Jessica Pino writes in a notebook in her hospital room at HUP

The U.S. health care system treats body and mind as separate. But in the hospital and in the clinic, a new integrated care model treats a person as a whole.

By S.I. Rosenbaum

Photos by Peggy Peterson

The patient was very sick, and she was alone.

Her name was Jessica Pino. She had been vomiting constantly for days; everything she ate made her ill. On one of her previous hospitalizations—there had been more than she could count; had it been 40 times now? 50?—she’d picked up a contagious bacterium, so now, on her first admission to the Hospital of the University of Pennsylvania (HUP), she was in an isolation room.

Usually her mother would be by her side, even in an isolation room. But her mother was sick too, this time. She was on her own.

It’s hard to be in the hospital alone—even for someone like Pino, who had gotten good at keeping it together. For six years Crohn’s disease had wracked her gut, disrupting every aspect of her life.

When she was first diagnosed, doctors told her not to stress: “Stress is the worst thing for you,” she remembers them saying. Pino almost wanted to laugh the first time she heard that line. How could she not be stressed when she was vomiting for hours? When doctors were slicing open an inflamed cyst on her face? When she was too nauseous to get out of bed?

And there were other sources of tension too—not just tension, but trauma. There was her father’s sudden death from an overdose when she was just 18. There was the abusive relationship she fell into afterwards, and the years she spent extracting herself from her ex’s control. The years she spent estranged from her mother: talking about it still makes her cry. The doctors who first treated her gastrointestinal symptoms didn’t ask her about any of this.

No patient arrives at the hospital a blank slate. In addition to whatever physical crisis they’re experiencing and the upheaval of being in the hospital, they all come with troubles and pain of their own, issues their doctors may know nothing about and be underequipped to help them navigate.

“There’s a backstory,” Pino says. “Everyone has a backstory.”

Now she was in the room by herself, feeling sick, trying not to stress, writing down questions for her doctors.

There was a knock on the door. For Pino, and for a growing number of patients who may experience a broad array of mental health challenges in the hospital, things were about to change.

The Psychiatrist’s Challenge

Years earlier, when Cecilia Livesey, MD, GME’16 was a resident at HUP, she worked on the inpatient psychiatric consult rotation. She saw the mental health challenges patients like Pino faced, on the same inpatient medicine floor where Pino would later stay, from the other side. She met many of those patients in the midst of a far more acute psychiatric episode than Pino’s creeping anxiety, though. Often she would get a call from a provider in the middle of the night, when a patient was in the midst of a psychiatric crisis.

Since MEND was implemented on two internal medicine units, 44 percent of patients have received a proactive consultation for support with their mental health needs—a nearly eightfold increase.

“By the time the provider called me, both patient and provider were fairly worked up,” she recalls. “I’d think, ‘Why has this become a crisis? If only I’d known about this two days ago... if only we could have intervened before the situation escalated.’”

Because she wasn’t on the ward until she was summoned, she often felt that she was being thrown into a scene already in progress. There were the dramatic cases, with hospital security on the scene, trying to calm a patient who was attempting to leave against medical advice. And there were also subtler ones: “Sometimes a medical team will call in a psych consult and say they have been trying to get a patient to go to dialysis for five hours. It turns out this patient is just anxious; if we’d known the story and known about the frustration and anxiety ahead of time, we could address that proactively and avoid these delays in care.”

Livesey found the lack of information ineffective: she often found herself having to react to the patient’s distress without useful context for how it had developed.

There was a massive unmet need. Only three percent of patients across HUP were getting psychiatric consults, consistent with the national average—but research has shown that 25 to 40 percent of patients in the hospital have an acute or chronic psychiatric issue.

Yet, in medicine, physical and mental health have long been treated as distinct disciplines, divided from each other geographically, educationally, and financially.

Primary care doctors who wish to refer patients to mental health care must often contend with entirely distinct insurance and computer information systems, if the mental health practitioner takes insurance at all. In hospitals, mental health and physical health are separate departments; if a patient’s upcoming gallbladder surgery prompts a psychological crisis, there’s no mental health provider stationed on the ward to help them.

But for actual people, mental and physical health is rarely separable the way hospital architecture and insurance billing systems would have it.

Livesey wished the system were different. She wished she could help more patients, and help them sooner.

Cecilia Livesey, MD, is chief of Integrated Services in Psychiatry at the Perelman School of Medicine.

The Conversation

Leona Pierce was outside the door of the isolation room at HUP.

