Unmet health-related social needs are associated with pain outcomes among midlife and older adults

1 Introduction

Chronic pain prevalence among U.S. adults 65 years and older has risen from 30.5% in 2021 to 36.0% in 2023, a larger increase than in any other age group (1). In this population, chronic pain [pain lasting for over 3 months (2)] is associated with impaired mobility and physical functioning (3), anxiety and depression (4), cognitive impairment (5), frailty (6), loneliness (7), and mortality (8, 9). Despite this substantial and increasing hardship, chronic pain remains widely underassessed and undertreated (10, 11), with the heaviest burden often falling on those most affected by structural inequities, including racism and poverty (12–15). These avoidable differences in pain prevalence, management, and treatment are dynamic and stem from multiple intersecting historical and present-day inequalities in power, access to opportunities and resources, wealth, and healthcare (14).

A growing body of work has called for attention to the intersecting and dynamic social factors that shape pain prevalence, experience, and management over the life course (16, 17). Research on the social predictors of chronic pain has largely focused on distal indicators, including education, socioeconomic status, and neighborhood environments. For example, a socioeconomic gradient has been identified whereby lower education predicts more frequent pain and greater pain severity and interference (18). Other studies have reported that living in areas with fewer socioeconomic opportunities is associated with higher mean pain interference (19, 20) and greater frequency of chronic pain among adults (20, 21). Persistent exposure to socioeconomic inequity across the life course, such as working in manual labor, further elevates the odds of pain in older age (22).

Comparatively less is known about more proximal, individual-level social needs, which are the daily non-medical requirements for living [e.g., food, housing, utilities, transportation, and social support (23)] that arise from upstream social determinants of health. These needs represent one way in which broader structural inequities are translated into individual-level health consequences such as pain (17). For example, a nationally representative cross-sectional study found that food insecurity was associated with higher odds of chronic pain after accounting for socioeconomic factors (24). Research has also demonstrated a link between social isolation and increased pain interference with activities of daily living (25). One potential pathway to explain these relationships is chronic stress: prolonged exposure to psychosocial stressors, such as those caused by unmet social needs, can increase vulnerability to chronic pain (26, 27), consistent with the weathering hypothesis (28). However, to our knowledge, studies have not examined how a wider array of cumulative and individual unmet social needs relate to pain outcomes. Identifying these associations can inform modifiable, non-medical targets for more equitable pain interventions and help providers and policymakers prioritize non-medical contributors to pain in this population.

The present study examines the associations between unmet health-related social needs and multiple pain outcomes among midlife and older adults with chronic pain living in an underserved urban community. We assessed the prevalence of unmet social needs and evaluated whether the cumulative number of unmet social needs and individual challenges with food, housing, utilities, transportation, or social support are associated with pain outcomes.

2 Method

2.1 Study sample

We pooled pre-intervention data from two randomized controlled trials of community health worker-led behavioral interventions for adults aged 50 years and older living in or near Detroit, Michigan. The first (Seniors using Technology to Engage in Pain Self-management, or STEPS) focused on chronic pain self-management (n = 193) (29). The second (Re-engaging in Self-care, Enjoying Today, or RESET), from which we drew a subset of participants reporting chronic pain at screening (n = 245), tested a positive psychology intervention to enhance psychosocial functioning. Participants were included in the present analysis for STEPS and the RESET subsample if they completed the social needs screener, reported experiencing pain in their muscles or joints on most days for at least 3 months, rated their pain level over the last week as 4 out of 10 or higher, and had at least one day in the last month where pain made it hard to perform their usual daily activities.

STEPS and RESET participants were recruited primarily through health fairs, information sessions at senior centers and senior housing facilities, social media, and letters through a partnering healthcare system (for STEPS only). Participants in both studies provided verbal informed consent over the phone, with the option to receive the consent document by mail or email to review ahead of time, or by mail after verbal consent. Research staff read the consent form aloud, paused after each section for questions, and audio-recorded this process with participants' permission. Participants earned up to $65 for completing all surveys, and control group participants were offered a version of the STEPS or RESET intervention after their final survey.

Demographic and pain data were collected during screening or baseline assessments. Social needs screener questions were administered in-person or via telephone during participants' first meeting with a community health worker. This meeting took place after randomization and before starting the intervention for the STEPS and RESET intervention group members. RESET control group members were administered the screener in a post-randomization “wellness check” phone call. STEPS control group members did not receive the social needs screener unless they opted to participate in a one-time pain management workshop following final data collection; data from this group are not used in the present analysis. Participants were provided with resource referrals to help address unmet needs. Follow-up data were collected for both studies (at 2 and 12 months for STEPS and 2 and 8 months for RESET) but are not used in the present analysis.

