In October 2014, my father was startled to receive a letter announcing the retirement, in a month’s time, of our family physician. Both he and his doctor were in their late 60s by then, and their relationship went back about 30 years, to the early 1980s, after my father followed his father and paternal grandparents, all from the Midwest, to Southwest Florida. How they began seeing the doctor is beyond memory, but as my father’s grandparents grew increasingly frail, his father frequently drove them to their doctor for checkups. At one of them, in the mid-’80s, the doctor suggested that it might be less strenuous for my great-grandparents if he met them in the parking lot. From then until they died, he came downstairs from his seventh-floor office with his black bag and climbed into the back seat of their yellow Oldsmobile 88 to give them their physicals.
More than a decade later, my grandfather, who was 78 by then and had chronic obstructive pulmonary disease, developed pneumonia. When he was admitted to the county hospital, it was the same doctor — now my doctor, too — who came to his room to describe his options: go on a ventilator and go home, where he would most likely remain bedridden for the remainder of his life (probably months), or let the illness run its course and die within a few days. My grandfather asked if the latter would hurt much. The doctor said no, he could make him comfortable. All right, then, my grandfather replied.
By the time the doctor abruptly tendered his retirement notice, I had long since moved away. I had lived with high-deductible health insurance for years and, after a few surprise bills I could not afford, I avoided medical professionals whenever possible. Often, when I didn’t have a primary-care physician or couldn’t remember his or her name (which changed with my ever-changing insurance policy), I filled out forms with my hometown doctor’s name instead: George P. Fitzgerald III, M.D. But until my father told me about his letter, I had not reflected on his role in my life. I’d rarely seen him more than once a year, and any problems I did have were unremarkable, easily treatable by someone else. And yet his departure felt like a loss.
My father, after his final checkup, reported why the doctor was leaving: He objected to the increasing industrialization of medical care, in particular a new Medicare mandate that he update patients’ medical histories electronically — not because he was old-fashioned or contrary (though he was both), but because the idea of dividing his attention between a computer screen and a patient offended him. This shift to electronic health records, intended to improve communication between a patient’s various physicians as well as “automate and streamline provider workflow,” was the latest in a series of incentives that the health-insurance market created over the past 20 years to maximize efficiency. In this payment system, the time-consuming process of cultivating a relationship with a patient is essentially worthless, because doing so is not a finite, billable procedure. For the doctor, however, the fourth generation of his family to practice medicine, it was vital to his craft.
The question of what the role of a primary-care physician should be, and how it should be valued, has perhaps never been more urgent. That figure, typically a general practitioner, family doctor or internist, is a patient’s first and often most personal connection to the rest of the health care system. But well-known corporations are betting that Americans would prefer to have health care “delivered” by a trusted brand rather than a trusted physician. At least 10 of the nation’s largest tech companies, including Apple, Microsoft and Alphabet, Google’s parent company, have designed tools and software for medical use and increased their involvement in health care start-up deals in recent years; the size of those deals went to nearly $3 billion in 2017 from under $300 million in 2012. Amazon, in partnership with Berkshire Hathaway and JPMorgan Chase, has announced its intent to disrupt the industry — among other ways, by reinventing primary-care delivery. Walmart now offers primary-care services in some of its stores.
Surprisingly little is known, though, about what the relationship between a patient and his or her primary-care doctor is actually worth, in terms of that patient’s overall well-being or medical costs, regardless of who bears them. In fact, David Meltzer, an economist and a primary-care physician at the University of Chicago, may be the first and only researcher in the country trying to quantify that relationship’s value in a randomized clinical trial, the most rigorous scientific method.
“I’m a complete believer that it’s going to change health care in America,” he told me nearly two years ago. Meltzer continues to believe as much, now that his trial has produced preliminary results, which he plans to present at the end of June at the annual research meeting of AcademyHealth, a health-policy organization. His experiment suggests that strengthening the patient-doctor relationship can significantly reduce patients’ hospitalizations and expenses (to hospitals and Medicare) and improve their mental health; it remains to be seen whether that will be enough to influence the plans that corporations, drawing on their expertise in efficiency and customer service, have to redesign the health care system.