There was a lot Pierce, a licensed clinical social worker, already knew about the woman inside, a 36-year-old named Jessica Pino. She knew, for example, that she had been hospitalized many times before for her Crohn’s disease at other hospitals in the region. She also knew that Pino had previously taken medication for anxiety and depression, and that she had a psychiatric history of complex trauma.

She recalls that she was “prepared to talk to someone that didn’t really have a tool kit” to deal with stress. But Pino surprised her.

“I’ve been in therapy almost my whole life,” Pino told her.

They sat together, talking.

Pino told Pierce about her first hospitalization, years before, when her then-boyfriend had left her at the hospital door and driven away. She talked about how, before she was finally diagnosed with Crohn’s, some doctors had told her she was just having “stomach pain” and dismissed her. How her mother had been with her every step of the way, and how hard it was to be in the hospital without her. How she missed her father every day and sometimes felt she could hear his voice.

“She shared her history, soup to nuts,” Pierce says. “And from that point on every time she came back, which I think was four or five times, I was able to dive right in: ‘How's your anxiety, how’s your Crohn’s, how are you feeling?’”

In the past, Pino’s statements about her symptoms had been second-guessed or overlooked by doctors. Now, Pierce was able to help her communicate clearly with her physicians by vouching for her sound judgement about her own medical condition.

Only three percent of patients in the hospital receive a psychiatric consultation, on average—but research has shown that 25 to 40 percent have an acute or chronic psychiatric issue.

“I was able to stand next to her and say, ‘Yes, I know her, we’ve gone through it,’” she says. “I was a witness to her admissions, I was a witness to her mental health care and to her medical health care and medical literacy. Sometimes that's all you want—you want someone there watching you and validating you.”

More than just advocating for Pino, Pierce was another key clinician on her team. In their therapy sessions, she was able to help Pino manage her anxiety and stress while in the hospital. As her mental health improved, so too did the symptoms of her Crohn’s disease.

The Program

The pilot program that brought Pierce and Pino together is called the Mental Health Engagement, Navigation & Delivery (MEND) program. It was spearheaded, along with its cousin, the primary-care-based Penn Integrated Care (PIC) program, by Cecilia Livesey, now the chief of Integrated Services for Psychiatry at the Perelman School of Medicine. MEND aims to join mental and physical health and their respective practitioners in a model known as integrated care, making it easier for patients to access mental health care where and when they access physical medicine. These programs are among the latest examples in a movement in mental health care that has been slowly unfolding. “The integration of behavioral health and general medical services has been the focus of intensive resources, planning, and education efforts for at least a decade,” the American Psychiatric Association wrote in a 2016 report.

The need for a program like MEND was obvious, given that research shows the number of patients who experience psychiatric issues in the hospital could be as much as ten times higher than the number receiving psychiatric consultations. Some of these patients, like Pino, have past traumas and anxiety. Others have active substance use disorders, schizophrenia, major depressive disorder, eating disorders—the whole gamut of mental health difficulties people can experience outside of the hospital will naturally stay with them once they’re admitted.

When she was offered the opportunity to proactively intervene on mental health in an inpatient medical setting, Pierce says, “I thought, ‘This is perfect ... there is going to be a way for collaboration between disciplines and the attention that has never fully been given to the most important part of our bodies, our minds.’”

Beginning in January 2019 Pierce and a team of psychiatric health providers were assigned to be actively embedded in two general medicine units at HUP, comprising 78 beds. Patients on these units have a variety of non-surgical medical needs, from Pino’s Crohn’s disease and other gastrointestinal disorders, to complex cardiac and pulmonary conditions. As new patients enter these units, their medical case histories are scanned by an innovative computer algorithm that rates them by how likely they are to benefit from mental health services.

Eleanor Anderson, MD, GME’13, MEND’s lead psychiatrist, then meets with the other members of the team—a social worker, a psychiatric nurse practitioner, and occasionally a rotating psychiatry resident—each day to go over the patients whose files have been flagged. They agree on which patients are good candidates. Then Anderson and her team go to the unit to talk with the rounding doctors from the medical services.

As a psychiatrist for MEND, Eleanor Anderson, MD, meets daily with a nurse practitioner (Colleen McKelvey, CRNP, at right) and a licensed clinical social worker to review cases flagged by the MEND algorithm based on keywords in their health records. They round together on the inpatient floor to discuss these patients with their medical providers.

“There's no substitute for that face to face interaction,” Anderson says. “One of the key points of this system is the collaborative feel of it; [doctors] need to feel we're there and available, so they don’t feel alone, they feel like we're true collaborators … We’re not just called in and then we leave, which is the standard consult model.”

