Social Determinants of Health in Practice: What It Takes to Move From Data to Action

Decades of research have shown that where you live, what you eat, and whether you can pay your bills shape your health just as much as the care you receive in a doctor’s office—if not more. In fact, research results consistently show these social determinants of health (SDOH) as responsible for between 30% and 55% of health outcomes.¹

Although knowing this is a step in the right direction, action is needed to change the landscape. So, what are clinics, hospitals, and community organizations doing with this information, and is it working? What does it really cost to include SDOH in clinical pathways, how do we measure success, and how do we know whether the most vulnerable patients are truly being reached?

Why the Current Landscape Is Different

Over the past few decades, the field has grown from awareness to a body of research documenting the outsized role of factors such as income, housing, and food security in shaping health. However, the transition to action has been slow, and measuring clinical outcomes remains a challenge.

Two programs operating in Houston, Texas, offer a ground-level view of where the field stands today. Harris Health Ben Taub General Hospital is the only public safety-net hospital in the Texas Medical Center, and it is home to the Patient Discharge Initiative (PDI), which screens emergency department (ED) patients for social needs using undergraduate volunteers from Rice University.

Link Health, a nonprofit with operations in Houston and Boston, Massachusetts, embeds credentialed patient navigators inside federally qualified health centers to connect patients with public benefits they are eligible for but are not receiving. Ar’Sheill Monsanto, Link Health’s executive director, explained that each year, more than $142 billion in public assistance goes unclaimed nationally. The program exists, she said, to “get those funds into the pockets of patients.”

Together, these programs represent the full arc of SDOH work in action, focusing on identifying gaps, connecting with resources, and following through.

Michael Jaung, MD, an assistant professor in the Department of Emergency Medicine at Baylor College of Medicine in Houston and a physician adviser to the PDI, reflected on what changed within his own institution. “There was a lot of local awareness that [SDOH are] a big driver for ED visits and also health care utilization in general for our patient population here in Harris County,” he said. The program’s ability to demonstrate early results, such as reduced ED revisits and improved connection to resources, helped build the internal case for sustained investment.

A Closer Look at the Programs

At Ben Taub, the day-to-day reality of SDOH integration looks like this: Undergraduates meet patients in a discharge corridor, using a brief screening tool to identify needs related to food security, transportation, financial assistance, immigration, and legal support. Conversations run 5 to 15 minutes. The most-used resources, Jaung notes, are food security assistance, transportation, and Harris County’s financial assistance program, which is a gateway to the broader safety net.

“When it’s 1 or 2 resources, usually the patients express that they’ve been able to at least start the path to get connected,” Taub said. When patients have 4 or 5 identified needs, only 1 or 2 are typically addressed by the time follow-up calls are made, he said.

At Link Health, navigators move through clinic waiting rooms, initiating brief screening conversations that take under 2 minutes to determine eligibility for programs such as the Supplemental Nutrition Assistance Program (SNAP), Medicaid, and utility assistance.

“For about 4 hours, these patient navigators roam the clinic,” Julissa Barrios, operations coordinator at Link Health, explained. “They go up to patients and ask them if they have heard about Link Health. They then use our dashboard, where we have screening questions that take under 2 minutes.” From there, navigators provide end-to-end application support and manage patients’ cases through approval or denial.

Both programs rely on trained, community-embedded personnel such as volunteers or credentialed navigators who speak patients’ languages and understand the local benefits landscape. “We work really closely with the clinical partners to identify the top 2 to 3 popular languages [in] their neighborhoods, and then we train on that,” Monsanto said.

The Cost Question

The PDI’s cost structure illustrates one end of the spectrum. Because the program runs on volunteer labor, Jaung said, direct human resources costs are minimal and cover supplies such as paper clipboards, a secure bin for patient information, and annual training materials for incoming volunteers. However, he noted that PDI has not yet conducted a full cost analysis of formally employing people in these volunteer roles—whether as community health workers, medical assistants, or patient navigators—though he added that “it would be interesting to see.” Evidence from other SDOH-focused programs in the country shows a favorable cost-benefit analysis. A Health Affairs analysis of the Individualized Management for Patient-Centered Targets, or IMPaCT, community health worker intervention, which is a rigorously evaluated, standardized model tested in 3 randomized controlled trials, found that every $1 invested returned an average of $2.47 to the Medicaid payer within the fiscal year, driven primarily by a reduction of hospitalizations.²

Other analyses of SDOH-focused community partnerships project that sustained programs could avert hundreds of millions of dollars in medical costs and productivity losses over a 20-year period.³ The caveat is that these returns often accrue to payers, while the program costs fall on health systems, community organizations, and grant-funded nonprofit institutions. This misalignment is central to the sustainability problem.

Link Health is navigating this directly. The organization is primarily philanthropy-funded but holds government contracts as a state-certified SNAP outreach partner in Texas and Massachusetts. It is also exploring fee-for-service arrangements with payers, building on CMS billing codes that now allow reimbursement for community health integration services.

“We’re having these conversations with our partners, and even some payers, where they’re noticing an uptick of families that are having housing issues,” Monsanto said, “and they’ve brought in Link Health to liaise.”

Measuring Outcomes

Neither PDI nor Link Health fully knows the actual clinical outcomes of the programs yet, as these data are difficult to find. This data chasm is a key gap in bringing SDOH into the mainstream. The PDI currently measures process outcomes, including ED revisits at 30 and 90 days, follow-up appointment adherence, and patient-reported resource use. Jaung acknowledged wanting to go further.