In 2016, I spent three weeks over four months at the University of Chicago Medical Center observing Meltzer’s experiment. Its centerpiece was a primary-care practice that consisted of five physicians, a small staff and, at its peak, 1,000 high-risk Medicare patients. I saw more than 90 patient-doctor interactions there; in every instance, the university’s media-relations manager made sure that the patients understood they could quite reasonably refuse me access. Yet almost everyone allowed me to sit in and write about whatever took place. The receptiveness to the presence of a stranger — potentially representing millions of other strangers — in these private moments puzzled me but did not seem to surprise Meltzer. Only much later did the question of why they said yes come to seem central to both his project and mine.
How much distance doctors should maintain from their patients is a debate that goes back at least as far as the invention of the stethoscope in 1816. It was among the first tools that made a physician privy to observations that were inaccessible to the patient; the device physically separated them by eliminating the doctor’s need to press an ear directly to a patient’s chest. This special knowledge and apartness earned new respect for what was then a trade. After the Civil War, as transportation improved and more people moved to cities, often without their families, agrarians and urbanites alike increasingly turned to doctors instead of neighbors or relatives for help when they were ill. By the turn of the century, Americans’ growing psychological dependence on doctors — whose actual capabilities were still extremely limited — gave those doctors “authority,” Paul Starr, the Princeton sociologist, writes in his 1982 book, “The Social Transformation of American Medicine.” Political organizing enabled doctors to leverage this authority into what he describes as “legal privileges, economic power, high incomes and enhanced social status.”
But not long after physicians coalesced into a professional class, advances in research and technology made it increasingly untenable for any practitioner to master the entire medical canon. At the same time, many experienced physicians argued that medical schools were churning out experts in ever-narrower portions of the body who lacked empathy for the whole person. In 1927, Francis Peabody, a professor at Harvard Medical School, famously defined the trade-off in The Journal of the American Medical Association. “The treatment of a disease may be entirely impersonal; the care of a patient must be completely personal,” he wrote. “The significance of the intimate personal relationship between physician and patient cannot be too strongly emphasized, for in an extraordinarily large number of cases both diagnosis and treatment are directly dependent on it.”
Ultimately, these unique relationships empowered physicians to reject intervention by any regulatory third party, including national government. But doctors’ control over pricing gave rise to inflated costs and gradually sowed anger and mistrust among the public. Demand for reform led to the passage of Medicare in 1965, but the program, eager to avoid angering doctors and hospitals, embraced the same reimbursement practices of most private insurance plans, arbitrarily assigning flat fees to services, like outpatient checkups, regardless of the benefit to patients. For primary-care doctors — paid the same for a simple exam of a relatively healthy patient as for a more involved consultation with a chronically ill one — this created a perverse incentive: see as many healthy people and as few sick people as possible.
As health-insurance providers prompted physicians to reorganize their patient pools, they simultaneously pushed patients to switch doctors by creating “networks” of physicians whose services were covered. (Those physicians, in turn, agreed to care for patients in accordance with the insurance companies’ policies.) Consequently, and as a result of shifting benefits, as more working Americans signed up for employer-sponsored health insurance, many of them — 41 percent, in one survey — “severed established relationships with their physicians in order to seek new providers,” according to a 1996 study in The American Journal of Public Health. The same study estimated that more than a third of Americans on Medicare who were 65 and older and who had a regular physician had been seeing him or her for a decade or more — and those with the longest ties had lower medical costs and were less likely to be hospitalized than those with the shortest.
Yet concrete proof that the patient-doctor relationship was responsible for these better circumstances remained elusive. “Despite a consensus that longitudinal care is important,” the study’s authors, Linda J. Weiss and Jan Blustein, then of the Columbia College of Physicians and Surgeons, wrote, “remarkably little is known about the actual value — in terms of health care processes, outcomes or costs — of long-term provider-patient relationships.” Reliable data on such relationships simply didn’t exist.
“One of the tragedies of American medicine is that the money doesn’t really align with value,” Meltzer told me one afternoon in August 2016. I’d been trailing his doctors for two days by then. Much of what I’d seen, and would continue to see, was as remarkable as it was tedious: The tedium itself seemed noteworthy. During clinic appointments, which typically lasted 30 minutes, and hospital rounds, the doctors appeared to be fascinated by trivial-sounding details and tangents: a patient’s description of the shortcomings of the spaghetti at her physical-therapy center; the stops on another’s bus route; the reasons a third believed an eviction notice violated his civil rights. Only gradually (if ever) did the medical relevance of these conversations reveal themselves. The first patient needed a high-calorie, easy-to-swallow meal supplement; the second, a recovering alcoholic, could avoid passing a liquor store by switching routes; the third was not discharged until his doctor called his building manager to make sure he could get into his apartment and plug in his oxygen.
