. The company’s 6 million Medicaid members produced $43 billion in 2018, almost 20% of total revenue.
It’s a profitable business overall. But the most expensive patients, who often present a complex blend of medical, mental health, and social challenges, cost UnitedHealth vastly more than it takes in to care for them. “Can you imagine people living on the street with these disorders? Heart failure, COPD. They’re rolling around with oxygen tanks, crazy stuff,” Brenner says. It isn’t hard to find people living in similar distress around Phoenix or any other American city. And despite their extreme costs, these patients often get poor care. “This is just sad. This is just stupid,” Brenner says. “Why do we let this go on?”
Sitting in a vacant studio apartment on the second floor of one of the complexes, Brenner shows me data on a patient named Steve, a 54-year-old with multiple sclerosis, cerebral palsy, heart disease, and diabetes. He was homeless before UnitedHealth got him into an apartment. In the 12 months prior to moving in, Steve went to the ER 81 times, spent 17 days hospitalized, and had medical costs, on average, of $12,945 per month. In the nine months since he got a roof over his head and health coaching from Brenner’s team, Steve’s average monthly medical expenses have dropped more than 80%, to $2,073.
After testing the idea in Phoenix, Milwaukee, and Las Vegas, UnitedHealth is expanding Brenner’s housing program, called MyConnections, to 30 markets by early 2020. It’s a business imperative. In January, after the company announced a $12 billion profit for 2018, Wall Street analysts pressed Chief Executive Officer Dave Wichmann on the performance of its Medicaid business. The return, he acknowledged, was “not at our target margin range of 3% to 5%.” Wichmann said it would hit the target next year.
Patients like Steve wind up in the ER because they don’t fit into the ways we deliver health care. The U.S. system is engineered to route billions of dollars to hospitals, clinics, pharmacies, and labs to diagnose and treat patients once they’re sick. It’s not set up to keep vulnerable people housed, clothed, and nourished so they’ll be less likely to get sick in the first place.
The U.S. spends 18% of its gross domestic product on health care, vs. 8.6% in the other 35 countries in the Organization for Economic Cooperation and Development. America’s outsize spending on health care contrasts with much paltrier investments in social support—housing, food, education, cash assistance, and care for children and the elderly. Other nations in the OECD spend $2 on social services for every $1 they spend on health care, according to The American Health Care Paradox, a 2013 book by Elizabeth Bradley and Lauren Taylor. In the U.S., each dollar of health spending is matched by only 60¢ of social support.
That a for-profit conglomerate like UnitedHealth is in the business of taking taxpayer money to care for poor people reflects the peculiarity of U.S. social policy. Medicaid was created in 1965 in tandem with Medicare—public insurance for older Americans. Congress has since expanded eligibility for Medicaid, most recently through the Affordable Care Act, and the program now insures 72 million people, more than 1 in 5 Americans. It pays for 42% of all births.
States split the cost of Medicaid with the federal government, but it takes up an ever-larger portion of their budgets—after education, it’s usually a state’s biggest expense. To keep down costs and avoid the difficulty of running a health-care system, most states contract with UnitedHealth and its competitors to establish what are called Medicaid managed-care programs. In 2017, $264 billion, almost 50¢ of every Medicaid dollar, went toward care for the 54 million people on private Medicaid plans.
Few entities outside the government exert as much influence over health care as UnitedHealth, based in Minnetonka, Minn. The company’s health-insurance unit, UnitedHealthcare, provides benefits to 43 million Americans. About 50,000 physicians work for its health-services unit, Optum Inc. UnitedHealth also owns pharmacies and a bank and Brazilian hospitals. Its revenue last year, $226 billion, surpassed that of all but five U.S. companies; it’s told shareholders to expect long-term earnings growth of 13% to 16% annually.
Brenner, a smiley and cerebral 50-year-old, is an unlikely insurance company man. He studied neuroscience at Robert Wood Johnson Medical School in New Brunswick, N.J., and anticipated a career in research. After a stint at a free student-run clinic that served homeless people and undocumented Central American refugees, he switched to the less prestigious field of family medicine. He did his residency in Seattle and then moved in 1998 to Camden, N.J., at the time the poorest city in the U.S. Brenner started at a small practice with three exam rooms and eventually split off to practice solo. Almost all his patients were on Medicaid. He’d get up in the middle of the night to deliver babies.
