Volume 3

Having an IMPaCT on Health Care

University of Pennsylvania program sends health care workers into disadvantaged neighborhoods, improves care, lowers cost


Ask people from disadvantaged communities about interacting with the health care system and they’ll voice several common complaints:

“I can’t afford the care.”

“It’s hard to get to medical appointments.”

“Medical staff, they’re so busy. They don’t listen and try to understand what I’m going through.”


A groundbreaking Philadelphia program is tackling those concerns by placing community health workers (CHWs) in hospitals and medical clinics. Their job: to learn how people in economically disadvantaged communities want to improve their health, and to help them make progress in achieving their goals.


The University of Pennsylvania Health System program, IMPaCT (Individualized Management for Patient-Centered Targets), has gained national recognition for its impressive results. Rigorous research demonstrates that its CHWs connect vulnerable patients to ongoing sources of medical care, are valued by community members, and yield a substantial positive return on investment.


Understandably, other health systems are interested in replicating IMPaCT’s achievements, and the program is turning its attention to disseminating its unique, carefully designed model more widely.


Socioeconomic factors shape health


Consider a statistic, not well known. The wonders of modern medicine contribute surprisingly little to the prevention of premature death in the United States — about 10 percent. By contrast, the impact of nonmedical factors — education and income, adequate housing and food, whether people exercise, what they eat, and whether they smoke, drink too much alcohol, or abuse drugs — is over five times greater.


Yet the health care system too often ignores the social, economic, and behavioral factors that shape people’s health.


Most affected by this narrow, medicine-is-all-that-matters approach are low-income individuals and families living in neighborhoods that suffer from a lack of resources and, often, trauma and violence. A sizeable body of research shows that these individuals die earlier, have a higher burden of illness, and receive less-than-satisfactory care more often than people higher up on the income ladder.


Enter IMPaCT, a program launched by Dr. Shreya Kangovi, a visionary physician at the University of Pennsylvania’s Perelman School of Medicine. Its focus: sending CHWs into disadvantaged Philadelphia neighborhoods to try to improve residents’ health. Its challenge: creating a model that delivers measurable, positive results that can be sustained over time in a variety of health care settings.


Origins and background


IMPaCT began modestly in 2010 as a research project, with $125,000 in funding.   In 2013, after research began to demonstrate its effectiveness, the University of Pennsylvania Health System helped to support the creation of the Penn Center for Community Health Workers, to spearhead research and dissemination of the model, as well as to offer IMPaCT to more patients.


“We have a lot of patients who need support with social needs as well as medical needs,” said Ralph Muller, chief executive officer of the health system, also known as Penn Medicine and IMPaCT’s first sustained funder. “It’s in our interest to help those patients get to the doctor, take their medications, eat well, stop smoking, maintain a healthier lifestyle, and not use hospital beds or the emergency room if their needs can be dealt with outside the hospital.”


Today, IMPaCT employs 30 community health workers who are integrated in medical teams at the Hospital of the University of Pennsylvania and Penn Presbyterian Medical Center, as well as a half a dozen Penn Medicine primary care clinics. “Our medical staff is very grateful to have their support,” Muller said. “It’s not the best use of medical staff’s time if patients aren’t taking their medications or coming to appointments, and workers’ familiarity with our patient population is very helpful in those areas.”


The passage of the Affordable Care Act (ACA), groundbreaking legislation with new incentives to reform health care practices and deliver value-based care, in 2010 was a catalyst for IMPaCT. Motivated by the new legislation, Kangovi pulled together a small group of researchers, health system leaders, and community

members to consider a significant question: Were there unmet opportunities to benefit vulnerable residents, as well as Penn Medicine, while achieving the so-called triple aim — better care that improves health at a lower cost? If so, what were they?


The group set four priorities. Residents and the health care system would be better off if more people had access to reliable, high-quality primary care and fewer people were admitted to hospitals for preventable reasons. Similarly, increasing people’s satisfaction with their health care and better controlling prevalent chronic illnesses such as diabetes and hypertension would help at-risk residents, as well as Penn Medicine, whose reimbursement was being shaped by new financial incentives coming from the ACA.


A geographic analysis turned up five ZIP codes in West and Southwest Philadelphia in which problems (higher-than-usual hospitalization rates, poor access to primary care, poor satisfaction with care, a high prevalence of chronic illness) were concentrated. Reaching residents who lived in these areas who were uninsured or on Medicaid, struggling with chronic medical conditions, and having a hard time navigating the health care system became a priority. IMPaCT subsequently expanded its service area to eight ZIP codes in Philadelphia.


Once this target population was identified, researchers set about building a program from the bottom up, rather than designing a program from the top down.


