Patients Redesign the Questions Hospitals Ask About Their Social Needs

Every year, hospitals across the United States ask patients a set of deeply personal questions: Can you afford food? Is your housing stable? Do you have a way to get to your appointments? These questions about health-related social needs, or HRSNs, have become a cornerstone of modern medicine’s effort to look beyond the exam room and address the conditions that actually make people sick. Health systems have poured an estimated $2.5 billion into screening and referral infrastructure, driven by federal mandates and a growing recognition that housing instability, food insecurity, and transportation barriers are as consequential to health as any prescription. Yet there is a striking flaw at the heart of this enterprise: the screening tools themselves are often designed with little or no input from the very patients they are meant to serve. A new initiative at Boston Medical Center shows what happens when that oversight is corrected, and the results offer a blueprint for how hospitals everywhere might rebuild trust into the act of asking for help.

Boston Medical Center, New England’s largest safety-net hospital, occupies a unique position in American medicine. More than 70 percent of its patients identify as people of color, over half live in households earning less than $20,400 a year, and roughly one-third speak a primary language other than English, most commonly Spanish or Haitian Creole. This is precisely the population that bears the heaviest burden of unmet social needs. Prior research has documented that racially and ethnically minoritized groups experience significantly higher rates of food insecurity, housing instability, threats of utility shutoff, and cost-related barriers to medical care compared with White adults. These disparities are not accidents of circumstance; researchers describe them as manifestations of structural violence, the social arrangements embedded within political and economic systems that systematically place certain populations in harm’s way. Any screening tool that fails to account for this reality risks missing the needs it was built to find.

The hospital’s screening program, known as THRIVE, had been integrated into clinical workflows since 2017. But as the years passed, process evaluations revealed a persistent problem: patients frequently struggled to interpret and respond to specific questions during primary care visits, resulting in incomplete or inaccurate data capture. The tool worked on paper but faltered in practice. So between February and August 2023, a team led by researchers at Boston Medical Center and Boston University Chobanian & Avedisian School of Medicine took a fundamentally different approach. They recruited 37 patients who had completed the THRIVE screener during routine primary care visits and invited them to become co-designers, not merely feedback providers. The participants spoke English, Spanish, or Haitian Creole, and the design sessions were conducted in their preferred languages by bilingual facilitators trained in qualitative methods and community engagement.

The mechanics of the initiative reveal how deliberately the team dismantled the usual barriers to participation. Six two-hour design sessions were held in hospital conference rooms after 5:30 in the evening to accommodate work schedules. Patients were recruited through existing advisory councils and clinic-based networks, then supported with transportation arranged through Uber Health, paid childcare when needed, follow-up reminders, and $50 compensation per session plus refreshments. Two sessions were facilitated in English with ten participants, two in Spanish with fifteen, and two in Haitian Creole with twelve. Crucially, the researchers framed participants as experts with lived experience whose time and expertise deserved financial compensation, a framing that stands in sharp contrast to the extractive dynamic that characterizes much clinical research with marginalized communities.

Each session followed a structured and deceptively simple format. After introductions and background on the THRIVE initiative, participants received a printed copy of the existing screening tool and were asked to complete it as though they were sitting in their primary care appointment. This mock screener exercise simulated the real clinical encounter and surfaced confusions that might never appear in a developer’s office. Facilitated discussions then tackled the language, tone, layout, and usability of the form, along with preferences for opt-out options, expectations about how the information would be used, and reactions to new items on social isolation, employment, and education. Participants provided verbal recommendations in their primary languages, documented in real time, with visual mockups and plain-language summaries supporting accessibility across literacy levels.

Three themes emerged with remarkable consistency across all three language groups, and the first was the most sobering: trust and psychological safety. English-speaking participants voiced fears that certain answers could trigger punitive responses, including reports to child protective services or loss of insurance coverage. Spanish-speaking participants noted that some family members, particularly men, might avoid answering altogether due to stigma or shame. Haitian Creole-speaking participants were especially concerned about how their responses could affect their immigration status or legal standing. Across every group, participants demanded clear explanations of how their data would be used and protected. The lesson is uncomfortable but essential: a screening question is never just a question. For patients whose lives intersect with child welfare systems, immigration enforcement, and precarious insurance, the act of disclosure carries perceived risks that designers who have never faced those systems cannot anticipate.