Brenner also treated victims of violent crime, which led to an interest in developing an accurate picture of Camden’s crime. It wasn’t going to come from the city government, he learned, because so many victims didn’t file police reports. He went to the hospitals instead.
The data he saw there illuminated a gross imbalance in health-care spending: A tiny sliver of patients accounted for a large part of spending. In Camden, 1% of patients made up 30% of the cost. Brenner spotted patients who went to the ER hundreds of times a year, including a handful of individuals who cost the system millions of dollars each. “Like, for 1% of the spending here, we could open up 10 primary-care offices all over the city,” Brenner says.
He had to shutter his solo practice when he was unable to sustain it on Medicaid’s payment rates. (Medicaid pays doctors and hospitals about 30% less than Medicare does; Medicare in turn pays significantly less than private insurers.) Meanwhile, hospitals were expanding. “The system had become so distorted that it felt like a microcosm of what was going on in America, which is if you don’t take good care of people, they’ll get sick,” Brenner says. “Then you’ll need more hospital beds and hospitals to take care of them.”
In 2002 he founded the nonprofit Camden Coalition of Healthcare Providers. The group used hospital claims data to identify outlier patients and hot spots of medical spending, then tried to help people before they landed in the most costly settings, ERs and hospital beds. That work brought Brenner national prominence, including a New Yorker profile by Atul Gawande, the surgeon and MacArthur “genius” grant recipient, in 2011. Two years later, Brenner received a MacArthur fellowship himself.
UnitedHealth supported the nonprofit and eventually approached Brenner about a job helping the company with its own strategy to address patients’ social needs. “I said no, and said no a couple of times,” he says. But in 2017, convinced that UnitedHealth’s commitment was serious, he joined to test his ideas on a vastly larger stage. The company has 80 times as many Medicaid members as Camden has people.
Brenner, whose title is senior vice president for clinical redesign, manages a staff of 65. The team was a bit larger before a recent broad round of company layoffs; UnitedHealth says the reduction won’t affect the housing program. By early next year the company expects to house 350 homeless Medicaid patients whose annual health-care spending, while they’re on the streets, exceeds $17 million. The goal is for them to “graduate” within a year to paying their own rent. (Most get a disability check; those who don’t get help from MyConnections to apply.)
Insurers, including UnitedHealth, generally try to reduce costs by restricting medical care. They require prior authorization for expensive procedures, deny claims for care deemed inappropriate, and limit the drugs available on prescription plans. This is partly why the industry has a bad reputation—the perception that insurers are middlemen that profit by withholding needed care without adding value. It’s behind the argument Senators Bernie Sanders and Elizabeth Warren make for replacing private insurance with “Medicare for All.”
Brenner aims to reduce expenses not by denying care, but by spending more on social interventions, starting with housing. How to do it is still largely uncharted. “I don’t think we’ve figured any of this out,” he says. “We’re at a hopeful moment of recognizing how deep the problem is.” A trip to any big-city ER reveals the magnitude of the challenge.
Kara Geren is trained to detect what’s about to kill you. The 40-year-old attending physician pulls eight-hour shifts in the emergency department at Valleywise Health Medical Center, a 325-bed public hospital north of the Phoenix airport. The unit has a low dropped ceiling, Formica countertops, and a motley collection of curtains that separate beds packed close together. Geren has the kind of calm yet focused demeanor you’d hope to encounter if you found yourself wheeled into the ER. She isn’t rushing, nor is she wasting any time.
“In emergency medicine you always assume the worst,” Geren says. “What’s going to kill this person in the next five minutes? What’s going to kill this person in the next hour?” Valleywise has two trauma bays and a landing pad for medevac helicopter ambulances. As a Level 1 trauma center, it has to be prepared for any unexpected medical crisis that might arrive at any hour of any day.