2.2 Predictors

Social needs predictors were drawn from standard screeners related to the social determinants of health (30), and were restricted to domains for which the community health worker could readily provide resources. Items were measured as no = 0 and yes = 1, and were as follows: (1) “Do you need support with other health or medical expenses like vision, dental, or medical equipment?”; (2) “Are you experiencing any challenges related to getting food?”; (3) “Do you need help with housing or housing concerns?”; (4) “Do you have a hard time paying your utility bills?”; (5) “Do you have access to the internet/Wi-Fi where you live?” (reverse-coded); (6) “Do you need help finding or paying for care for yourself or a loved one?”; (7) “Do you have trouble with transportation?”; and (8) “Do you have the social support that you need?” (reverse-coded). Each item was retained as a binary indicator and a cumulative unmet social needs score was also created by summing the eight items (range 0–8). Given the skewed distribution of cumulative unmet social needs, with fewer participants reporting three or more needs, and consistent with other studies (31, 32), cumulative unmet needs was collapsed into four categories: 0, 1, 2, and ≥3 unmet needs.

2.3 Outcomes

Three pain outcomes were examined. Pain intensity was assessed as average pain in the past 7 days on an 11-point Numerical Rating Scale (0 = no pain, 10 = worst imaginable pain) (33). The standardized Patient-Reported Outcomes Measurement Information System (PROMIS) Pain Interference 4-item short form assessed the extent to which pain interfered with daily activities in the past 7 days, measured on a 5-point Likert scale (1 = not at all, 5 = very much) (34). Items were summed and converted into T-scores using conversion tables (35); T-scores have a population mean of 50 and standard deviation of 10. The 2-item version of the Pain Self-Efficacy Questionnaire (PSEQ) was scored by summing responses to the items: “I can do some form of work despite pain” and “I can live a normal lifestyle despite the pain” (7-point Likert scale: 0 = not at all confident, 6 = completely confident) (36). The PSEQ-2 was collected in the STEPS study only.

2.4 Covariates

Covariates included: age (years), gender, race (Black or African American, White, Other), education [high school graduate/General Educational Development (GED) or less, some postsecondary education (including vocational/technical, some college, or associate's degree), and bachelor's degree or higher] and number of self-reported chronic conditions (whether a doctor or other healthcare professional had ever diagnosed arthritis, low back pain, high blood pressure, diabetes, heart attack, angina, high cholesterol, depression, cancer, stroke, asthma, chronic obstructive pulmonary disease, migraine, heartburn, or osteoporosis). One participant self-identifying as transgender male (Other) was coded as male in regression analyses.

2.5 Analysis

Sample characteristics were summarized using means and standard deviations for continuous variables and counts and percentages for categorical variables. Bivariate tests assessed the relationships between cumulative unmet social needs or individual needs and pain outcomes (pain intensity, pain interference, pain self-efficacy). Analysis of variance (ANOVA) was used to compare mean scores of pain outcomes across the four levels of cumulative unmet social needs (0, 1, 2, and ≥3), and Tukey's honestly significant difference (HSD) was applied to identify significant pairwise differences. Individual unmet needs of interest (food, housing, utilities, transportation, and social support) were each treated as a binary indicator (no, yes), and mean pain outcome scores were compared between groups using Welch's t-tests.

Separate linear regression models examined the associations of cumulative unmet social needs and individual social needs (food, housing, utilities, transportation, and insufficient social support) with each pain outcome. All regression models adjusted for age, gender, race, education, and number of chronic conditions. Analyses included participants with available data for each outcome (pain intensity: n = 436; pain interference: n = 437; pain self-efficacy: n = 191). Missing values for categorical covariates were retained in regression models by including a Missing category for race (n = 13), education (n = 7), and gender (n = 2), such that adjusted and unadjusted models used identical samples. Multicollinearity was assessed using generalized variance inflation factor and all adjusted values were under 1.2, suggesting no problematic multicollinearity. Analyses were conducted in R version 4.5.3 with p < .05 considered statistically significant.

3 Results

Among 438 participants, the mean age was 67.5 years (SD = 8.1). Most (87%, n = 380) were female, 77% identified as Black or African American (n = 338), and over half (54%, n = 237) had one or more unmet social needs (Table 1). The unmet needs most frequently reported were housing (20%, n = 89), transportation difficulties and health or medical expense needs (both 18%, n = 81), Difficulty paying utility bills (17%, n = 74), and food access challenges (16%, n = 70). Fewer participants reported insufficient social support (11%, n = 46), needing assistance with childcare or eldercare (6%, n = 25), or internet access needs (5%, n = 24).

Table 1. Sample characteristics (N = 438).

Includes participants with three (n = 41), four (n = 22), five (n = 9) or six needs (n = 3).

Higher scores indicate worse pain interference in daily life. Items were summed and converted into T-scores.