To gather input from community members, they hired a woman from West Philadelphia who was trained as an outreach worker and went into the neighborhoods. Three essential insights emerged from those conversations, according to Kangovi: “People wanted support from someone with whom they could relate and who had experience solving real-life problems; people were finding it incredibly difficult to get high-quality, affordable primary care,” which they wanted; and people felt they were being “set up to fail” because health care providers set overly ambitious goals and didn’t understand the reality of their lives.


Program design decisions flowed from the community’s input. IMPaCT would rely on community health workers, people who came from the same neighborhoods as the program’s target population, who could serve as support people, health coaches, problem-solvers, navigators, and advocates. Connecting people with high-quality primary care and other resources would be a key objective. And realistic goal-setting would be a fundamental strategy.


As Kangovi put it, “The program would be about helping patients decide what

they want to get done and then helping them do it.”


“It's not the best use of medical staff's time if patients aren't taking their medications or coming to appointments

A patient-centered approach


The notion that patients are in the driver’s seat, instead of doctors, nurses, social workers, or other professionals, is the exception rather than the rule in health care. Although there’s a great deal of talk about patient-centered care in health care circles, few programs achieve this in practice. IMPaCT does.


Three IMPaCT initiatives offer different levels of service to about 2,000 patients a year at Penn Medicine:

IMPaCT Transitions. This was the organization’s first program, created to help patients make a safe and effective transition home when they’re released from the hospital. CHWs help patients understand discharge instructions (which are often confusing), obtain medications, address their need for help at home, and schedule a follow-up appointment with a primary care doctor within 14 days, among other issues. This is a short-term program, lasting two to four weeks.


IMPaCT Outpatient (launched 2013). To be eligible for this six-month program, patients in primary care clinics must be uninsured or on Medicaid and have at least two chronic medical conditions, one of which is poorly controlled. Improving outcomes for at least one of these conditions is a primary goal. CHWs have weekly contact with patients, who are invited to join an ongoing peer support group.


IMPaCT Home (launched 2014). Some patients who’ve been discharged from the hospital have highly complex needs that can’t be addressed in a few weeks. They may have overlapping mental health and substance abuse problems, broken families, unstable living arrangements, and few reliable relationships. CHWs address those issues, providing a more intensive level of support over the course of three months in this program.


Essential program elements cross all three initiatives. Notably, every time CHWs interview patients, whether in a hospital or a medical clinic, they end up asking patients what they think they need to improve their health. “It’s an incredibly powerful question because it elicits what patients want and it gets buy-in,” Kangovi said.


The action plans CHWs create rely on patients’ input, taking into account their

personal circumstances. It’s common for multiple plans to be created and revised. Time and again, the plans’ goals testify to the disruptive impact of poverty,

unemployment, unaffordable housing, food scarcity, social isolation, and confusion about how to navigate the health system or connect with community resources.

In a report published in 2014, Kangovi and colleagues at Penn Medicine, the University of Pennsylvania, and the Philadelphia Veterans Affairs Center for Health Equity Research and Promotion described the issues patients identified and how CHWs attempted to address them:

A 42-year-old man with heart failure told a health worker he couldn’t afford public transportation to his primary care physician — even a fare of $2 was too much — and waited to become sick before calling an ambulance. That worker arranged van services through the patient’s insurer so he could see his doctor instead of going to the emergency room.


A 46-year-old patient with hypertension couldn’t afford a $65 co-payment for medications he’d been prescribed after being discharged from the hospital. His community health worker asked the doctor if it would be possible to prescribe generic medications without a co-pay.


A socially isolated 62-year-old woman who’d been hospitalized repeatedly for panic attacks and chest pain ended up going to a local recreation center, accompanied by her health worker. It became a place where she felt at home and able to interact with other people.


A depressed 53-year-old man with an ulcer had been self-medicating with marijuana but wanted to stop. His CHW arranged for him to enter an addiction services program. He went regularly and began the process of recovery.

What CHWs don’t do is educate patients about their medical conditions (that’s a job for health educators) or attempt to offer clinical advice (that’s well beyond their training and raises liability issues).


Recognizing the difficulties of the job — helping people whose lives are chaotic presents all kinds of challenges — IMPaCT leaves little to chance. Everything from performing initial patient interviews, which are in part scripted, to creating action plans to conducting home visits to tracking patients’ progress has been outlined, step by step, in a comprehensive program manual.


The objective: to create a standardized process, rather than a seat-of-the-pants approach, that works for patients in a wide variety of circumstances and that can be replicated across health care settings. And, to create a set of tools that maximizes opportunities for CHWs to be successful.


“This is a high-stakes job, and we need to be sure it’s done right,” Kangovi said.


