Social Risks Shift Over Time in Specialty Clinics, Study Finds

Patients do not arrive at a rheumatology clinic as a fixed snapshot of circumstance. Jobs are lost, housing situations change, transportation falls through, and food insecurity can appear or disappear between appointments. A new research report published in the Journal of General Internal Medicine by a team led by Daphne Lew of Washington University School of Medicine examines precisely this dynamism, tracking how social risk factors, the non-medical circumstances that shape health and health care, change over time within a specialty outpatient population. The study’s central message is deceptively simple but consequential for anyone designing screening programs: a patient’s social risk profile at one moment is not a permanent label, and screening done only once risks missing both emerging needs and needs that have been resolved.

Social risk factors, often discussed under the umbrella of social determinants of health, encompass a range of conditions including housing instability, food insecurity, transportation barriers, difficulty paying for utilities or medications, unemployment, and limited social support. Decades of research have linked these circumstances to worse outcomes across chronic disease. In rheumatoid arthritis, systematic reviews have documented associations between social deprivation and higher disease activity, while studies of type 2 diabetes have shown that social determinants influence glycemic control and complication rates. Population-level analyses across multiple US states have found that arthritis burden clusters among people with more social risk factors. The implication is that a clinic treating inflammatory disease cannot fully understand or treat its patients without understanding the social terrain they navigate daily.

The new study emerged from a quality improvement initiative in rheumatology clinics at Washington University in St. Louis, where patient-reported social risk factor screening was embedded into routine outpatient care. Rather than treating screening as a one-time intake exercise, the clinics collected responses repeatedly, generating longitudinal data within the electronic health record. The research team, which included investigators from Washington University, its Bursky School of Public Health, the University of Wisconsin School of Medicine and Public Health, and Yale School of Medicine, extracted de-identified screening data to ask a question that most cross-sectional studies cannot answer: how stable are social risk factors in a specialty outpatient population, and what does that instability mean for health systems?

The technical rationale for longitudinal screening rests on how social needs are measured and used. Health systems increasingly adopt standardized screening instruments, prompted in part by federal momentum such as the Accountable Health Communities model described in the New England Journal of Medicine and by Centers for Medicare and Medicaid Services guidance encouraging social risk screening. These programs typically rely on patient-reported answers to short questionnaires covering domains like food, housing, transportation, and utilities. When such data are recorded once, they enter the chart as static fields. Downstream uses, from risk stratification to resource referral to quality metrics, then inherit the assumption that yesterday’s answer describes today’s reality. The Washington University analysis challenges that assumption by examining within-patient change across repeated screens.

Why would social risk status change in a specialty population? The patient mix offers one explanation. Rheumatology clinics serve people with chronic autoimmune conditions such as rheumatoid arthritis and lupus, diseases that themselves can destabilize employment, exhaust financial reserves, and erode support networks over time. A patient who was securely housed and employed at an initial visit may experience disease flares, medication side effects, or out-of-pocket costs that push them into material hardship months later. Conversely, a patient who screened positive for food insecurity at intake may connect with community resources, recover employment, or stabilize with family support, and no longer report that need at a subsequent visit. Both directions of change carry clinical meaning: new risks signal emerging barriers to adherence and control, while resolved risks indicate that prior interventions or circumstances have shifted.

The study’s methods reflect its origins in real-world clinical operations rather than a controlled trial. Data were collected as part of the clinics’ screening initiative and extracted from the electronic health record in de-identified form, an approach that received exempt institutional review board approval from Washington University. Because the analysis draws on routine care data from a single specialty population, its findings speak most directly to health systems operating similar screening programs, particularly in rheumatology and other specialty outpatient settings where longitudinal relationships between patients and clinics create natural opportunities for repeated assessment. The work was supported by an Arthritis Foundation 2022 DEI Research Award and a grant from the National Institute of Arthritis and Musculoskeletal Diseases, and it was presented at the American College of Rheumatology’s annual meeting and the American Medical Informatics Association Annual Symposium before publication.

What makes the findings resonate beyond rheumatology is the growing infrastructure built around social needs data. Health systems across the United States are investing in screening workflows, referral platforms, and documentation standards, often under pressure from payers and accreditation bodies. If social risk factors fluctuate meaningfully within patients, then several operational conclusions follow. First, the timing of screening matters: a single positive or negative screen captures a moment, not a trajectory. Second, rescreening intervals become a design decision with real consequences, since too-infrequent screening will miss incident needs while overly frequent screening may burden patients and staff. Third, longitudinal social risk data could enable more sophisticated analytics, allowing systems to distinguish persistent hardship, which may require intensive intervention, from transient need, which may respond to short-term support.

The study also touches on an ongoing conceptual debate in the field, crystallized in work by Alderwick and Gottlieb on the meanings and misunderstandings of social determinants terminology. Social risk, social need, and social determinants are related but distinct constructs, and health care systems often conflate them. Screening in a clinic measures reported risk at a point in care, not the full architecture of disadvantage shaping a life. The Washington University team’s longitudinal lens adds nuance to this lexicon by highlighting temporality: a risk factor is not merely present or absent but has a duration, an onset, and sometimes a resolution. Capturing those dynamics requires treating social data more like vital signs, measured repeatedly and interpreted in context, rather than like demographic attributes recorded once.

For patients with chronic inflammatory disease, the stakes are tangible. Social deprivation has been associated with higher disease activity in rheumatoid arthritis in systematic literature reviews, and arthritis-related disability is unevenly distributed across social risk strata in state-level analyses. A patient struggling to afford transportation may miss infusions; one facing food insecurity may have difficulty maintaining nutrition that supports overall health; one without stable housing may struggle with medication storage and adherence. When clinics detect these barriers early and repeatedly, they can connect patients with social resources, adjust treatment plans, or coordinate with community organizations. When they screen once and move on, the chart may silently misrepresent the patient’s circumstances for years.

The broader lesson for medicine is that social information in the electronic health record deserves the same attention to change over time as laboratory values or vital signs. The Washington University study, grounded in a real specialty clinic population and a functioning screening program, provides an evidence base for that shift in practice. As federal guidelines push wider adoption of social risk screening and health systems build the informatics infrastructure to act on the results, the finding that social risk factors change over time argues for dynamic, repeated assessment as the default. A chart that updates a patient’s social reality as faithfully as it updates their disease activity is a chart that can finally support whole-person care, and this study offers a concrete, clinic-tested step in that direction.

Subject of Research: Longitudinal changes in patient-reported social risk factors in a specialty outpatient rheumatology population

Article Title: Changes in Social Risk Factors Over Time in a Specialty Outpatient Population

Article References: Lew, D., Seidler, L., McQueen, A., Bartels, C. M., Garg, S., Eisen, S., & Kim, A. H. J. (2026). Changes in Social Risk Factors Over Time in a Specialty Outpatient Population. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10786-9

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10786-9

Keywords: social risk factors, social determinants of health, rheumatology, outpatient screening, rheumatoid arthritis, electronic health records, health disparities, population screening, patient-reported outcomes, health services research, food insecurity, longitudinal study