After a successful pilot in the Pacific Northwest, Kaiser Permanente is rolling out a community health worker (CHW) program across its entire enterprise. Leaders of this KP social health program and their partners at Impact Care recently spoke with Healthcare Innovation about how a randomized trial helped demonstrate the impact and build the business case for scaling up CHW efforts.

Nicole Friedman, national director of social health at KP, recalls that six or seven years ago KP had many clinical care management programs, all well intentioned, but they weren't seeing what they expected in terms of health outcomes and reduced utilization. “We had this problem that I think a lot of health systems struggle with, which is we built these programs around distinct domains or problems. We had a high-utilizer program, chronic heart failure case management, a falls programs, one for depression and anxiety. They’re all well-intentioned and probably getting good results in their specific domain,” she said, “but when you have patients with co-occurring chronic needs and they struggle with housing insecurity, food insecurity, and financial strain that persists across their clinical complexity, that's where the challenge is. So we stepped back and said, you know, we're doing the right things, but we're not really managing the whole-person care.”

That led KP execs to reach out to Shreya Kangovi, M.D., because they had heard about the work she had done at UPenn validating the impact of a community health worker program. She also started a public benefit corporation called Impact Care to hire and deploy CHWs in partnership with health systems and payers across the country. 

In 2020, during the height of the pandemic, KP Northwest worked with Impact Care to train and redeploy the patient navigators it already had on staff as community health workers. “We risk-stratified our population. We found a group to really focus on,” Friedman said. “We gave the CHWs standard caseload supervision and evidence-based workflows. We built all of that in Epic and worked with Impact on all the training, oversight, and population identification. That's how we saw some of the great outcomes that came from the Northwest, where we did this as a randomized controlled trial, so we could really isolate the impact of CHWs. We saw statistically significant reductions in hospital admissions and length of stay that not only paid for the program but saw a significant return that helped us then scale that across KP Northwest.”

The timeline of the study was a year, but when they got early indicators at nine months that this was working, KP leaders decided to begin spreading the program. “We went from a pilot of nine community health workers to now there are about 34 in that market,” Friedman said. 

“Importantly, we spread it not just for ambulatory care, but into other departments as well,” said Esme Cullen, M.D., M.P.H., chief health equity officer for national Medicaid and state programs at KP. “We recognized that this would actually be impactful for maternal health, for our emergency departments, for specialty clinics, even oncology, where we saw that there might be benefit for folks in particular periods of vulnerability. So I think another exciting part of the initial business case was its implications for broadening the population.”

Investing in the work force

"We know that the real drivers of health are social, behavioral, and economic factors that shape our daily lives,” Kangovi said. “So why do we keep investing in an American health workforce that doesn't really point at the problems we're trying to solve? That's the big picture, and I think that applies really nicely to Kaiser. But it also applies to any payer, any value-based provider who can capture the value of preventing poor health and costly utilization by going upstream.”

Kangovi said in the randomized trial KP saw cost savings and improvement in quality. They were able to validate with actuarial payer analysis what Impact Care had earlier proven in randomized control trials. KP saw significant reductions in hospital days per member. “They knew that most of what was ailing these members was things outside of biomedicine. They were things that were happening in individuals' day-to-day lives.”

The KP and Impact Care teams spoke about their work at a recent NCQA summit meeting, and Kangovi said that was a perfect forum to talk about this. NCQA has a white paper called "Effective Community Health Worker Programs: Critical Inputs for Supportive Systems,” which offers a framework for quality for these CHW programs, she said. It points to core components of effectiveness. 

“When we came into the partnership with KP after they had tried other approaches, we were excited to be able to roll up our sleeves,” Kangovi said. 

Doing a randomized control trial takes a lot of time, resources, and thoughtfulness,” Cullen said. “There are much faster analyses you can do, but they're not as rigorous. The benefit of conducting controlled trials is that you’re randomizing folks in such a way that you're able to prove causality, so we see huge value in these rigorous evaluations.”