Hiring, training, supervision


IMPaCT’s innovations extend to the hiring process. Instead of listing job openings on Penn Medicine’s “careers” site or on internet bulletin boards, the program reaches out to community development organizations, block captain associations, churches and interfaith organizations in search of “natural helpers,” people already helping other people in their neighborhoods. Instead of focusing on qualifications such as a college degree, the program seeks people with personal attributes such as reliability, problem-solving and communication skills, initiative, empathy, and compassion — qualities that community members said they valued.


“It’s all about getting the right people in the door and hiring for the right qualities,” said Casey Chanton, one of IMPaCT’s first employees, now the Assistant Director for Training and Development at the Penn Center for Community Health Workers.


Extensive classroom training, which focuses on handling real-life situations, qualitative interviewing, and role-playing, is provided after CHWs are hired. Over the course of a month, 140 hours of instruction cover everything from building trust with patients and helping them set goals to operating IMPaCT’s custom-built, computerized database and workflow system, known as HOMEBASE, to bringing work with patients to a close. Everything the CHWs will need to perform their work, including standardized forms that they can go over with patients, is described in the program manual.


Ensuring safety is one of the topics featured. CHWs always go out with another person when they conduct a home visit. If they encounter a situation that seems dangerous — for instance, someone who’s dealing drugs out of his home and who admits that he’s been targeted by gang members (a real story) — CHWs have to leave the premises immediately and report to their managers, who arrange a “safety huddle.”


“Any time a CHW feels unsafe working a patient, we all get on the phone or meet in person that day to talk about the issue and make a plan about how we’re going

to handle it,” said Jill Feldstein, chief operating officer at the Penn Center for Community Health Workers. For instance, a CHW might be told to meet a patient only in a public place, such as a coffee shop or a library, instead of going to his home (a strategy recommended with the patient described above).


Strong supervision is another program element. After classroom training, CHWs get weeks of on-the-job training, overseen by an experienced worker. When they’re working in the field, CHWs meet with managers at least once a week.


Every month, managers track outcomes for CHWs. Did they call all their patients regularly? Did patients make progress on their goals? What portion of patients saw primary care providers within a specified time frame? Managers interview at least one of the CHW’s patients each month to see how things are going.


“As a manager, I’m tracking performance really closely,” Chanton said. “If something goes off track, I can see it quickly and respond.”


At least once a year, IMPaCT hosts an all-staff “Design Jam” to brainstorm about program improvements. “We bring everyone together in a big room and go through our program manual page by page,” Kangovi said. “Everyone in the room has an opportunity to say ‘I don’t like that; it doesn’t work’ or make suggestions about what we could do better.”


In 2016, CHWs identified a group of patients that they were seeing but hadn’t previously singled out — people who were at the ends of their lives, who needed extra support, and who might have benefited from being referred to hospice or palliative care. Leaders subsequently arranged for a bereavement counselor to talk to staff, part of specialized training that occurs on a monthly basis.


Community-based research


In addition to building a rigorous infrastructure for the project, high-quality research has been a priority for IMPaCT from the outset.


IMPaCT’s research has focused on two key areas: understanding the needs and preferences of the communities it serves and rigorously evaluating the program’s outcomes.


One of leadership’s most innovative decisions was hiring Tamala Carter, a resident of West Philadelphia, to go out and talk to people about their lives and their experiences with the health care system at the start of the program. Carter,

IMPaCT’s community research coordinator, had previously served as an outreach worker with the Enterprise Center Community Development Corp. in West Philadelphia. Kangovi calls her a “natural ethnographer — someone who gets on the bus, stands next to someone and ends up listening to that person’s life story.”


Carter’s work took her to patients’ bedsides in hospitals and clinics, as well as libraries, stores, and homeless shelters. Ultimately, she interviewed 65 people over 22 months. “People were very open, which was surprising to me,” Carter said.


“Finally, they had someone who would sit there and listen to them and not judge them. They really wanted to share what was in their hearts and minds, and they really wanted the support.”


When Carter asked people what prevented them from taking care of their health, she got an earful. “Some of them had mental challenges; others had substance abuse challenges. There were financial issues — people were like, ‘I had to pay my rent so I didn’t have enough money to pay for medicine.’ Insurance was a barrier; some people didn’t have it. Not being employed, I heard a lot about that.”


Those needs pointed directly to the kind of assistance community members said would be most valuable: getting health insurance, referrals to substance abuse and mental health programs, aid with transportation, getting help to pay electric bills or rent, being directed to food banks, feeling that someone had their back,  and being connected to other resources that could provide ongoing assistance.


In addition to shaping IMPaCT’s design, conversations with community members helped leaders figure out that CHWs could forge trusting relationships with patients and conduct get-to-know-you patient interviews. At first workers were goal-directed in these conversations, but it soon became clear that could be off-putting. Now, CHWs start by asking people to tell their stories: where they grew up, what their family origin was like, what activities they enjoyed, and important events in their lives. Only then do CHWs begin to ask people about their current circumstances and issues they’d like to see addressed.


