A petite woman, Dr. Shreya Kangovi, has an outsize vision.
As the founding executive director of the Penn Center for Community Health Workers, a national center of excellence, Kangovi wants to translate nationwide the success Penn Medicine has seen in implementing the community health worker program she helped develop.
“If you look at the map of the United States there are hot spots where where we can clearly see life expectancy being determined by where you live and your income. So If you drill into those hot spots, you see real life issues – things like housing, transportation, trauma, child care.” Kangovi, who is trained physician, said in a recent interview in her office in downtown Philadelphia. “My center’s catchphrase to get us out of bed in the morning is ‘we want to make those red hot spots blue.’ ”
Philadelphia is no stranger to this “hot spots” phenomenon with life expectancies varying disturbingly between zip codes barely a few miles apart. And as Kangovi began researching community health worker programs in 2010, what became clear is that people represent the cornerstone of any successful, evidence-based, scientifically-developed CHW program that aims to address the social determinants of health. There aren’t any tech-based silver bullets.
The limitations of a tech-based approach
At a time that startups in Silicon Valley and around the country lionize digital health as the messiah that will rein in chronic diseases and save U.S. healthcare, a conversation with Kangovi illuminates the limitations of technology.
“A lot of these apps are based on a fundamental, theoretical premise — knowing your numbers is a good thing. I am going to tell you your sugar, your weight, or your blood pressure and you will take that information and say, ‘I ate a lot of cookies last week and this is motivating me to to get back into the gym.’”
Such people have a learning response. But others can have a quite different reaction.
Dr. Shreya Kangovi, founding executive director, Penn Center for Community Health Workers, is passionate about healthcare equity
“People who face failure and adversity a lot may have to develop avoidance as a coping mechanism and you know those of us who can throw money at the problem, we’re learners because we can solve those problems,” Kangovi said. “If you check your glucometer and it says 300 everyday and it’s because you are short on your rent and it’s so stressful that it will make you eat, that glucometer is not going to pay your rent for you.”
She added people have different “behavioral phenotypes” and on bad days even wealthy people with means may choose to avoid rather than confront the status of their health.
Another healthcare executive, who is well aware of Kangovi’s work in the field and applauds her achievements, echoed her.
“I would urge us to not use simple shorthands like people with socioeconomically challenged situations need X and people who are wealthy need Y,” said Dr. Sachin Jain, president and CEO of CareMore Health, a division of Anthem, in a phone interview last week. “I think the most important observation is that people are people. There’s lots of different kinds of people within all income categories and so there are wealthy people that can benefit from community health workers and poor people who can benefit from community health workers.”
Jain also broadly agreed with Kangovi that the potential of digital health as conceived by techies smacks of preconceived notions of about how actively engaged people want to be with their health.
“The digital health movement has largely grown out of assumptions about patient behavior by nonsick people,” Jain declared. “A lot of the people who are developing the solutions are building solutions for what they would need, what they would want, but not recognizing that they might want different things when they are actually sick. Many patients don’t want to be active patients. ”
There are other, structural reasons why digital health may fail with certain patient communities. To describe those, Kangovi launched into a lengthy explanation:
Why don’t lower income people who really have struggles with diabetes use these apps? No. 1, there’s still a digital divide. We act like everybody has the same tools. They don’t. We’re doing a study where we are focusing on individuals that have diabetes who come from high poverty zip codes and the diabetes is in poor control. Only 30 percent of the people have a smartphone in that population. We initially were thinking that we would combine community health workers support with some smartphone-based glucometer tracking but we couldn’t. People don’t have smartphones. And even if they have the device they may not have the data.
So what can community health workers do to help a stressed patient? They can develop a meaningful relationship with them, earn their trust and address these challenges. For the person worried about rent, the community worker can connect him or her to the right person who can advocate for the patient with the patients’ landlord for instance, Kangovi said.
Avoiding the 5 pitfalls of community health workers program
The success of any CHW program is largely dependent on how well you address the reasons why such programs fail. Before developing the IMPaCT program at Penn — whose model has been described in JAMA showing hospital admissions rates falling 65 percent and a ROI of $2 for every dollar spent on the program — Dr. Kangovi thoroughly researched the subject. Her background in global health and research made her dive deep into the medical literature through which she identified five causes that have hamstrung CHW programs throughout history:
- high turnover of workers, sometimes as high as 50-77 percent;
- lack of a proper infrastructure to support the community health workers;
- too diseases-specific or clinical focused;
- no balance between medical and community-based work
- flawed metrics to judge program success
One of the first things Kangovi and her team did was to identify the needs of the community that they were trying to lift up. They conducted 1,500 patient interviews and then based on those needs, they tried to hire the community workers who would become full-time employees of the program. Choosing the right candidate is key.
“We don’t post a job on Indeed.com. We circulate it through soup kitchens or churches where there are volunteers. And then we have a very easy paper applications but we do it mostly on meet and greet where we can observe interpersonal skills,” Kangovi said.
She added that personality tests, behavior assessment can also be conducted and the IMPaCT program has also used organizational psychology “which many other industries have but healthcare has lagged” to hire community workers. Kangovi also made sure that proper training and support is available for the community workers. As a result, the IMPaCT program at Penn has had a turnover of only 1.7 percent compared with at least 50 percent experienced by other programs, according to a NEJM paper that Kangovi authored.
But an equally important aspect of the program is staying away from a doc/nurse focus who, even with the best intentions, are not well suited to relate to patients’ life experiences.