One afternoon I watched one doctor, Ram Krishnamoorthi, examine L.A. Kizer, who had just turned 70. He was wearing a new-looking ball cap with a trout on it and pulling an oxygen tank. “Did you go fishing?” Krishnamoorthi asked. He had been Kizer’s doctor for more than two years, but he could never tell when such an innocuous question might turn out to be significant: If Kizer said yes, Krishnamoorthi could try using the activity to motivate healthful behavior, or he might find out that someone who could be tapped as a source of support had accompanied Kizer on the outing.
“Used to,” Kizer said. He suffered from congestive heart failure and was still breathing heavily from the trip down the hall.
“Let me see you walk over,” Krishnamoorthi said. “This is a lot worse than before. What happened?”
“I saw you two weeks ago,” Kizer said, “and it feels like I woke up and I can’t move.” He added that he had gained four pounds. Krishnamoorthi rolled up his pants legs and probed his ankles. He pressed a stethoscope to his back. A doctor who had not seen Kizer so recently would have no reason to investigate his sudden decline (and Kizer, who said he didn’t like doctors, might not have told an unfamiliar doctor that anything had changed). “How’d you get here?” Krishnamoorthi asked.
“It’s fluid,” Krishnamoorthi said, referring to Kizer’s added weight and difficulty walking and breathing. He asked if Kizer wanted to be admitted to the hospital, where he could get additional medications; Kizer declined. Krishnamoorthi eventually agreed to let him go home if he turned off his tank, waited for 10 minutes and had his blood-oxygen level checked. This would take up more of Krishnamoorthi’s time — without generating additional reimbursement — but it would mean vast savings for the health care system if it helped him prevent Kizer from having an emergency on the bus later. Heeding Kizer’s preference probably made him more likely to keep future appointments.
“Why am I doing so bad?” Kizer asked as Krishnamoorthi clipped an oxygen meter to his finger. “Is it all the meds?”
“I don’t know. Maybe.” Krishnamoorthi read the meter display: “98 percent. I think you’re safe. Don’t croak on the bus.”
“If I do, you won’t find out.”
“I’ll find out. They call the doctor first. What’d you do for your birthday?” A physician in a hurry might not have noticed the date — or accepted his first answer.
“What I’m doing now. Nothing.”
“Did you have guests over?”
“What did you do?”
“Did you eat chips while you were playing cards?”
“A couple. You know I do.”
“Do you think that might be part of the reason for the weight?” Extra sodium can cause fluid buildup.
What I couldn’t quite believe was that these lengthy conversations were economically sustainable. Nationwide, primary-care physicians often can’t afford to spend more than 15 minutes with each patient, because of the way the health care system values their time. I had seen Krishnamoorthi and colleagues take that long to sort through patients’ medications, often more than a dozen, which they asked patients to gather up and bring to appointments. During one visit, Grace Berry, another doctor, Google-mapped a patient’s new address to locate a nearby pharmacy, transferring her prescriptions there and helping her secure reliable transportation to and from the clinic.
Meltzer insists that doctors spending more time with their patients actually saves money. After a year in his clinic, for instance, patients have 20 percent fewer hospitalizations than their control-group counterparts. Because hospitalizations make up the greatest portion of these patients’ annual cost to Medicare, which averages $50,000 to $75,000 each, that reduction is worth several thousand dollars per person in the first year, or a combined several million dollars; by comparison, the doctors’ annual salaries add up to less than $1 million.
In a traditional primary-care clinic, patients’ medication — prescribed to them by multiple specialists who are not in communication with one another — might not have been systematically sorted by a single person. Or if it had been, it might have been checked by a professional-care coordinator in consultation with doctors. In Meltzer’s view, that increases labor costs, introduces inevitable miscommunications, as in a game of “telephone,” and runs the risk of not seeming as important to the patient as instructions personally delivered by the physician.