Higher scores indicate increased self-efficacy in functioning in the presence of pain; data only available for STEPS dataset.

3.1 Bivariate associations between unmet social needs and pain outcomes

Bivariate tests examined associations between pain outcomes and cumulative and individual unmet social needs (Table 2). Cumulative unmet social needs were significantly associated with pain intensity (p = .002), pain interference (p < .001), and pain self-efficacy (p = .013). Tukey's HSD post-hoc comparisons indicated that participants with three or more unmet social needs reported significantly higher mean pain intensity than those without unmet needs (M = 6.61 vs. 5.49, p = .001) or one unmet need (M = 6.61 vs. 5.71, p = .033); the comparison with those reporting two unmet needs did not reach statistical significance for any outcome. A similar pattern emerged for pain interference post-hoc tests: participants reporting three or more unmet needs had worse interference than those with none (M = 63.26 vs. 58.65, p < .001) or one unmet need (M = 63.26 vs. 59.66, p = .015). Lastly, participants reporting three or more unmet social needs had lower pain self-efficacy than those with none (M = 6.15 vs. 8.69, p = .009).

Table 2. Bivariate associations between social need domains and pain outcomes.

Dataset contains participants from the STEPS study only.

ANOVA; Tukey's post-hoc test results are not reported in the table. All other results describe t-tests.

Regarding individual social need domains, transportation difficulties and insufficient social support were associated with higher pain intensity and interference, and lower self-efficacy (all p < .05). Housing needs and food access challenges were significantly associated with higher pain intensity and interference but were not associated with pain self-efficacy. Difficulty paying utility bills were not significantly associated with any pain outcome (all p > .05).

3.2 Multivariable regressions examining unmet social needs and pain outcomes

In regression models adjusted for age, gender, race, education, and number of chronic conditions (Table 3), participants reporting three or more unmet social needs had significantly higher pain intensity (B = 0.75, p = .010, 95% CI: 0.18, 1.32), pain interference (B = 3.24, p = .002, 95% CI: 1.24, 5.23) and lower pain self-efficacy (B = −2.04, p = .006, 95% CI: −3.50, −0.59) compared to those without unmet social needs.

Table 3. Multivariable regression results for unmet social needs and pain intensity, interference, and self-efficacy. Models adjusted for age, gender, race, education, and number of chronic conditions.

Only available in STEPS data.

Reference: None reported.

Reference: Sufficient social support.

In adjusted models examining individual needs, participants with food access challenges or insufficient social support reported higher pain intensity than those without these challenges (food access: B = 1.08, p < .001, 95% CI: 0.54, 1.61; social support: B = 0.74, p = .024, 95% CI: 0.10, 1.38). Those with food access challenges, transportation difficulties, or insufficient social support reported greater pain interference than those without (food access: B = 2.53, p = .009, 95% CI: 0.65, 4.42; transportation: B = 2.06, p = .026, 95% CI: 0.25, 3.86; social support: B = 2.44, p = .035, 95% CI: 0.17, 4.71). Participants reporting food access challenges or transportation difficulties also had lower pain self-efficacy compared to those without these challenges (food access: B = −1.54, p = .048, 95% CI: −3.07, −0.01; transportation difficulties: B = −1.82, p = .014, 95% CI: −3.27, −0.38). There were no significant relationships between housing or utility bill payment difficulties and any pain outcome, transportation difficulties and pain intensity, or insufficient social support and pain self-efficacy (p > .05). Unadjusted model results are included in Supplementary Material 1.

4 Discussion

This study examined associations between unmet social needs and pain outcomes among midlife and older adults with chronic pain in an underserved urban community. More than half the sample reported at least one unmet need; most commonly housing, health or medical expenses, and/or transportation. In adjusted regression models, only the highest category of cumulative unmet social needs (three or more versus none) was significantly associated with worse pain outcomes. This pattern may reflect co-occurring challenges that deplete the time and resources needed for pain self-management (37). Studies have reported that lower income is associated with worse pain self-efficacy scores (38) and there is an association between a greater number of unmet social needs and lower health confidence (39). More unmet social needs have been associated with higher perceived stress (40), which may in turn exacerbate pain. Together, these co-occurring challenges reflect a high level of disadvantage that may act as a fundamental cause of poor health (41, 42).

This stress pathway is exemplified in the weathering hypothesis, which proposes that chronic psychosocial stressors, including structural racism and economic hardship, contribute to physiological stress dysregulation and accelerated aging (28, 43); this dysregulation has been linked to increased risk of chronic pain (26, 27). This framework may be relevant to our sample: participants lived in or near Detroit, a city that has experienced structural inequities, including residential segregation, redlining, and disinvestment (28, 44). Most participants in this sample identified as Black or African American, a population disproportionately impacted by such policies and practices. However, experiences of stress are heterogeneous, shaped by intersecting social identities, life course exposures, and access to protective social resources and community supports (28).