The randomized trial to test the impact of the CHWs is something that plays to KP’s strengths as a learning health system, Cullen added. “We've done a very similar thing with, for instance, colorectal cancer screening and figuring out how to optimize FIT tests and how to reduce disparities there and isolating different interventions to understand that,” she said. “Also, with remote monitoring and blood pressure, we have worked to understand where and how best to scale that based on initial studies done in pilots. This is actually what we're set up best to do: isolate and understand where is the evidence really, and what do we want to scale, and where do we want to put our nickel down in a value-based care system?”

Friedman stressed that it was important that the CHWs were well-integrated with the clinical teams. “You can have the best intervention, but if you don't integrate the intervention into clinical operations and care delivery then you will fail,” she said. “And as important as that is the change management that comes along with the physicians, social workers, nurse care managers, pharmacists embracing this role.”

She said Impact Care calls this blueprinting and KP calls this co-design, but it's the same sentiment. They brought frontline managers, nurses, and social workers to the table to say this is what Nicole's going to do. This is the scope of her work as the community health worker. What are the referral workflows? When you are both managing a patient, what can Nicole do as your field partner in the community to enable you to do X, Y, and Z? How do you all communicate in Epic via secure chat? 

“We built standard clinical escalation protocols because community health workers can't give medical advice,” Friedman explained. “If you know there's a situation around interpersonal violence, they escalate to a licensed clinical social worker. We co-designed those workflows and have been able to adopt those similar types of workflows in all of our our markets.”

Cullen said health systems are still trying to figure out how to bring in the community-based providers in a meaningful and effective way. She added that it is important that the CHWs are included in team discussions of complex cases and are able to reach out to the clinical team when they need to escalate something.

“What has been beneficial is sharing the results of that initial evaluation broadly across KP, across our states, is that a lot of the leaders came to us excited,” Cullen explained, “because I think the gaps are known for populations that are struggling to either get the access that they need, the care they need, for all sorts of reasons that we really can't control in the clinical spaces.”

Spreading CHWs to other KP regions

Cullen said they saw huge opportunities when they looked at the populations in Mid-Atlantic, in Georgia, in Colorado, and in Southern California. Starting in 2024 they started to think about expanding into those areas. 

But because each of the populations is unique, Cullen said, there is a balance of fidelity and fit. There are core standard quality components they are going to keep the same from that original KP Northwest model that they know will lead to the quality outcomes they want to see, but there are also local adaptations they need, and a lot of those come from local input. “There's this critical period of talking with local leaders, trying to understand what this would look like here, and what's unique about these populations, whether because of the department we're in or just because of the market. Atlanta is not the same as Denver is not the same as Baltimore is not the same as West LA.”

One serendipitous event happened a few years ago when the president of KP Northwest became the president of KP Mid-Atlantic and brought the Impact CHW model to the Mid-Atlantic states in close consultation with KP Northwest. That was also the first market to test Impact’s full-service model, where Impact is the employer of the CHWs and integrates as a vendor into KP’s clinical care delivery model. They’ve seen good results with that and have now signed their third yearly contract. 

Following the Mid-Atlantic states, KP Georgia was next. It currently has 30 community health workers. “We were able to work with Impact and get that done rapidly,” Friedman recalled. “There started to be this natural groundswell. We realized not only is this working, but we can scale this quickly.” 

• Having good governance is critical to ensure support. Physician champions have been critical in all of KP’s deployments in embracing the role and helping lead the change management.

• Have a clear measurement and evaluation strategy so that you can demonstrate that the outcomes are attributable to the community health workers. 

• Avoid creating a community health worker program as just another program on top of programs. CHWs are enablers of other clinical care management population health strategies and should be part of a holistic population health model vs. stand-alone, Friedman said. When you integrate them into other clinical disease care management programs, it goes much smoother and the member outcomes improve.