Last week, the Camden Coalition hosted the national webinar, “The ROI of resource navigation: Findings from the Accountable Health Communities Model,” to provide expert perspective on the Center for Medicare and Medicaid Innovation (CMMI)’s third evaluation report for the Accountable Health Communities (AHC) Model in November 2024.

The AHC Model assessed the strategy of health-related social needs (HRSN) screening and navigation among a national participant population of one million-plus Medicare or Medicaid beneficiaries in healthcare settings across 28 communities. The Camden Coalition assumed leadership of the AHC Model navigation rollout in the participating South Jersey community — providing us firsthand knowledge on the implementation strategies and real-world outcomes associated with the screening and navigation strategy.

As previously highlighted in the webinar, the findings from the third CMMI evaluation were mixed: although hospital visits and healthcare expenditures decreased among patients with HRSN who received navigation, their connections to community services and rates of social need resolution did not improve after the intervention.

Throughout the hour-long webinar, our experts deliberated on the Model’s design, delivery through South Jersey, and the implication of its findings. The panel included:

• Moderator Kathleen Noonan, CEO and President of the Camden Coalition

• Dawn Alley, former Chief Strategy Officer at CMMI and Head of Scale at IMPaCT Care

• Marisol Caban, Associate Director of Care Management Initiatives at the Camden Coalition

• Dawn Wiest, Director of Research and Evaluation at the Camden Coalition

At the conclusion of their discussion, the panel fielded questions from the audience on matters of the evaluation outcomes, the makeup and duration of the Model assessment, and the means to designing a successful community-based navigation program going forward. Their responses to the audience questions are transcribed below, with edits made to clarify their statements.

In case you missed the webinar or want to watch it again, you can now watch the full recording on our website. You can also read a summary of the key takeaways from the webinar here.

Do you think if we were able to have studied the AHC Model longer that some of these social issues would have been able to have been resolved? Was it an issue of timing, in terms of the evaluation and how long we gave social issues to resolve?

Dawn Wiest: The timeline was pretty long for this evaluation. For example: in the assistance track, the outcomes were 12 quarters after screening. So, that’s a pretty long frame of time — that’s three years for Medicaid beneficiaries, and then four years (16 quarters) after screening for the fee-for-service Medicare beneficiaries. But the question is really important because we know from the results of our randomized controlled trial of the Camden Core Model and some of the secondary analyses we’re looking at that the timeline for change could be quite a bit longer than what you anticipate it would be. And regarding community capacity building — that is a long-term effort. It’s probably a lifelong, perpetual effort.

But again, the timelines for this were long. There’s still more to learn, because the claims data for this third evaluation only goes up to a certain point and when they update the Medicaid and Medicare claims data, they’ll be able to look at more long-term outcomes, as well as other groups that may not have been included beyond a year or two into this evaluation.

Kathleen: Regarding the Camden Coalition’s RCT, we would tell you that we thought the timeline that we created was too short. It is true that for some of these studies and some of these assessed health-related social needs, 90 or 180 days (may be too short) — especially if you’re talking about someone that needs housing, which is so core to an individual’s ability to resolve so many other issues. So, you really do have to think about the timeline.

Could you talk about the findings showing that people with substance use disorder were not able to resolve HRSNs as quickly, and what from your experience may explain that trend?

Marisol: It was at times hard to stay connected (to those patients). Loss to engagement or follow-up was a major issue during the rollout because there can be issues with these patients often not having a telephone. It’s important to consider the broader context of their circumstances. The specific challenges they are facing at the time, as well as their stage of recovery, can significantly influence their ability to engage and follow up effectively. It’s important to consider the broader context of their circumstances. The specific challenges they are facing at the time, as well as their stage of recovery, can significantly influence their ability to engage and follow up effectively. When you’re engaging during a screening at the hospital or in outpatient services, it’s easier to have that conversation because they have the time to engage with you. Once you do your follow-up about 2 – 3 days later, there’s a lot that can happen between that time. But nobody ever left without something. So, even if they qualified for navigation and we couldn’t connect with them, we made sure that they had resources around their area. They at least left with something that they can connect to quickly, if and when they needed it.

Kathleen: I think there’s a reason why the Camden Coalition’s navigation demonstrations with people with substance use issues are in-person. That’s where we’re really focusing on providing a CHW paired with an attorney and a health team in a recovery setting. I think it really depends, but certainly we think that that is something that’s critical: that in-person strategy.

Could you expand on how to create an infrastructure for navigation?

Dawn Alley: Absolutely. I think what you’ve heard in some of the great examples that Marisol has shared speak to the fact that this doesn’t just happen. It’s not just a matter of hiring folks from the community and then saying, “Go do good work.” If we want to achieve outcomes and provide person-centered care that is also structured, then we need the infrastructure that relates to hiring the right people, giving them the training they need, giving them workflows, having the analytics so that they can effectively manage caseloads, keep track of just how long someone has been in working with a CHW, and be able to report to funders or have the infrastructure required for billing.

I’ll share an example I recently learned that broke my heart, which is that when the state of South Dakota created a new Medicaid billing code for CHW services, it was five years before anyone was actually submitting claims using that code. And unfortunately, we see that over and over again where policymakers are making very well-intentioned efforts to create sustainable payments for these services, but dramatically underestimating the amount of workforce development that it takes, and the amount of capacity development — whether that’s capacity directly in community-based organizations to be able to participate in this, or the capacity to form those partnerships with clinical sites. What we see at IMPaCT is that if you want to create successful scale, it’s not just about billing for the individual services, it’s about how do you support the workforce and the CHW employers to be able to offer these high-quality services.

Is there another way to conduct navigation-based screening that includes more options than “yes” or “no” for beneficiaries to consider?

Dawn Wiest: The binary option is tricky. Obviously, a “yes” means that the HRSN exists for an individual. But does that mean that each person has the same magnitude of need? Absolutely not. Does it mean that that need has persisted for the same amount of time? Absolutely not. And I think that question of persistence would be really important to address, because some of the critiques I’ve heard about the AHC Model and the strategy of screeners are (regarding the binary outcomes of intervention). If somebody has a food need, they receive a referral to a (nearby) food bank. Does that resolve a persistent food need, or does that only resolve a moment of need? So, I think that is really important to the success of these types of models, is to have a deeper understanding of the magnitude and the persistence of the need identified.

Dawn Alley: And I’m going to make the provocative argument as someone who helped create the AHC’s screening tool, that we should really be moving away from this approach in general. I see a couple of things happening. One is an incredible amount of effort going into ways to increase the rate of screening, which is important because we know that the people who probably need the services the most are going to be people who didn’t get screened — whether it’s because they don’t actually have an encounter with a provider, or because they didn’t complete the entire pre-visit history included in the screening. And another piece is that we’re screening for this very narrow slice of five different needs. This screening tool was created to satisfy an evaluation need. It was not actually created to become the standard of care in the U.S.

So, we need to have room for clinicians to identify a person who needs to have a navigator, even if they didn’t screen positive on one of those five things. And we may want to look at (high-risk) individuals where we see that a person has used the emergency department three times in the last six months or has some other constellation of indicators that they probably need navigation services, but they may never have completed a screening or may not have screened positive for that particular set of needs that we’re looking at. So, I would ask all of us to focus a little less on these narrow screening strategies and a little more on actually asking people what they need.