The second theme, clarity and usability, was more prosaic but no less consequential. Participants consistently preferred a single-page screening tool and viewed longer or multi-page forms as burdensome and confusing. They requested larger fonts, cleaner formatting, and clearer instructions about which sections were optional and which were required. Recurrent confusion about the form’s flow, particularly where sections ended or whether both sides of a page needed to be completed, emerged as a persistent barrier. Participants also pushed for plain language and the avoidance of overly technical or repetitive terms. These may sound like minor design quibbles, but in a safety-net setting where a third of patients are navigating the form in a second language, every point of confusion translates directly into missing data and unmet needs.

The third theme, cultural and contextual relevance, produced some of the most concrete redesign recommendations. English-speaking participants asked for the removal of infantilizing language and the reframing of questions that felt accusatory. Haitian Creole-speaking participants stressed the importance of asking about work authorization in a non-threatening manner and called for clarity around the potential immigration consequences of accessing benefits. Spanish-speaking participants highlighted confusion for those who help patients complete forms but are not formal caregivers. On social isolation, participants across all groups supported the inclusion of new questions but suggested alternative phrasing, such as asking whether patients feel connected to their community, to improve clarity and reduce the sting of the inquiry.

The hospital acted on what it heard. The revised THRIVE tool received a rewritten disclosure statement clarifying purpose and confidentiality, questions reframed to begin with the supportive phrasing Do you need support with, and new items capturing housing quality issues, immigration status, financial hardship, and digital health needs. The tool was shortened to 19 data points to reduce respondent burden, its layout was redesigned to guide patients through a more intuitive sequence, and it was paired with multilingual resource guides offering curated, actionable referrals across 15 key domains. The team is now conducting content and psychometric validation of the co-designed instrument, though the published report focuses on the co-design process itself. The authors acknowledge a key limitation: no post-implementation evaluation has yet assessed how the redesigned tool performs in actual clinical practice, leaving an important question for future research.

The implications extend well beyond one Boston hospital. Existing screening instruments, even those described as validated, have not been adequately tested in safety-net settings or in racially and ethnically diverse populations, and there is evidence that some tools under-identify needs among Black patients compared with White patients despite broad clinician consensus on their value. Patients, the research consistently shows, are more willing to disclose sensitive social information when the process feels trustworthy, transparent, and supportive rather than bureaucratic or extractive. The Boston findings align with national priorities including Healthy People 2030 objectives, the Centers for Medicare & Medicaid Services Accountable Health Communities Model, and National Committee for Quality Assurance standards for social risk adjustment. Scaling this approach, the authors caution, will require real infrastructure: multilingual facilitation, participant compensation, and institutional commitment to iterative refinement based on community feedback. But the core insight is simple enough to fit on a single-page screener. Screening for social needs is not a data-collection exercise; it is a relational process built on trust, cultural responsiveness, and transparency. Hospitals that treat patients as co-designers rather than data points can identify the barriers of psychological safety, clarity, and cultural relevance before they undermine large-scale implementation, and in doing so, they may finally capture the full picture of what their patients need to be healthy.

Subject of Research: Community-participatory co-design of a multilingual health-related social needs screening tool at an urban safety-net hospital

Article Title: Centering Multilingual Patient Voices in the Co-Design of a Health-Related Social Needs Screening Tool in an Urban Safety-Net Hospital

Article References: Khemraj, U. D., Solomon, P., Eddy, R., Olesinski, E. A., Varela-Poole, K., Phicil, S., Muoto, I., Sandel, M., & De La Vega, P. B. (2026). Centering Multilingual Patient Voices in the Co-Design of a Health-Related Social Needs Screening Tool in an Urban Safety-Net Hospital. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10758-z

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10758-z

Keywords: health-related social needs, social determinants of health, participatory design, safety-net hospital, health equity, screening tools, multilingual patients, patient engagement, Boston Medical Center, primary care, cultural relevance, psychological safety