By contrast, Meltzer’s doctors run their clinic in partnership with two nurses, a social worker and a project manager; the leanness of the staff means that most of its internal communication can take place during a 30-minute group meeting each morning. That permits more face time between doctors and patients, according to Meltzer. “And face time is possible if you decrease back-room time,” he says. “People fail to recognize that when you have a $200 person here and a $200 person here and a $100 person communicating between them, the real cost of that is all of their time.” In this setup, which describes a coordinator relaying a message from a doctor in one setting to a doctor in another, the cost is $300 per hour of conversation, or $600 total. What Meltzer has done, in effect, is consolidate the first two physicians into one, simultaneously doubling their time with patients and eliminating the need for a go-between, cutting out that $600 cost altogether. To accomplish this feat, he employs the oldest, slowest, least-innovative health care delivery model imaginable: general practitioners who oversee almost all aspects of their patients’ care and, in doing so, come to know them personally.
Meltzer, who is 54, grew up in Chicago in the 1960s and ’70s, attended Yale and then returned home to simultaneously pursue a Ph.D. in economics and a medical degree at the University of Chicago, becoming an assistant professor in its departments of medicine and economics and in its graduate school of public-policy studies in 1996, as well as a primary-care doctor with his own small practice at the medical center. (He still sees many of the patients that he began seeing then.)
In the early 1990s, primary-care doctors typically made daily rounds at their hospitals to oversee the treatment of their patients there. But this was about to change. In 1996, an influential article in The New England Journal of Medicine by Robert M. Wachter and Lee Goldman, then physicians at the University of California, San Francisco, predicted that the “explosive growth of managed care” emphasizing efficiency would lead to “the rapid growth of a new breed of physicians we call ‘hospitalists’ — specialists in inpatient medicine — who will be responsible for managing the care of hospitalized patients in the same way that primary-care physicians are responsible for managing the care of outpatients.”
As an economist, Meltzer read their forecast with interest. The appearance of hospitalists was, he recognized, a textbook example of specialization or “division of labor,” whereby a task is split into simpler ones that workers can perform more quickly, as on an assembly line. In this case, the split eliminated the need for a primary-care doctor to travel to and from the hospital, enabling him to move patients through his clinic faster; meanwhile, the hospitalist, with his experience treating acute illness and getting the most out of his hospital’s unique facilities, would presumably be able to care for patients admitted there more efficiently and effectively. When, soon after the publication of the article, the University of Chicago Medical Center hired its first two hospitalists, Meltzer volunteered to study their impact. His plan was to compare how patients fared when they were cared for by a specialist (a hospitalist) versus a generalist. He began keeping records — which physicians treated which patients, for example — and hired two students to interview patients in the hospital and then again over the phone a month after they were discharged to learn about their health.
On its face, Meltzer’s research and that of others seemed to show that hospitalists were a great boon. They reduced patients’ medical expenses (by nearly $800, he as well as Wachter, Goldman and their colleagues found), shortened their hospital stays (by about half a day) and modestly decreased mortality rates. But the effects of the nascent hospitalist specialty were only being analyzed rigorously at teaching hospitals like Chicago’s, where the attendings (supervising doctors who spend most of their time caring for ambulatory patients) did not know the patients any better than the hospitalists did. No one was investigating how the hospitalist model performed beyond academic settings, where it was replacing patients’ primary-care doctors. Meltzer noticed, though, that when the Chicago hospitalists got the weekend off and an attending took over their patients for just 48 hours, the handoff completely erased the cost benefits the model otherwise seemed to generate. It stood to reason, then, that separating patients in community hospitals from their longtime doctors might be having a significant negative impact.
But running a clinical trial in a community hospital to directly compare the hospitalist model to the primary-care model it was rapidly replacing was widely considered impossible. How many longtime patients of a general practitioner who was still making hospital rounds would agree to swap her for a hospitalist so that researchers could see whether doing so was good for them?
By 2010, the hospitalist model had become the norm. That year, Meltzer published a theorem describing why the model had taken hold so quickly; in the past 15 years, the number of practicing hospitalists had jumped from several hundred to 30,000. (Meltzer practices as a hospitalist himself.) The math itself was basic. But, staring at his final equation, Meltzer had a revelation: increasing one variable — the likelihood that a primary-care physician’s patients would be hospitalized — would decrease the odds that the physician would cede his hospital duties to a hospitalist out of financial necessity, halting, at least mathematically, the specialty’s rapid spread.
