The Accountable Health Communities Model

Find all supporting materials at the Hunger Vital Sign explainer series website.

This episode features an interview with Katherine Verlander, Deputy Division Director at the Centers for Medicare & Medicaid Services (CMS)

Part One of this series featured Children's HealthWatch - find their Hunger Vital Sign materials and background research here.

Part Two of this series featured Hunger Free Vermont and their work implementing Hunger Vital Sign in Vermont.

Part Three of the series introduces the screening, referral, and navigation services evaluated as part of the Accountable Health Communities Model at the CMS Innovation Center.

Audio Editing and Post-Production Provided By Evergreen Audio

Transcript

LABUN:

Welcome to Episode 11 in our series of short explainers for the Hunger Vital Sign tool. We’re partway through our exploration of Hunger Vital Sign out in the wild – in this case as it appears in the Accountable Health Communities model. We have a guest expert, Katherine Verlander.

VERLANDER:

 My name is Katherine Verlander and I am the Deputy Division Director for a Division in the CMS Innovation Center that houses the Accountable Health Communities model.

LABUN:

Last episode. . . well, last episode we basically got through explaining Katherine’s introduction. We went over what CMS is, the goal of their Innovation Center, and the basics of the Accountable Health Communities model. 

This episode, we’ll go deeper into the screening tool connected with the AHC Model. The key things to know are that the Hunger Vital Sign appears in a screening tool that covers a range of social risk factors. And this tool is embedded in a process of screening, referral, and navigation to services carried out by entities called Bridge Organizations. 

Now, we always knew it was artificially narrow to look only at food insecurity – health care providers and their patients deal with a whole range of health-related issues. To be able to account for all factors shaping a person’s health would require omniscience. Pretending that just one factor is the key doesn’t match real life experience. So, how did the CMS Innovation Center find a middle ground and choose their core areas for screening? 

VERLANDER:

Yeah, there are many non-medical needs that impact health and we only selected five core health related social needs domains for focus in the Accountable Health Communities Model. And we selected those based on the following criteria. One, there had to be high quality evidence existing that linked the health-related social need to poor health or increased healthcare utilization and cost. Second, the need has to be something that can be met by community service providers. And third, the need is not already systematically addressed by healthcare providers.

LABUN:

Basically, this ties the core domains back to the central thesis of the model from last episode –there are gaps in health care currently (needs not addressed by health care providers); these gaps could theoretically be filled through referral and navigation (they’re something that can be met by community services providers); there is evidence to suggest that integrating these services could meet the Innovation Center’s underlying purpose of improving health quality while reducing health care costs. 

Following this construct the researchers chose 5 core domains for Bridge Organizations to address. 

VERLANDER:

. . . The health-related social needs screening covers a minimum of five core domains. They were allowed to add extra, but the five core domains are housing, food security, transportation, utilities, and interpersonal violence

LABUN:

I’ll point out that this run of logic wasn’t inevitable. An example of a different logic model might be that a community has food security as a top priority, and so chooses risk domains they see as most closely related to that priority – maybe that’s food insecurity plus transportation, physical conditions that limit mobility, access to family or community support for daily tasks, and emergency preparedness.  

Next episode we’ll get into resources that help communities implement their own approach while still following common guidelines and benefiting from the model’s lessons learned. 

For now, let’s focus on screening questions to go with the 5 core domains. We already learned in the Hunger Vital Sign process how difficult it is to build validated questions from scratch, and by 2016 when the Innovation Center was setting up this AHC Model there were already a lot of screening tools in use. 

VERLANDER:

So through this process, over 50 screening tools, totaling more than 200 questions were compiled. And in order to refine this list, we consulted this technical expert panel and the panel consists of a diverse group of tool developers, public health, and clinical researchers, clinicians, population, health, and health systems, executives, community based organization leaders, and other federal partners.

LABUN:

The panel relied on another set of three guiding principles to refine the list. I’m sure there’s a statistics students somewhere who can take the two sets of three guiding principles and the five core domains with the fifty screening tools and figure out how many alternate universes might have been created with a different set of decisions. In this universe, which I like to imagine is the optimal universe, here’s how it went:

VERLANDER:

In developing the AHC HRSN screening tool for the five core domains, we followed three guiding principles. First the tool needed to consistently identify the broadest set of health related social needs that could be addressed by community service providers. So ideally the tool would allow clinical delivery sites to identify broad needs and then navigators and community service providers who are better suited to identify the more specific needs and address the individualized issues that underlie the unmet need.