Besides, she says, “Sometimes the primary medical teams might not even know how we could be helpful yet. We had to give them an idea of, ‘Here is what you can use us for.’”

When she runs into a doctor whose patient MEND has flagged as a good candidate, Anderson checks in to see if the doctor agrees. Sometimes they’ll say that the patient seems fine, is about to be discharged, or for some other reason doesn’t need mental health support. But usually they’ll agree to dispatch a MEND team member to see the patient. Which team member depends on the individual patient’s needs: A social worker might focus on a brief talk-therapy consultation, while a nurse practitioner may weigh in on a diagnostic workup question or managing psychiatric medications. Anderson, the medical doctor, works with the patients who have the most complex needs. It has turned out only one in five MEND patients need that level of care.

Before MEND was implemented, about six percent of patients on the two units received psychiatric consultations. Since MEND was implemented, 44 percent of these units’ patients have received a proactive consultation for support with their mental health needs—a nearly eight-fold increase.

The Dramatic Change

Emmanuel King MD, SFHM, FACP, a professor of Clinical Medicine at the Perelman School of Medicine, co-leads the medical units piloting MEND. He says the MEND program has transformed the ward.

Before MEND, “the patients weren’t really having their needs met,” King says. Often he and other doctors would call to request a psychiatric consult only after a patient was already in distress and acting out—on average, the fifth day they spent in the hospital.

He remembers being grabbed by his lapels by one patient, dragged close and sworn at, called derogatory names. But that patient might not have been so agitated when they were first admitted—his delirium or disorientation and other symptoms might have escalated over a period of days before he became violent.

Other times a patient might shut down, refusing to accept medication or skipping physical therapy sessions. It might take some time before the clinical team noticed these quieter patients who weren’t improving as quickly as they should.

Even in such situations, however, King says they had to weigh whether a patient’s behavioral health problem was serious enough to require the consult service to intervene. It was also likely to be a different person picking up the phone every time, as the position rotates among psychiatry residents, fellows, and attendings, so there would be little continuity or familiarity with patients.

“When you have a system where there are so many steps to get a patient seen, it becomes much more challenging and things break down,” King says. It was hard on patients and doctors alike, he says, since patients in distress took up the bulk of doctors’ time and tended to lead to burnout and fatigue as doctors struggled with crises they weren’t trained to handle. Even when the traditional psychiatric consult team was called in, their capacity to help was limited—in order for a handful of psychiatrists to serve the entire hospital, they needed to limit their time to mainly focus on concrete questions such as medication advice or assessing a patient’s decision-making capacity.

“It's very different to have a team on the floor looking to help, proactively screening but also never turning away a request,” King says. Now, according to statistics gathered by the MEND team, mental health providers on average meet with patients in need on their first day of stay in the hospital, not the fifth.

MEND has expanded the scope of mental health care that is available to patients on a medical unit. With the “old model,” King says he probably wouldn’t have consulted psychiatry for a patient like Pino: her situation couldn’t be condensed into a request for medication or help with a singular crisis that required a psychiatrist’s intervention. But MEND draws from a broader multidisciplinary bench, and they are able to treat not only diagnosable psychiatric illnesses but the psychological traumas and stresses of hospital stays.

Because Pierce was able to establish a rapport with Pino, her anxiety level dropped—and her next hospitalization was less stressful, King says. And her physical illness responded to her anxiety and mood. “The social worker would see her, catch up with her, and make sure she was doing ok,” he says. The team also floated the possibility of connecting with an outpatient psychiatrist that had a special interest in patients with gastrointestinal disorders. “That brought her a lot of optimism,” King says. “It was very different than her stay otherwise would have been before the service was there.”


Connecting patients with ongoing treatment is a big part of MEND, and it’s often crucial that that process begin while the patient is in the hospital, King says. For example, the MEND team was able to help one patient wean off methadone while she was in the hospital for a cardiac condition, and get on a waiting list for a drug recovery program she could participate in with her young child. She could leave the hospital with a support structure already in place.

King singled out the case in a letter he sent to HUP and Penn Medicine senior leaders, asking that MEND be made permanent and expanded throughout the hospital.

King and his colleagues could vouch for its measurable benefits beyond individual anecdotes: with MEND assisting in the care of patients with mental health needs, early data showed that fewer patients were being discharged to inpatient psychiatry and that fewer patients were readmitted within three months. Most remarkably, the unit experienced a decrease in length of stay for all its patients—not just the ones MEND clinicians interacted with.      