This gave Meltzer an idea. In 2012 he won a grant from the Center for Medicare and Medicaid Innovation, an organization created by the Affordable Care Act to test novel “payment and service-delivery models,” to set up an experimental outpatient clinic at the university medical center. For patients, he recruited Medicare recipients who had been hospitalized in the past year and were therefore statistically likely to require future hospitalization. In practice this meant that most of them, like half of Americans, were managing at least one chronic illness; most were also from the impoverished and violent neighborhoods of Chicago’s South Side. Their average annual income at the time of enrollment was $10,000 to $15,000.
If they volunteered, they were told, they would be randomly assigned to either a primary-care doctor for outpatient needs and a hospitalist in the hospital, or they would join Meltzer’s new clinic, which he called the Comprehensive Care Program, and receive inpatient and outpatient care from the same doctor. Meltzer’s team would directly compare the two groups using a medley of metrics: how often patients were hospitalized and for how long; whether they did not refill their prescription medications because of expense; and their total costs to Medicare. A standard survey would assess their overall health. Finally, to try to understand how patients’ relationships with their doctors might differ between the two groups and affect the outcomes, the team devised a survey in which patients rated their doctors in categories like trust and interpersonal relationships. Did he or she know their values and responsibilities? Give clear instructions? Seem caring and kind?
Meltzer’s biggest fear, at first, was that the patients’ existing primary-care physicians would accuse him of stealing them. It turned out that many hospital patients his team approached did not have one. Within four years, the researchers had signed up 2,000 subjects. “In a funny way, the study that we’re doing now was the same study that should have been done in 1999,” Meltzer told me. “The difference is that we could never have gotten people out of having their own doctor until now, when they don’t have it, and we offer it back.”
On a Monday morning in August 2016, I went on hospital rounds with Krishnamoorthi, as he performed the same duties a hospitalist would with one key difference: He already knew the patients. “Because you intensely see them, and by intensely I mean frequently, the relationship building happens faster,” he said. “The awesome part of it is, I don’t have to look stuff up about my patients. I don’t have a perfect memory; it’s just that I’ve seen them a bunch of times. So it’s quicker, it’s faster to just come in their room and say, ‘So, what’s happening this time?’ ” Krishnamoorthi, now 40, wore a thin maroon tie and scuffed brown loafers. His first job was at the Veterans Affairs medical center associated with Loyola University, where he did his residency, as a primary-care physician for patients with mental health problems. He was responsible for at least 700 of them — not many, compared with traditional primary-care physicians, who nationwide manage 2,000 to 3,000 patients apiece. Still, he was in the office until 9 every night, making phone calls and filling out paperwork. He began to worry that he would make a mistake, forget to order a test. After a year, he switched to a job as a teaching hospitalist, supervising residents; the schedule was less extreme, and when he clocked out, his responsibilities and his interactions with the patients he was treating ended. The work was less emotionally draining and more lucrative (hospitalists tend to make about $20,000 more than family physicians) but also felt less meaningful. In 2013, at a party, he was introduced to Meltzer, who was just beginning to staff his clinic. When Krishnamoorthi heard that its five doctors would manage 200 patients each and work regular hours, he decided to apply.
That Monday morning, his first patient had head and neck cancer and needed an M.R.I. Krishnamoorthi had him practice lying flat and slid pillows under his head and adjusted the bed angle until they found a position the man could tolerate in the scanner that afternoon. (If the patient balked, he would have to be hospitalized longer, increasing his risk of developing an infection and reducing the potential profit for the hospital, which earns money when it admits a patient and loses it for every additional day that patient stays.)
Then, Krishnamoorthi jogged upstairs to check on Jamal Johnson, a 34-year-old with cystic fibrosis. Johnson was sitting on the edge of his bed, hunched over his lunch tray. Krishnamoorthi dragged a footstool over and plopped down on it, stretching out his legs. “When I woke up, I felt kind of draggy,” Johnson said. But physical therapy, which consisted of wearing a vest that vibrated his chest to break up sputum, had helped. “See?” Krishnamoorthi said. “It always works.”