To illustrate why this is important, Carter told of an 80-year-old diabetic woman whose CHW was encouraging her to eliminate certain foods from her diet. “What the CHW didn’t realize was this patient had grown up dirt poor and her goal in life was to eat because she didn’t have food when she was young. Telling someone like that to give up food just wasn’t going to work.


“Based on situations like this, we realized that, in order to have CHWs and patients on the same page, CHWs needed to learn more about the patients,” Carter said. “So now we ask them to talk about their lives before we start talking about other issues.”


Academic studies


In initial interviews, community members made it clear that they believed their recovery from an illness or a medical crisis depended more on gaining access to high-quality primary care than avoiding another hospitalization. That became the primary outcome investigated in IMPaCT’s first scientific research report.


That study, published in JAMA Internal Medicine in April 2014, provided strong initial evidence of IMPaCT’s effectiveness. It looked at 446 patients who were admitted to two Penn Medicine hospitals from April 2011 to October 2012. The patients were randomly divided into two groups — those who got routine hospital care and those who got routine care plus assistance in the hospital and for up to two weeks following discharge from an IMPaCT community health worker.


Results showed that patients who received help from IMPaCT workers were 52 percent more likely to visit a primary care physician within two weeks of being discharged from the hospital. Also, patients reported improved mental health, a stronger sense of “activation” (feeling able and motivated to take care of themselves), and better communication with hospital staff at the time of discharge.


Positive findings didn’t extend across the board, however. Even with help from IMPaCT workers, patients didn’t report a meaningful effect on their physical health or their satisfaction with the medical care they received. Overall, patients who received IMPaCT services were just as likely to be readmitted to the hospital at least once, although multiple readmissions were lower in the group that received IMPaCT services.


The study’s conclusion: IMPaCT’s hospital transition intervention achieved a primary purpose — connecting vulnerable patients with a reliable source of follow-up medical care. At the same time, the study noted, “the intervention offers [the] health system a scalable strategy for implementing the principles of patient-centered care,” a much-discussed element of health care innovation.

Quantifying this impact was a significant accomplishment. “The million-dollar question in this field is ‘can you achieve hard outcomes by doing what patients want?’” Kangovi said. “Well, we showed you could.”


A second scientific report, published in August 2017 in the American Journal of Public Health, examined IMPaCT’s second intervention for low-income patients with multiple chronic illnesses who seek care in outpatient settings. Just over 300 patients participated in this study from July 2013 to October 2014. About half were randomly assigned to receive “usual” outpatient care, while the other half received usual care plus six months of weekly support from an IMPaCT community health worker and the opportunity to participate in a weekly support group.


Once again, the study’s findings highlighted IMPaCT’s benefits. Patients who set concrete health goals and an actionable plan to achieve them with the help of an IMPaCT CHW were less likely to be hospitalized than the control group and more likely to believe they’d gotten high-quality primary care that helped them manage their medical conditions. Again, patients’ mental health improved — an important finding because depression, anxiety, and other mental health conditions can sap motivation and exacerbate underlying medical conditions.


Financially, IMPaCT is a winner. Both evaluations documented a robust return on investment, estimated at $2 annually for every $1 put into the program.

Still, the difficulty of IMPaCT’s work shouldn’t be underestimated. Even with support, only 17.2 percent of patients whom IMPaCT worked with achieved the health goals they articulated at the start, according to the study. At best, they demonstrated “incremental improvements” in their medical conditions. And nearly two-thirds ended up not participating regularly in the peer support group that leaders had hoped would help them establish new social connections.

“Our work isn’t over. There’s plenty yet to do,” Kangovi said.


Wider dissemination


Impressed by favorable research results, health systems across the country have expressed interest in IMPaCT’s work. Partners HealthCare in Boston, the University of Pittsburgh Medical Center, Lehigh Valley Health Network, and Family First Health, a group of six federally qualified health centers in three Pennsylvania counties, have begun creating or improving their own CHW programs with technical assistance and support from IMPaCT staff. Also, a major Medicaid managed care plan, Keystone First, is offering IMPaCT CHW services to its members in Philadelphia.


Disseminating IMPaCT’s model more broadly has become a new program priority, spearheaded by a new initiative, IMPaCT Communities. To aid in this effort, staff

created an online learning system, which includes five hours of interactive training content. Extensive resources are available on the organization’s website, chw.upenn.edu.


As for the future, “we’re pretty excited about the potential of community health workers to continue to improve health in high-risk populations,” Feldstein said.  “So many programs of this kind have been struggling, and we think we’ve learned a lot about the elements of success. We’re looking forward to sharing those more widely.”