It’s because doctors and nurses are spending less time tied up with patients exhibiting behavioral or mental health problems—problems beyond the initial medical need that brought them to the hospital, something that can take extra time, and sometimes make clinicians feel helpless or frustrated. With MEND team members taking the lead, in particular with behavioral crises, the doctors can focus more on the physical ailments within their core specialties. Medical residents in internal medicine have reported that they feel better educated on psychiatric care after rotating in the units with the MEND team. Overall, the doctors on these floors feel less fatigued and burnt out, they’ve told King. Many signed on to endorse the program and its impact on both patients’ care and their own job satisfaction in letters to HUP and Penn Medicine leaders.

The Model

For Livesey, MEND is a balm and a step toward the transformation to the system that troubled her since her residency rotations on the psychiatry consult service. That model of treating mental health crises in a reactive way in the hospital didn’t make any sense, she says, given how mental health can have ramifications for a person’s physical health, and vice versa. For example, someone who contracts a physical illness can lose their ability to work or function at home the way they’re used to. Being in pain can affect their temperament, and the stress of managing an illness can impact their relationships. For someone who is already prone to mental illness, this can be a disaster.

By the same token, a mental health condition like depression or bipolar disorder can take a toll on the body, disrupting sleep, exercise, diet, or treatment adherence for other health problems, leading to physical illness. And, the brain and the body share biological markers of distress, such as inflammation and hormone dysregulation. Congestive heart failure and depression involve a similar inflammatory cascade, for instance. “It’s a chicken and egg question,” Livesey says, “but they seem to be connected.”

All of this was on Livesey’s mind when she became medical director of strategy and integration for Psychiatry at Penn in 2017 with the support of Maria Oquendo, MD, PhD, the new chair of Psychiatry, and David Oslin, MD, the chief of Psychiatry at the Michael J. Crescenz VA Medical Center and a professor of Psychiatry at Penn. Her first project was what would become Penn Integrated Care, or PIC.

Similarly to how MEND embeds health care professionals in a hospital ward, PIC placed licensed clinical social workers in eight primary care locations starting in January 2018.

With MEND team members taking the lead with behavioral health crises, the doctors can focus more on the physical ailments within their core specialties—and they report feeling less fatigued and burnt out as a result.

When patients come to see their primary care provider, they can access mental health care at the same location—rather than have to contend with a separate practice in another part of town. If a patient needs it, their provider can even just walk them down the hall to the social worker’s office for a “warm hand off” so they can get immediate care.

She was lucky, she says, to partner with Matthew Press, MD, MSc’10. Press had just started as associate medical director of the Penn Medicine Primary Care service line, fresh from a stint working for Medicare—where he helped develop a set of new billing codes that providers could use to bill for mental health care in a primary care setting.

“The collaborative care model had been studied and shown to be effective, but very few health care systems used it because there was no way to pay for it,” Press explains. “Once Medicare created codes in January of 2017, it sets the example for the rest of the industry. Medicare creates the billing code, and then it’s up to individual insurance companies to decide to cover it or not.”

In the first year of PIC, Press and Livesey lobbied hard with insurance providers across the region to get them to accept the new codes. “When we went and made our pitch to them, we said ‘This isn’t just us asking you to pay for the newest tech or expensive drug,” Press says. “‘Those things are all important, but look, we’ve got in some ways a wonder drug, a program that is not only going to open access to your members to mental health and deliver better mental health outcomes to them, but it saves money overall.”

Matthew Press, MD, MSc, associate medical director of the Primary Care service line and an associate professor of General Internal Medicine in the Perelman School of Medicine, partnered with Livesey to bring Penn Integrated Care to primary care offices.

When the program began, the team expected 500 patients to be referred in the first year. Instead they got 6,000.

The familiarity and ease of access is what made the program so popular, Livesey says. “When people are there with their primary care provider who they already know and trust, at a location that they know and trust, and you say ‘I can offer you mental health services right here,’ they say, ‘Oh that's great!’”

“If you look historically, our national health care system has essentially created two separate systems, a mental health system and a medical system,” says Press. “How we pay and the amount we pay is different between mental and physical health, and coordination and communication between them is completely siloed.”

“But you can’t carve a person up into different problems. They’re one person.”

The Path Forward

As PIC was getting off the ground, Livesey and a multidisciplinary team, including Colleen McKelvey, CRNP and Christine Chevellier, LCSW, were applying to Penn’s Center for Health Care Innovation Accelerator Program—the organization’s program to rapidly design, test, and implement transformative care models. They’d been looking into integrated care in hospital settings, visiting other institutions who had programs similar to MEND, and Livesey had an idea.