“It hurts like hell, though. That’s something I can say.” Five years earlier, Johnson broke a rib playing basketball, making therapy treatments for his cystic fibrosis excruciating and initiating a cascade of hospitalizations. When he met Krishnamoorthi a year ago, he hated him instantly. Krishnamoorthi wanted him to consider a lung transplant; Johnson refused. He didn’t want to hear about further treatment options. When Krishnamoorthi broached the subject of staying in a nursing facility, Johnson told him that the doctors and nurses there were “just paid to like you.” The remark stuck. Krishnamoorthi asked around until he found a nurse he thought Johnson might connect with; eventually, Johnson agreed to enter her facility and came less frequently to the hospital. On occasions when he was admitted, Krishnamoorthi would, with his permission, call his mother with updates — an overture, she later told me, that none of the countless doctors who had treated her son over the years had ever made and for which she was especially grateful. Johnson never wanted to worry her.
The new arrangement was more beneficial for Johnson, less stressful for his mother, less expensive for Medicare and less expensive for the hospital, but it had taken Krishnamoorthi weeks to orchestrate. It was increasingly clear to him how rewarding — but equally taxing — it was to try to do more for his patients than what was absolutely required. “Last year, I would go home worried whether he liked me or not, if he would ‘fire’ me,” he told me after we left Johnson’s room. “It would have affected me a lot.”
The results of the first five years of Meltzer’s study support his theory that strengthening the relationship between patients and their doctors can decrease medical costs and improve patient health. In addition to being hospitalized less than their control-group counterparts, the clinic patients rated their satisfaction with their physicians within the 95th percentile of such scores nationally versus the 80th for the control group; and though their general health status (self-ranked on a scale from 1 to 5) was similar to those in the control group, their mental-health scores were significantly higher.
The next major test of the model will be whether it can be implemented in different sets of financial, geographic and demographic circumstances. Already, more than half a dozen organizations, from a senior-living community in Florida to the National University Health System of Singapore, have consulted Meltzer about creating versions of his program. One of them is the health-insurance company Kaiser Permanente. The Mid-Atlantic Permanente Medical Group, an organization of more than 1,400 primary-care physicians who serve members of Kaiser’s health plan, ran a pilot of the program in 2015. Two physicians were assigned to its 400 sickest patients. Unlike Meltzer’s team, which relies on hospitalists for after-hours care, Kaiser made the doctors available 24/7 and were surprised to find that patients very rarely called late or on holidays and weekends. Still, the number of times they were admitted to the hospital or E.R. fell drastically, by 80 percent, prompting Kaiser to expand the program to seven doctors and 800 patients. In the summer of 2016, Vanderbilt University Hospital started a similar pilot and observed a 20 percent decrease in hospital admissions. The school is also expanding its program.
“All of us, I think, are struggling with how we can provide care to an aging population with resources drying up left and right,” Bernadette Loftus, a clinician and the executive medical director of the Mid-Atlantic Permanente Medical Group, told me. “David and I both came of age at the birth of the hospitalist movement. It’s a great efficiency movement, but it takes about two decades of living in that environment to see the flaws in it.”
Meltzer has broadened his experiment, with a grant from the Robert Wood Johnson Foundation, to offer additional resources with the hopes of reaching his clinic’s patients who rarely engage with its services; the clinic can now refer them to an additional community-health worker and a neighborhood arts-and-culture program designed with Urban Labs, a center at the university that studies approaches to combating health problems and other social challenges. But he believes versions of the model could also apply to primary-care practices with patients who have a more traditional range of incomes and health issues. For instance, to increase patient-doctor familiarity and streamline communication, a practice might partner with a single hospitalist who takes care of all of its hospitalized patients. In Meltzer’s view, his model’s key mechanism, the prioritizing of the patient-doctor relationship, makes it flexible enough to export anywhere. “To have a relationship,” Meltzer says, “is to mean that the big incentive is to want them to do well because you care about them and to have the time to really understand what that means and deal with it, because sometimes figuring that out just takes time. When price is the only lever you have, you’re either paying people to do something or paying them not to do something. And it goes only one way or the other, more or less. Whereas when you have a relationship, there’s just so many more nuances.”