LABUN:

We’ve seen this principle before in Hunger Vital Sign, the idea that you want a “first step” screening that identifies a category of risk – like food access – but allows further specification, the diagnosis if you will, to happen in consultation with the individual.

VERLANDER:

So for example, the AHC HRSN screening tool would identify that an individual is at risk for food insecurity broadly, and then the care navigator or the community service provider would identify the individual's specific drivers of food insecurity, like difficulty obtaining food near end of the pay period, or maybe it's a daily need for food assistance.

LABUN:

Another point in this first principle is that it gives the individual being screened an opportunity, after the first broad questions, to say whether they want to go down the path of further describing their needs and receiving assistance. Not everyone wants that.

Okay, second principle. 

VERLANDER:

The second guiding principle for developing the tool was that the tool needed to be simple and streamlined to ensure that its questions were readily understandable by the broadest audience across a variety of settings, as well as to allow for inclusion of routine screening and busy clinical workflows. So Medicare and Medicaid beneficiaries represent a really diverse subset of the us population and include individuals of all ages and backgrounds. And AHC was implemented in all types of clinical delivery sites. So maybe a high volume emergency department or a low volume rural FQHC of federally qualified health center. So really had to have broad applicability. It's also important to note that they worked to make the tool accessible to beneficiaries regardless of their language, their literacy level or their disability status, which again is important in broadening its applicability. So additionally, because AHC clinical delivery sites had the option to allow beneficiaries to self administer the screening tool to so to take it themselves, without having somebody reading it allowed to them simplicity and the design and the language were key considerations in order to reduce the need for outside assistance.   

LABUN:

We discussed this principle in Episode 4 on reliability. A tool needs to work in a variety of settings, with a range of patients.

VERLANDER:

The last guiding principle was that the tool needed to be evidence based and informed by practical experience.  

LABUN:

We’ve been defining this final element for Hunger Vital Sign. A 2022 report from the Social Interventions Research and Evaluation Network, or SIREN, offers an overview of all the ways “evidence based” and “informed by practical experience” are defined in research on social risk screening. Or, as the authors put it, psychometric validity and pragmatic validity. Of course, the expert panel looking into possible tools in 2016 didn’t have the advantage of the 2022 SIREN report, which suggests one challenge in these long-term projects – CMS is studying a field that’s evolving, with or without a final Model evaluation. We’ll talk more about navigating the varying paces of change in the next episode. 

There’s another challenge hidden in this overview of guiding principles. The AHC Model requires that all screening be followed by connection to community resources that can address individuals’ expressed needs. But there’s also the same 5 core domains for screening in every location. How do you guarantee the needed services exist across those domains?

VERLANDER:

The reason the CMS Innovation Center exists is to help improve the quality and the value of healthcare. The underlying concept of the accountable health communities model test is that identifying and addressing health related social needs has the potential to improve healthcare outcomes and reduced total cost of care screening is a means to an end identifying needs so they can be addressed through a referral and or navigation services. Screening is not an end in and of itself. We don't ask these questions out of curiosity or so we can have another data element to measure or to put into an algorithm to attempt to predict outcomes. We ask these questions to identify needs that we can help with there's broad agreement in the field about this. And it also makes for good patient-centered care. You wouldn’t want to ask somebody about a personal hardship and then not do anything in response. I saw a toolkit come out from the American academy of pediatrics and the food research and action center that really drives this home. The toolkit is to help pediatricians address food insecurity. They recommend the hunger vital sign by the way, and they titled the reports screen and intervene. And I think that really sums it up nicely that, you know, the, the purpose is not just to screen, but you know, you always want to screen and intervene.

LABUN:
 One way that the Innovation Center knew the Bridge Organizations had services to bridge to is simple – there was an application process. Being able to address the full range of potential social needs was a requirement for participating. 

But that requirement was for the trial period. The screening guide built for communities beyond the initial Model participants includes options for modification to match the local environment. 

VERLANDER:

When you look at the screening tool on our website and you look at the screening guide on our website, you'll see that there are really two sections of the tool. There's one section of the tool that's for core needs. And there's another section of the tool that's for what we call supplemental needs. The core needs section includes questions related to the five health related social need domains that we determine to be core needs. All the accountable health community model awardee are required to screen for those five core domains. And those are living situation, food security, transportation, utilities, and safety Organizations using the screening tool outside of the AHC model can determine which domains to include in their screening based on their unique constellation of priority health related social needs in their communities, and also the availability of community resources to address them.

LABUN:

Part Two of this explainer series went over ways to use the screening process to build more resource connections, last episode discussed that element in the Alignment Track communities for AHC, and case studies on the AHC website provide more examples. 

Many of the early lessons learned from the screening process are useful even if your community would need to alter the domains to meet all steps in screening and connecting to resources. For example: 

VERLANDER:

.  . . Of those screened in the AHC model so far about 34% had a health-related social need had one of the core health related social needs. Food insecurity was the most commonly reported health related social need followed by housing and transportation. And the most commonly co-occurring needs were food and housing and food and transportation.

LABUN:

Complicated statistics are hard in audio format, but here’s one example - among patients offered navigation services, the rates of food insecurity were as high as 82% in some locations. Much higher than the general community levels. Also, among this group, almost 60% had multiple needs, with either transportation or housing supports as the most common second need. 19% reported all three. If I’m in a health care system new to addressing health-related social needs, information like this suggests that I could start with mapping community resources across food access, housing, and transportation. That sets a foundation for early screening and referral, and I’m targeting based on likely patient needs - so if gaps emerge during the mapping process, it’s reasonable to anticipate that investing in closing those gaps will be a useful service. The next episode goes into more examples of possible ways to use the early AHC Model results. 

VERLANDER:

Another finding is that the AHC model is effectively identifying higher cost utilization beneficiaries by screening for one or more health related social needs and two or more ED visits in the last 12 months.  

LABUN:

This is good news for implementation. Screening for social needs and emergency department visits can identify patients with overall high health care costs without running a detailed claims analysis for everyone who walks through the door. Last episode we noted that the costs we’re interested in are connected to avoidable use of services. I realize that “emergency department” implies something unavoidable – an emergency. But sometimes that emergency could’ve been prevented through earlier treatment – for example if transportation barriers lead to deferring care until it’s an emergency. Or it’s a health issue made acute through other complicating factors – like a serious asthma attack brought on through poor air quality in substandard housing. Or maybe the emergency department is the most accessible place to reach for non-emergency needs. It’s open 24 hours a day and it’s where the ambulances drive to. 

Also, the combined factors of self-reported Emergency Department visits and a health-related social need predicted higher expenditures better than either did when taken alone. This evaluation element is similar to the Hunger Vital Sign work, where the researchers tested how far they could simplify the process before the results were outside their parameters for validity. 

VERLANDER:

And beneficiaries eligible for navigation are accepting navigation at higher rates than anticipated. We anticipated around 40% of beneficiaries eligible for navigation would accept it. And actual acceptance of navigation was much higher, around 74%, according to the first evaluation report.

LABUN:

There are many reasons why someone would turn down navigation services. That’s not always a bad thing. Patients may have screened positive for a social risk, but don’t need assistance beyond a referral. Maybe they’ve already addressed the concern – remember, the food insecurity screen looks back a year. Or they have other places they already go for assistance – whether that’s a formal system, like an Area Agency on Aging or a Community Action Program, or an informal system across family, friends, and neighbors. Or the reason might be less optimistic, like fear of stigma or a fundamental distrust of the medical system. These are obstacles the Bridge Organizations can help address. 

Next episode will look in more detail at early lessons learned from the Accountable Health Communities Model and what we can expect to learn from future evaluations. To review this episode’s key points:

·      The AHC Health Related Social Needs screening tool has 5 core domains and 8 supplemental domains. 

·      The 5 core domains are housing, food security, transportation, utilities, and interpersonal violence. Hunger Vital Sign is used for the food security screen.

·      In the AHC Model screening does not exist by itself. All organizations applying to participate had to show they could follow the screen with offering referral and navigation to community resources.

·      Community resources needed to include the 5 core domains but didn’t need to be perfect – the Model assumed the process would reveal gaps in community resources. The Alignment Track, which we described last episode, required active engagement by the Bridge Organization in identifying those gaps and building a response to fill them.

Follow the link in our Show Notes to find the materials referenced in this episode.