“One thing I’d noticed was that the identification and engagement piece was laborious,” she says. Mental health clinicians triaging incoming patients would manually review their charts. “Wouldn’t it be great if we could identify these patients right when they hit the floor?”

The winning application proposed creating a smart computer algorithm that could score incoming patients on the likelihood that they’d benefit from mental health intervention, shortcutting the triage process and allowing providers to move fast. It also suggested a multidisciplinary team approach and a focus on navigating patients to care outside the hospital.

Of course, that was no easy task.

Livesey and her team dream big—of a world in which collaborative care is embedded in all medical practices, from cardiology to cancer care.

Once accepted into the accelerator program, Livesey partnered with Kelley Kugler, MSc, an innovation manager with the center. As a former consultant, Kugler had experience developing data-led services for use in government, retail, and electronics. As a pair, they discussed how these tools might meet certain challenges of mental health. The idea, Kugler explains, was to create something that worked similarly to a spam filter for email—except instead of picking up on keywords that indicate spam, they’d be teaching a computer to pick up on keywords that indicated a history of mental illness.

“We started out looking at things like what meds are someone on, or do they have a previous psychiatric diagnosis,” Livesey says. “That got us nowhere.” Because mental health care can be stigmatized and hard to access, many patients who needed it didn’t have those elements in their health records.

So instead, with the support of coder-clinician David Do, MD, an assistant professor of Clinical Neurology, they were able to let the computer teach itself to recognize patients the way health care professionals do.

Four clinicians trained in psychiatry spent hours reviewing about 300 patient charts and graded them on a scale of one to five on how likely they might be to benefit from mental health care, as well as the severity of care they’d require.

“I’d pretend I was seeing them for the first time: ‘OK, they're arrived on the floor, how likely are they to need a psychiatrist,’” recalls Anderson, who was one of the clinicians. She spent her own time, nights and weekends, on the project.

“It’s kind of a labor of love,” she says. “We really believe in the model. We wanted to create this tool for ourselves, and it has made things easier, so it was an investment.”

The algorithm looks for charts that share certain code words with the charts Anderson, Livesey and others evaluated—like “noncompliance,” “consequences,” and “victims.”

“There were over 2,000 words that were statistically significant,” Livesey explains. 

It’s about 70 percent accurate, she says, at the end of the accelerator program’s first phase, and the Penn Data Science team, led by Michael Draugelis, is at work on making it even more accurate. 

Both PIC and MEND have proved popular with the medical teams. “The broader acceptance has been tremendous,” she says. “It’s a relief for clinicians to be supported as they care for these really complex patients.”

Livesey points out that MEND can yield many benefits to the health system—addressing patients’ clinical needs, alleviating provider burnout, and even reducing costs by allowing patients to go home sooner. Plus, she says, there may be an opportunity for insurers to support MEND’s model for inpatient care, as they do in PIC for outpatients, through a case rate or the collaborative care codes.

Within the first two years of the PIC program, in eight practices, more than 14,000 patients were referred for mental health care. One in ten of these patients referred to care were identified as having had suicidal thoughts. PIC is in the process of expanding to six more practices with plans to support the whole primary care service line within the next 2-3 years. PIC is currently running a pilot in Penn Medicine’s 24/7 OnDemand telemedicine urgent care service to serve Penn Medicine employees with mental health needs. 

The model is growing, and Livesey and the PIC team dream big—of a world in which collaborative care is embedded in all medical practices, from cardiology to cancer care. While her main focus is the primary care expansion, Livesey’s next goal is to integrate mental health with outpatient medical specialists. This spring, PIC will be expanding to a Penn Medicine obstetrics and gynecology office. And in January 2020, a MEND team launched at Pennsylvania Hospital, which is also home to two inpatient psychiatry units, a crisis center, and a community mental health center. Offering MEND in a new context will test the model and help demonstrate what MEND would look like on a broader scale, Livesey notes.

For patients like Pino, the program’s support can make all the difference.

“Coming to such a big hospital, at first you’re kind of worried, wondering, ‘Am I going to get lost in the hustle and bustle?’”

“Having someone sit down and take the time, to where they’re not saying ‘Oh I gotta go,’ it makes you feel like, ‘Wow, it doesn’t matter how big the team is, the doctors care, they want you to fully get better,’” she says. “Not only physically, but mentally get better.”

Share This Page: