With the highest number of people experiencing homelessness in the country, California health care leaders and policymakers are seeking new ways to improve health outcomes for this population. Illumination Foundation, a nonprofit organization in southern California, created an integrated care model that offers recuperative care, emergency shelter, temporary and permanent housing, and primary care for their clients experiencing homelessness who have complex needs such as chronic diseases, infections, mental illness, and substance use disorders. Their care model uses risk profiles and social determinants of health (SDOH) data to improve care for this population.
The Better Care Playbook recently spoke with Pooja Bhalla, DNP, RN, Co-Chief Executive Officer, Illumination Foundation, and Ashish Abraham, MD, MBA, President and Co-Founder of Foresight Health Solutions, and a consultant to Illumination Foundation, to learn about Illumination Foundation’s approach to data analytics and how it can inform other organizations looking to improve care for people experiencing homelessness.
Can you describe Illumination Foundation’s integrated care model for people experiencing homelessness? How does it address SDOH?
"Our goal was to connect individuals to health services, and we very quickly realized that there really isn't a [health care] delivery system for people experiencing homelessness."
Pooja Bhalla, DNP, RN, Co-Chief Executive Officer, Illumination Foundation
P. Bhalla: When we talk about an integrated care model, it is about meeting individuals where they are. Illumination Foundation started off with a small presence in a homeless shelter. Our goal was to connect individuals to health services, and we very quickly realized that there really isn't a delivery system for people experiencing homelessness. Now, we provide care and services for individuals at shelters, on the streets, and in our medical respite program. Illumination provides the full continuum of care, integrating housing and health care to meet the complex needs of our clients. Say someone is coming in from the street — they don't have access to health care, they don't know how to get into housing, they don't have food or childcare. We're addressing SDOH when they enter into our system by providing them with transportation to medical appointments and access to primary care, psychiatric and mental health services, dental care, and substance use care, all within the same place. We also connect them to housing through our case managers and housing navigators. Once they move into housing, they still have access to their primary care doctor in the same building as the other services. The doctor can also connect with them via telehealth or do house calls.
What does a typical individual served by Illumination Foundation look like?
P. Bhalla: We see folks who have experienced a lot of trauma, have no access to health care, and have multiple chronic medical issues. These individuals often cycle through inpatient settings, hospitals, and emergency departments (EDs). They come to us either directly from the street, through a hospital, or through an outreach team. When they're coming to us, they have no idea they could even access health care or they don't know how to get into housing. About 85 percent of the population that we see has diagnosed mental health conditions. We cover the whole spectrum from people experiencing childbirth to palliative care. In our family programs, we see a lot of single moms with their kids that are working two or three jobs. We're also seeing a lot of seniors in our settings. We have 85-year-olds living in our shelters who were dropped off outside in a wheelchair.
Can you describe how Illumination Foundation uses data to identify care needs and guide care management interventions for its clients?
P. Bhalla: At first, we were collecting data on Google Sheets. Those data helped inform us how to build a better system and understand the populations to serve and what our data could show us. When I met Dr. Abraham, we joined forces and started looking at all the data we're collecting. We now have a methodical approach of collecting data versus before when we were just collecting as much as we could.
A. Abraham: Using data for informed care decision-making is an incremental journey. The first step in this journey is having a system that captures data. The minute you have that in place, you realize you start getting an overwhelming amount of data. And within electronic health records (EHRs), we find fragmentation.
The EHR that Illumination Foundation used was geared toward behavioral health and substance use, while other information systems were being used for housing. Illumination Foundation engaged Foresight to help the organization move toward becoming more data-driven in their decision-making and to integrate data across health and social needs. Artificial intelligence (AI) and machine learning analytic tools were important to guide this process in three ways:
- First, these tools helped us aggregate data from disparate sources and create a single, person-centered, chronological record that identified the full spectrum of needs.
- Second, we focused on the analytic capability to predict which clients are at the greatest risk for future adverse outcomes based on all of these very complex, sometimes interrelated drivers of health. This helped the team prioritize where they should focus time, energy, and staffing resources.