That vigilance makes it one of the most expensive places to get health care, and many patients who visit the Valleywise ER shouldn’t be there. Some are immigrants who don’t know how to navigate the U.S. system, so they walk into the hospital for routine treatment. Some are uninsured, so other doctors won’t see them. Some come to get out of the summer heat; temperatures in Phoenix can top 100F for weeks on end. The city’s growing homelessness crisis exacerbates the burden. The number of unsheltered homeless people in Maricopa County, which includes Phoenix and its suburbs, has almost doubled since 2016, to about 3,200.
Some patients are combative, especially if they come in drunk or high. Others are simply seeking shelter and a meal, and complaining of chest pain at an ER is a sure way to get both. Frequent flyers, as nurses and doctors call them, may visit a few times a week or daily. “Sometimes in the same shift, you’ll have a patient come back who you discharged a few hours earlier,” says Heather Jordan, Valleywise’s nursing director for emergency services. “They get a medical screening exam and maybe get a sandwich and a Powerade, and they go back out to where they started.”
Homeless patients have few good options when they’re ready for discharge. Sometimes the hospital pays to send them in taxis to city shelters, which are often full when they arrive. Some go to behavioral health centers for further treatment of mental illness or substance-use disorders. Others go to a respite center run by a nonprofit called Circle the City, where they get medical care along with a bed in a shared dormitory. There are never enough beds to meet demand.
Some people who no longer require hospital care stay at Valleywise simply because more appropriate quarters aren’t available. “There’s a couple of patients who live upstairs that have been here for months and months and months, because we can’t find a place, a safe place, to put them,” Jordan says.
The cost for their care—$3,825 a day—is paid by Medicaid or, for those with no insurance, absorbed by the public hospital and ultimately the taxpayers who fund it. “We could put them in a residence for a fraction of that, and then we can keep ourselves available for that burn patient, that ICU patient, the people, the patients that need us critically,” says Kris Gaw, chief operating officer for Valleywise Health.
Valleywise has been able to place a small handful of homeless patients with MyConnections in Maryvale. The developments were known for drugs and prostitution before UnitedHealth and its nonprofit partner, Chicanos Por La Causa, took them over a couple of years ago. The insurer gave the nonprofit a $21 million low-interest loan to purchase, rehab, and manage the 500 units. Fixing it up was a challenge. One property manager says she got death threats for evicting drug dealers. Eventually, the frequency of police calls dropped sharply, and kids started playing in the courtyards and using the pools.
Most of the apartments rent to the public at market rates, starting at $609 a month for a studio. But up to 100 units are set aside for formerly homeless UnitedHealth Medicaid members. One empty studio with new wood floors at the end of a row on the second story is an office for five “health coaches.” They serve as case managers, counselors, and companions who look after the patients in the program.
One of the coaches, Ray Torres, 50, used to work as a case manager at a county-run clinic for the homeless. Some of his current clients are people he knew from his old job. He’d refer them to services, but they’d frequently just disappear back onto the streets. “Here, we’re on-site, we connect them, we knock on doors,” he says. Torres keeps the medical appointments for his 18 clients in his calendar. He calls taxis for them and occasionally goes with them to the doctor. Sometimes a knock on the door is critical. The week before we spoke, one client had forgotten about an appointment for kidney dialysis. The man had no phone, and Torres’s check-in likely prevented him from going into kidney failure in his apartment.
Torres and his colleagues bring a reservoir of patience deeper than what the homeless typically encounter. Much of the U.S. social safety net conditions assistance on certain behaviors, in an effort to inspire or force people to change. In homeless shelters, people are often required to earn privileges such as a locker or a larger space, eventually to be rewarded with placement in a group home or further housing assistance. Many programs are predicated on first kicking drug habits or adhering to medication. If people act out, they may end up back on the streets. “It’s a little like playing Sorry,” Brenner says. “You go back to the beginning and start over again.”
Brenner, by contrast, advocates a model known as Housing First, which recognizes that getting off the streets is often a necessary first step for people to adhere to treatment for addiction or mental illness—not the other way around. Many of the patients he’s concerned with have experienced early trauma, which has lasting health consequences. Exposure to adverse childhood experiences is a strong predictor of problems such as chronic illness, obesity, smoking, substance abuse, and, not incidentally, health-care spending.
“There’s a whole thread in health care around personal responsibility that this work evokes in people. As though scolding them, they’re going to go, ‘Oh, you’re absolutely right,’ ” Brenner says. “All of these things that we talk about, you know, people not taking personal responsibility—things happen to people. And what we’ve learned is that if you’re very young and you’re exposed to toxic stress, that brain formation is very different. The way that you navigate the world is different. Literally some of your circuits are different.”
One of Brenner’s greatest challenges is deciding who should benefit from the program. Giving patients housing sounds beguilingly simple, but the economics are a high-wire act. Medicaid isn’t paying UnitedHealth anything directly for housing assistance. The company spends from $1,200 to $1,800 a month to house and support each member, so it must save at least that much to break even on Brenner’s program.
On average about 60 members are enrolled in the Phoenix sites at any given time. Once a week, Brenner and his team get on the phone to evaluate potential candidates—anywhere from 2 to 14 people whose names have surfaced in UnitedHealth’s data. They want patients who are homeless and whose medical spending exceeds $50,000 annually, with most of that coming from ER visits and inpatient stays. People living on the streets with less extreme medical costs may need a home just as much, but it doesn’t pay for UnitedHealth to give them one.
For patients above the $50,000 threshold, the reductions in medical costs should let the company at least break even on its investment in housing and services. But it’s not as simple as running the numbers. Brenner is looking for people who not only need help but are ready to accept it. “We want a storyline around, Why is the housing going to make a difference? What’s going on in there? And then what’s the exit strategy?”
It’s a difficult judgment, made more complicated by a statistical concept called reversion to the mean. Simply put, an outlier will tend to go back to the average over time. Some of the most expensive homeless patients spontaneously become less expensive. Maybe they move in with family or get help from another program; maybe they stop visiting hospitals after being mistreated. Brenner says that his team doesn’t fully understand the phenomenon and that the rate at which spending on high-cost patients declines is different in each city. Either way, the housing units he’s allocating are scarce resources, and he doesn’t want to give them to people who would have reduced spending on their own.
He also wants to make sure the program actually does help people reduce their hospital use, and it doesn’t work on everyone. Some people resist it and continue going to ERs even after UnitedHealth puts them in housing. Brenner shows me an analysis of the first 41 patients in Phoenix to get the intervention. The housing and support services proved cost-effective for the 25 most expensive patients, reducing their overall costs dramatically. For the other 16, total spending increased. “The return’s only going to work out if we target the right people,” Brenner says. That’s why UnitedHealth is starting with just 10 subsidized apartments in each new city where it’s introducing the program, even in places where there might be hundreds of homeless Medicaid members on its rolls.
Brenner’s bet is that he can break the cycle for people like Cathy, a 56-year-old who was homeless for several years. She remembers “moving around like a giant turtle,” with her belongings stuffed into bags latched to her electric wheelchair, which she’d plug in to charge overnight at the Sun Devil Auto repair shop in downtown Phoenix. For months, she visited ERs almost daily. One night she left St. Joseph’s Hospital after eight hours and went directly to another emergency department a few miles away. “I was going to keep going every day if I had to, because I was having pain in my chest, and they couldn’t tell me why,” says Cathy, who asked that her last name be withheld.
Her long list of ailments includes diabetes and asthma. A heart attack left her with a stent, and a series of infections almost claimed her foot. That’s on top of depression, post-traumatic stress disorder, and what she describes as “extreme anger issues.” Two years ago, Cathy moved into a subsidized apartment in Phoenix. Torres has witnessed her transformation. “She had that wall put in front of her,” he says. “She had no trust with anybody.” Now the two share wry jokes. “Ray kind of kept trying to be positive, be all sweet and nice, like he is,” Cathy says.
Housing hasn’t solved all her problems. She still has depression, and another heart attack left her hospitalized again earlier this year. But it’s made a profound difference. For one thing, she no longer makes a stop at the ER part of her regular routine. That’s good news for UnitedHealth. And then there’s this: “I feel human again,” she says. “Before, I didn’t.”