“We’ve thought about tracking blood pressure management, hemoglobin A_1c [HbA_1c], and more, but the apparatus and bandwidth we have in quality improvement [have] not allowed us to get to those clinical outcomes,” Jaung said. Data barriers compound the challenge. Additionally, patients who visit Ben Taub’s ED may follow up at a low-cost clinic outside the Harris Health system, making longitudinal tracking structurally difficult.

Even with the data, isolating the contribution of a single SDOH intervention from the broader continuum of care changes is difficult. Jaung noted, “I think it would take a really large data set to be able to sort of tease out all these different specific variables.” The logical connections are clear: Being able to afford medications, access nutritious food, and attend follow-up appointments supports treatment adherence. However, establishing rigorous causal chains in open, nonclosed health systems is a different matter entirely.

The broader literature reflects this ambiguity. A 2025 scoping review of 41 US-based hospital SDOH initiatives found that although 68% of studies assessing clinical outcomes showed improvement, nearly 25% showed no change, and 10% reported mixed results.⁴ Jaung noted that disease-specific programs, such as those targeting HbA_1c in diabetes or blood pressure in hypertension, tend to produce the clearest signals, but even there, results are inconsistent.

Link Health plans to track some of these biometrics to add quantitative data to the program by enrolling approximately 300 patients across Boston clinic sites and tracking clinical outcomes, such as lipid panels, HbA_1c, and cardiovascular risk scores, at 3 and 6 months, alongside dietitian support and structured grocery access.

Large-scale data efforts are just beginning to identify associations between SDOH and specific clinical risks, which are crucial for developing targeted interventions and, eventually, improving outcomes. A preprint analysis of the National Institutes of Health All of Us Research Program, which draws on data from more than 372,000 participants, used network analysis to identify 4 distinct SDOH subtypes among individuals with co-occurring social needs.¹ The highest-risk subtype, characterized by the simultaneous presence of low income, food insecurity, housing instability, unemployment, low literacy, and low educational attainment, carried an OR of 4.2 for depression compared with the lowest-risk group and was also significantly associated with delayed medical care and ED visits.

Crucially, this cluster was more likely to include patients living in states without Medicaid expansion, compounding an already heavy burden. What the analysis makes clear is that people rarely experience a single SDOH in isolation, as the median participant had 9 co-occurring social risk factors, and racial and ethnic minority participants had significantly more than their White counterparts.

The Equity Lens

A critical factor in SDOH programs and research is the ability to ensure equitable distribution, as higher-literacy, more engaged patients tend to get connected to resources first. For the PDI, a significant barrier is limited English proficiency, independent of health literacy. In related work at Harris Health, Jaung found that language, rather than ethnicity, was the primary driver of patients’ existing medical knowledge and interest in preventive services, which has clear implications for program design.

Jaung noted, “Presenting materials in the patient’s preferred language, communicating and counseling patients in their preferred language, [and] having native Spanish speakers as volunteers or health care workers” are not optional add-ons but rather necessities to ensure the program reaches those who need it most. He added that the basic screening model may not reach patients with psychiatric and substance use disorders, requiring wraparound services that go beyond what a discharge corridor conversation can provide.

Monsanto grounds Link Health’s equity strategy in proximity to patients. By placing navigators in federally qualified health centers in underserved zip codes, hiring staff who reflect the communities they serve, and training them not just in application processes but in community engagement, Link Health aims to serve the communities that need it most. However, rural populations represent a structural gap that is difficult to fully address. Monsanto noted that funding cuts to Medicaid and the consequent loss of hospital-based community health workers in rural communities created openings for organizations such as Link Health, but also illustrate the fragility of the broader ecosystem. When program staff disappear, so does the connective tissue between identified needs and available resources.

What’s Still Missing

Both programs, despite their longevity and reach, share a common vulnerability: no durable payment model. The PDI runs on volunteer labor and Harris Health’s institutional support. Link Health runs on philanthropy and government contracts that require constant renewal. Sustainable funding is necessary for the continued growth and support of these programs.

What these programs have demonstrated, however, should not be minimized. They show that SDOH integration is operationally possible at scale, even with limited financial resources, and that patients want and use social resources when they are made genuinely accessible. Although the medical field is no longer asking whether social care belongs in clinical settings, the questions now are more difficult and also more practical: Who pays for it, how do we prove it worked, and how do we make sure it reaches the people who need it most?

References

1. Bhavnani SK, Zhang W, Bao D, et al. Subtyping social determinants of health in All of Us: network analysis and visualization approach. medRxiv. Preprint posted online August 11, 2023. doi:10.1101/2023.01.27.23285125

2. Kangovi S, Mitra N, Grande D, Long JA, Asch DA. Evidence-based community health worker program addresses unmet social needs and generates positive return on investment. Health Aff (Millwood). 2020;39(2):207-213. doi:10.1377/hlthaff.2019.00981

3. Honeycutt AA, Khavjou OA, Tayebali Z, Dempsey M, Glasgow L, Hacker K. Cost-effectiveness of social determinants of health interventions: evaluating multisector community partnerships’ efforts. Am J Prev Med. 2024;67(6):916-923. doi:10.1016/j.amepre.2024.07.016

4. Rangachari P, Thapa A. Impact of hospital and health system initiatives to address social determinants of health (SDOH) in the United States: a scoping review of the peer-reviewed literature. BMC Health Serv Res. 2025;25(1):342. doi:10.1186/s12913-025-12494-2