Among individual social need domains, transportation challenges were associated with pain interference and self-efficacy. Pain self-efficacy is an important, modifiable target for behavioral interventions to improve pain management (45). This novel finding suggests that unaddressed transportation difficulties may undermine pain interventions. Primary care patients experiencing depression and pain have reported that limited transportation and financial resources, as well as inadequate social support, are barriers to pain self-management (37). These barriers are particularly salient for older adults, who may have greater healthcare needs, restricted mobility, or may no longer drive (46, 47). Transportation also likely intersects with other domains, including food access (48) and social support, which could exacerbate pain disparities.

Food access challenges showed the largest magnitude of association with pain intensity across individual social needs and were also associated with pain interference and self-efficacy, although the self-efficacy association should be interpreted with caution (p = .048). These results corroborate nationally representative evidence that food insecurity is associated with increased odds of pain among older adults (24). Mechanistically, food access challenges may worsen bodily function through obesity, insufficient nutrition, and inflammation (24, 49). The stress of insufficient food access or any unmet social need may also exacerbate pain, consistent with the stress pathways described previously. Insufficient social support was associated with pain intensity and interference, consistent with research findings that social support is directly linked to decreased pain intensity (50). Social support is a well-established protective factor against pain, operating through improved coping (51) and mental health, as well as reduced physiological stress responses to pain (50, 52).

In contrast, challenges related to housing and utility bills were not associated with any pain outcome in adjusted models. This finding aligns with a Midwestern study on emergency department patients that reported no significant associations between housing instability and pain intensity, interference, and self-efficacy (53). Alternatively, our screening item for housing needs may have lacked specificity regarding the type or severity of housing challenges. Future research should use more granular measures of housing circumstances to better assess these relationships.

This study addresses a gap in the literature by examining multiple social needs in relation to pain outcomes among midlife and older adults. Prior research has largely examined single social factors and pain in isolation, limiting our understanding of how multiple needs relate to pain. Unmet social needs arise from intersecting inequities, including structural racism, that can disproportionately shape the conditions and opportunities for Black, Indigenous, and other people of color (17) as well as economic stratification in society. Therefore, preventing and addressing chronic pain disparities requires multi-level interventions, including structural reform that honors community strengths, while addressing the root causes of inequitable resource distribution and discrimination in pain treatment. Understanding differences in the experiences and treatment of these adults can inform provider training on whole-person care, non-discriminatory clinical practice, and the development of culturally affirmed pain management interventions (54, 55).

Further, our findings have important implications for clinical practice. In this sample of adults living in a city shaped by structural inequities, more than half of participants with chronic pain experienced at least one unmet social need. While pain care seldom focuses on the social aspects of pain (17), these results underscore the need for greater integration of social needs screening and interventions into pain care. This may occur by embedding community health workers and/or social workers into interdisciplinary pain care models for improved pain management (56–58).

Beyond clinical care, these findings underscore the importance of programs and policies that prevent and address unmet social needs among midlife and older adults. While such programs, including the Supplemental Nutrition Assistance Program and food prescriptions, have been evaluated for health and/or economic outcomes (e.g., 59–61) their impact on pain is not well-established. Future evaluation of such programs should consider including multiple dimensions of pain experience as outcomes.

Study findings should be interpreted alongside limitations. This cross-sectional design cannot establish the directionality between social needs and pain outcomes; nor were we able to determine whether unmet needs directly affect pain or are indicators of another dimension of socioeconomic position that affects pain. For example, chronic pain and disability may contribute to changes in employment and financial burden (62). Additionally, the binary response format for social need questions may obscure heterogeneity in the severity of social needs. For example, participants indicating they need assistance with housing may be describing houselessness or home repairs. This variation may influence the association between social needs and pain outcomes. Similarly, participants reporting three or more unmet needs were collapsed into a single cumulative unmet need category, which obscures variation in needs within that group. The findings may also not be generalizable to other settings or age groups. Future studies should examine these associations using nationally representative samples with longitudinal designs.

Overall, the findings indicate that cumulative unmet social needs, food access challenges, transportation difficulties, and insufficient social support were associated with multiple pain outcomes. Future research should assess whether interventions targeting unmet social needs improve pain outcomes longitudinally, with particular attention to midlife and older adults with multiple co-occurring unmet social needs.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. The authors used generative AI (Claude Sonnet 4.6) to identify suggestions for improving grammar and sentence clarity, with a focus on the Methods section. After using this tool, the authors reviewed and edited the content and take full responsibility for the content of the publication.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpain.2026.1885908/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Tamara A. Baker, University of North Carolina at Chapel Hill, United States

Reviewed by: Janiece Taylor, Johns Hopkins University, United States

Catherine Clair, Johns Hopkins University, United States