“Doctors and nurses of a different background say, “Why is this person paying for TV when they don’t have $3 for a medication?” Whereas a person who has been there can say, “Well that’s the thing that is keeping them from blowing their brains out right now, so I get it.’ So that is why this concept of the community health worker and the concept of shared experience is important.”
Jain pointed at another aspect for why community health workers is so vital. For a variety of reasons, certain populations lack trust of the healthcare system at large. He observed this as CareMore Health began serving Medicaid patients in Tennessee four years ago.
“The trick in Medicaid is getting people to want to trust you, to want to use the resources that are available to them,” he said. “People who are from the community, know the community and are of the community can often times play a valuable role in bridging the gap of trust that exists between various patient populations.”
Developing a scientific method of evaluation
But developing a good CHW program doesn’t simply begin and end with hiring people from the community and expecting that the connection they share will magically yield great returns in terms of good medical outcomes. That’s where a good evaluation system play a role. Kangovi said that what plagued CHW programs is that health systems would take a group of very sick patients, intervene with some community health worker program and see hospital admission rates fall among this group. And when they fell, program leaders patted themselves on the back for a job well done. Ultimately though, the hospital rates would rise again leading organizations to abandon CHW programs.
Yet the real problem was wrong metrics and not recognizing that only comparing sick people against themselves can lead to the mathematical problem of regression to the mean.
“They’ve had a bad year, they’ve been in the hospital 365 times and so there’s no way to go but down,” she explained of hospital rates falling back after being high for a period. “Life goes in ups and downs and if you choose an outlier point and you do nothing at all they will regress to the mean.”
What the IMPaCT program has done by contrast is run randomized control trials where patients are divided into treatment and control arms solely on the basis of whether they are receiving the CHW intervention, all other things being equal. Three RCT studies done on IMPaCT program show that hospital rates fell 65 percent among those who received that CHW compared to those who didn’t.
Kangovi and her team have also developed a cloud-based application that allows community health workers to input patient information through their phones or laptops after interacting with them. Supervisors can monitor individual performance of those community health workers and make adjustments where necessary.
To date, the IMPaCT program has served more than 10,ooo people in the Philadelphia region. The Penn Center for Community Health Workers, which officially launched in 2014, has 57 full time equivalents, 30 of whom are community healthcare workers. The Center has received grants from the National Institutes of Health as well as the Patient Centered Outcome Research Organization(PCORI). Penn Medicine has invested $3 million and the CHW program is part of its annual budget though Kangovi did not say how much it invests per year. The Center has an annual budget of $4.5 million.
Included within its budget, is a revenue stream through partners. These are essentially other integrated health systems, payers, state Medicaid organizations, or other institutions, that are using the IMPaCT model to run a scientifically-validated community health workers program in their own regions. Kangovi’s team at the Center can help external organization with planning, hiring, training, launching, troubleshooting and evaluating their programs. Kangovi said the model is being used by 30 organizations across 15 different states including the Veterans Health Administration.
Disseminating the Penn model
One such organization that wanted to upgrade its community health worker program is Siloam Health, a Nashville-based Christian nonprofit. In a recent phone interview, Amy Richardson, Siloam Health’s chief community health officer said that the organization looks after nearly 4,000 patients, 90 percent of whom are foreign born, through a primary care clinic. Most of these patients do not have health insurance.
Siloam launched a CHW program about four years ago and the goal was to improve the health of patients in four different communities — the Latino, Arab, Burmese and Nepali populations. But over time it became clear that Richardson was managing four different CHW programs rather than a single, cohesive one. The program had part-time community health workers and about a year ago the decision was made to bring them in-house and make them full-time. But at around the same time, Richardson was introduced to Kangovi through a Siloam board member.
Siloam has now replicated the IMPaCT model and Richardson couldn’t be happier. The program is more streamlined and structured and her community health workers feel less exhausted. For instance, before the IMPaCT model was instituted, the community health worker in charge of managing patients in the Latino community had to adjust schedules constantly to make sure that she could serve patients some of whom worked days and some nights. Now, she has set hours and her patients have also adjusted to know when she is available. The format can be adjusted when needed.
“It’s still flexible enough to vulnerable populations’ schedules but it’s not letting the community run the health worker to the ground,” Richardson said. “So it’s something sort of simple, but was revolutionary to the way we are taking care of our employees.”
Siloam also serves the Nepali-speaking population struggle with loneliness and lack of relevance, especially the older generation who have come to the U.S. mainly as refugees. They are also unaware of how to take their medications and refill prescriptions, Richardson said. The community health worker assigned to this group helps by not only helping to educate them and their family members to navigate the health system but also make them reconnect with each other.
While no formal study has been launched to evaluate the one-year program, Richardson reports no turnover among the four full-time CHWs and improving obesity rates and A1c levels among the program participants.
“It’s not surprising that these medical outcomes would improve as we focus on the social factors affecting health, but it’s still revolutionary in the healthcare world,” Richardson said.
She added that one of the most appealing things about the IMPaCT model — aside from the rigorous framework and the assistance from Kangovi’s team — is the cloud-based data collection system.
“Before, I was tweaking our data collection system every year trying to figure out the best way to track our community health workers activities and then patient outcomes,” she said.
Ultimately, though, Richardson, like Kangovi and Jain, pointed to the power of the human connection as the vital element in any CHW program.
“We are influenced by those around us whether we are a more educated health care user or someone who has very low health literacy,” she said. “We change because of relationships.”
Photo: kate_sept2004, Getty Images and Arundhati Parmar