Current metrics are still too crude to capture the intricate effects of these relationships, however, even as they point to their collective impact being quite large. By contrast, day-to-day, patient-to-patient progress has often seemed glacial at best and demanded considerable energy from the clinic doctors and staff. As I observed Meltzer’s team at work, I couldn’t help wondering how long its young, idealistic caregivers would last. Unknown to Meltzer, several of them had lent patients cash for bus money or delivered medications on their way home — gestures they hadn’t felt inspired (or pressured) to make in previous jobs. When, that October, Jamal Johnson passed away, Krishnamoorthi, upon visiting his bedside to say goodbye, was surprised to find he couldn’t stop crying. “At the end of the day, I don’t feel finished,” Lauren Wiklund, the project manager, told me one particularly hectic morning. “I just feel like I have to carry on with the rest of my life or I will die.”
Why, then, did they do it? “I became disenchanted with primary care because I was constantly having to cut people short,” Joyce Tang, who spent seven years at her previous practice, told me. “I was always an hour late. Always. My patients just knew to expect that.” One morning, I watched her examine a man who seemed to prefer the accommodations of the hospital to those of his assisted-living facility and habitually took a cab to the emergency room. Tang had brainstormed with him, his sister, his insurance company and a nurse at the facility, about possible solutions. She was carrying a housing application for a faith-based facility she thought he might like better, but so far, nothing had worked to change his patterns. In cases like his, how did she measure success?
“There are traditional metrics: fewer hospitalizations, better compliance with medicines,” she said. “But then for patients there are more specific goals — so understanding those.” She gave the example of a patient who was immobilized by her weight. “For her, the big intervention was I got her a motorized wheelchair, and the next time I saw her she was just smiling. She said it totally transformed her outlook just to be able to get outside and get some fresh air.” Another woman had been on large doses of opioids. For a year she and Tang saw each other every two weeks and gradually reduced her dose until, recently, she flushed the rest of her pills down the toilet. “Now when she comes to the hospital, she doesn’t want IV pain meds,” Tang said. “She’s so happy. She’s really proud of herself. That’s a big success story. A lot of people don’t come back when we say ‘taper.’ I asked her, ‘What made you come back?’ She hated me when I said that. But she said: ‘I trusted you. I knew you wanted what’s best for me.’ ”
Velma Winslow was 89 when she died in the hospital in the middle of the night after her doctor, Anshu Verma, went home. Several times, I had seen Verma crouch beside Winslow’s bed when it was long past her ability to respond and speak in a bright voice: “Hi, Ms. Winslow! Can you look over here at me? Hey, Velma?”
The morning after Winslow’s death, Meltzer hosted a weekly meeting with the clinic and research teams. Once everyone had settled in with a paper-wrapped sandwich, he said, “Case of the week?” and the conversation turned to Winslow. She had been hospitalized with minor injuries after a fall and developed an infection; together, her son and Verma had wrestled with when to stop treating it. “Her son last night asked: ‘Do you think I’m doing the right thing? What do you think I should do?’ He said, ‘I know you care about her, so I trust you.’ Not, ‘I trust you because I know you’re a doctor, or I know you’re a professional’ — and that doesn’t happen with a hospitalist.”
“It’s really amazing that the struggle was more with your own feelings,” Meltzer said, and Verma agreed.
By then, I had watched her and her colleagues confront a breathtaking variety of indignities that age and sickness force upon human beings: impotence, incontinence, immobility, dementia, death. It occurred to me that perhaps the explanation for patients’ allowing me to witness these conversations was as simple as wanting, even at their most vulnerable — perhaps especially then — to be seen and heard, and that this was also precisely what they prized in Meltzer’s doctors. What both doctors and patients were after, it seemed, was not just an assignation of “value” but of meaning. The better they knew one another, the harder their relationship was to define wholly in medical terms. “Everything’s just too late sometimes, and then what do you do?” a farmer who had recently learned he had metastatic cancer told Tang one morning. “It’s just kind of sad to me and the wife, too. It’s just like a death sentence.”
“I know,” Tang said softly, “I know.” There was little else she could do medically. And yet, it was difficult to imagine the farmer saying the same to a friend or a relative or a social worker or a priest and receiving the same affirmation.
Meltzer characterized this singular connection for me one afternoon. “Being able to even just touch in for a little bit, have the conversation and talk about it, there’s an intimacy to it,” he said, shortly after telling one of his own primary-care patients in the hospital that she had cancer. The subconscious swap of “touch in” for “check in” evoked the laying on of hands — an antiquated, unbillable and, in the face of cancer, useless gesture, but quite possibly not a worthless one.