- Finally, we used prescriptive analytics to combine all of the interventions and the actions that are being tracked in the health record, and to evaluate which of the services that are being provided in this integrated model have the best likelihood to help mitigate or reduce risks. That's where you truly become a data-driven organization that is able to not only identify which of the clients are most at risk, but also pinpoint the specific actions and interventions that can influence at both the cohort and individual patient level.
That's the true holy grail — I think everyone is trying to influence health in a prescriptive fashion, using the power of data to help support impactful care decisions.
What are you learning from combining health care and SDOH data?
A. Abraham: Capturing SDOH data is not easy. When a client is being asked many questions directly, sometimes they are very hesitant to share some of these deep social vulnerabilities that they struggle with. But the relationship that the health navigators, case managers, and housing navigators at Illumination Foundation have established with these patients builds trust. In those trusting relationships, patients are revealing so much information around SDOH outside of the clinical screenings. Illumination Foundation has started using an AI and machine learning tool called natural language processing to scour through the text of the notes. It’s staggering that sometimes 40-50 percent more social risk factors are identified through those relationship-based conversations compared to standardized clinical assessments.
"It’s staggering that sometimes 40-50 percent more social risk factors are identified through relationship-based conversations compared to standardized clinical assessments."
Ashish Abraham, MD, MBA, President and Co-Founder of Foresight Health Solutions
Another learning that’s important from a policy perspective is the whole idea of risk adjustment for reimbursement models. There's been a strong push to have risk adjustment as a key component in the methodology for rate determination for Medicaid and Medicare, yet models that are used for risk adjustment today are predominantly based on diagnostic risk factors. One fascinating finding is when we started incorporating SDOH factors in predictive risk analyses, the model became far more accurate than when based upon diagnostic and demographic factors alone. We see accuracy jumping to 90-93 percent of actual observed costs when we incorporate SDOH data. For Illumination Foundation’s patients, SDOH factors like homelessness and food insecurity are significant drivers of health care complexity and unless reimbursement rates account for these risks, the rates don’t adequately cover the full continuum of services these patients require.
What have been the impacts of Illumination Foundation’s integrated care model, including on cost and utilization?
P. Bhalla: Overall, our clients experienced decreased ED utilization, increased primary care utilization, and more connections to social services. Our integrated model improves care and reduces cost to the overall health system.
A. Abraham: There was a 70 percent increase in primary care utilization between 2014 and 2019. There is value in providing this historically marginalized group with primary care access, as people experiencing homelessness have traditionally not had access to primary care in the past.
What does that mean in terms of the actual savings that a system could gain? Using claims data over a three-year period for a population of about 1,250 clients, the study evaluated both cost and utilization results over this longitudinal period. The results measured client medical costs and acute care service utilization one year before entering the program and compared those to costs and utilization levels one year after exiting the Illumination Foundation program. We saw costs before they entered the program of around $25 million, and after they completed services, their costs were $17.7 million. That's a total cost savings of over 30 percent after Illumination Foundation’s intervention, driven by reductions in acute care. Hospital admissions went down by about 26 percent and ED visits by 22 percent. However, getting to these cost savings requires providing much needed medical, behavioral, and social services to be delivered to clients that are marginalized.
Do you have any advice to share with other organizations working to provide housing and health care to people experiencing homelessness?
"I think we should focus on patient advocacy as the foundation of our models, but we need to balance that with an equally strong focus on effectiveness and efficiency of care services."
Ashish Abraham, MD, MBA, President and Co-Founder of Foresight Health Solutions
P. Bhalla: I think first and foremost, collecting the right data is critical. Second, I would say building partnerships with your health and managed care organizations is really important. And third is figuring out how to use the data to inform the care model. We're on the right track, but still have some work to do.
A. Abraham: In this journey toward becoming a true data-driven, decision-making organization, everyone is at different stages. I think it takes critical leadership, ownership, and really driving that culture within the organization. It is not easy, especially in the social services arena. It's very much driven by objective cost parameters because your entire business is at risk. On the provider side, I think we should focus on patient advocacy as the foundation of our models, but we need to balance that with an equally strong focus on effectiveness and efficiency of care services. Investments in both the right staff, systems, and analytic tools are needed to be able to allow this framework to succeed.