This study was a part of a parent project investigating ethnic/racial and socioeconomic differences in chronic low back pain (ERASED) (56, 57). The parent project examined racial differences in biopsychosocial factors that influence pain. The data in this study were collected between November 2017 and July 2022. Interested participants completed a telephone screening to determine eligibility, and two laboratory-based sessions spaced one week apart. Questionnaires assessing depressive and insomnia symptoms, and perceived stress were completed in week 1, and questionnaires assessing pain severity and interference were completed at weeks 1 and 2. For this study, we used the BPI-SF scores that were completed at week 2. Total compensation for the completion of both experimental sessions was $400. Specifically, participants received $150 per session, $60 for the completion of at-home sleep monitoring, and $10 for each follow-up telephone call. Inclusion criteria were as follows: (1) between the ages of 18–85; (2) able to read, write, and comprehend English, (3) self-identify as Black/African American or White/Caucasian, and (4) report low back pain for 3 consecutive months that was present for at least half of the days in the past 6 months, with the primary complaint being non-specific low back pain. Exclusion criteria were as follows: (1) low back pain attributed to other factors such as ankylosing spondylitis, infection, malignancy, or compression fracture, (2) surgical intervention or accident/trauma in the past year, (3) presence of any systemic rheumatic conditions, (4) uncontrolled hypertension, cardiovascular, or peripheral artery disease, (5) poorly controlled diabetes, (6) neurological disease, (7) psychiatric hospitalization within the past year, and (6) pregnancy. Written informed consent was given to the participant and signed. This study was approved by the local IRB.

Participants in the Birmingham, Alabama community were recruited via flyers posted at the UAB Pain Treatment clinic and surrounding areas. A total of 208 participants were recruited for this study. Full demographic details will be presented in Table 1.

All analyses were performed using SPPS v.28 (IBM, Corp.). According to prior theory, listwise deletion can be used when less than 15% of data are missing, and data appears to be missing at random (78). Given these circumstances, we used listwise deletion to conduct our analyses. Normality tests were also run to determine whether parametric or non-parametric analyses should be employed. Apart from the Insomnia Severity Index and the Pain Severity variable, none of our data was normally distributed. Consequently, we used robust analyses (i.e., Spearman's rho, mediation with bootstrapped sampling) to accommodate for potential normality violations. Sequential mediations using PROCESS macro (Model 6, 95% BCI, 5,000 bootstrapped samples) were conducted to evaluate the indirect effect of discrimination on pain outcomes, and vice versa. We chose age, gender, race (Black vs. White), education (some vs. no college), income level (below or above 30,000), and medication status (taking vs. not taking medications) as covariates. Prior literature states that disparities in pain vary based upon these sociodemographic characteristics, and behavioral patterns such as using prescription or over the counter medications for pain or related conditions (79–82). All categorical variables were coded as 0 and 1 for inclusion in analyses. Stress, depressive and insomnia symptoms all independently served as mediators.

In our sample, 62% (N = 129) of our participants identified as Black/African American, and 38% (N = 79) identified as White/Caucasian. The percentage of women in our sample was 55.3% (N = 115), and 44.7% (N = 93) of our participants identified as men. The average age for this sample was ∼45 years old, and the average income ranged from $30,00–$34,999. Moreover, 41.7% (N = 83) of our participants had lower incomes (e.g., below 30,000), and 58.3% (N = 116) had higher incomes (e.g., above 30,000). Additionally, 77.9% (N = 162) of our sample had some college experience, followed by 22.1% (N = 46) of those with no college experience. Almost half of our participants (N = 89) were currently taking non-steroidal anti-inflammatory drugs (NSAIDS), and 33.7% (N = 70) of our participants had high blood pressure. Of the 208 participants, over half had moderate pain severity 52.5% (N = 106), followed by 20.8% (N = 42) with severe pain, 17.8% (N = 36) with mild pain, 5.9% (N = 12) with very severe pain, 2.5% with very mild pain, and 1 person who marked “no pain” on the BPI-SF. Cutoff scores from the ISI also demonstrated that 33.7% (N = 70) of our participants had subthreshold insomnia symptoms, followed by 29.3% (N = 61) with moderate symptoms, and 10.1% (N = 21) with severe insomnia. Importantly, 26.9% of our participants did not meet cutoff scores for clinically significant insomnia. A vast majority (46.8%, N = 101) of our participants met the cutoff criteria for clinically significant depressive symptoms. Full sociodemographic and clinical characteristics are shown in Tables 1, 2, and responses to the everyday discrimination scale will be shown in Supplementary Figure S1.

As seen in Table 3, greater discriminatory experiences were associated with male gender (ρ = −.20, p = .004), greater perceived stress (ρ = .31, p < .001), symptoms of depression (ρ = .38, p < .001), and insomnia (ρ = .26, p < .001), higher pain severity (ρ = .23, p < .001), and greater pain interference (ρ = .25, p < .001). Greater perceived stress was associated with younger age (ρ = −.18, p = .009), greater symptoms of depression (ρ = .35, p < .001), and insomnia (ρ = .35, p < .001), higher pain severity (ρ = .19, p = .005), and greater pain interference (ρ = .28, p < .001). Symptoms of depression were associated with having a lower income (ρ = .20, p = .004), greater insomnia symptoms (ρ = .62, p < .001), higher pain severity (ρ = .43, p < .001), and greater pain interference (ρ = .57, p < .001). Lastly, insomnia symptoms were associated with higher pain severity (ρ = .45, p < .001) and interference (ρ = .57, p < .001). While most correlations were moderate (less than 0.5), the associations between sleep, mood, and pain outcomes were stronger (<0.6).

To examine the associations among psychosocial risk factors and self-reported pain severity, hierarchical linear regressions were conducted. Sociodemographic variables (age, race, gender, income level, some or no college experience, medications) were entered into Model 1, explaining 15% of the variance in pain severity (F = 5.40, p < .001). Specifically, identifying as Black (β = .23, p = .001), having a lower income (β = .16, p = .02), and no college experience (β = .15, p = .03) were associated with greater pain severity. When added to the model, psychosocial risk factors contributed an additional 24% of variance in self-reported pain severity (F = 11.65, p < .001). Specifically, identifying as Black (β = .23, p < .001), greater depressive symptoms (β = .18, p = .02) and greater insomnia symptoms (β = .29, p < .001) were associated with greater self-reported pain severity. While overall significant (F = 10.54, p < .001), Model 3 didn't contribute additional variance to self-reported pain severity (p = .95). All regression coefficients will be shown in Table 4.

Additional analyses were conducted to examine the associations between psychosocial risk factors and pain interference. Model 1 included demographic covariates and explained 14% of the variance in self-reported pain interference (F = 5.01, p < .001). Specifically, identifying as Black (β = .17, p = .02), having lower income (β = .16, p = .02), no college experience (β = .16, p = .02), and taking medications (β = .16, p = .02) were associated with greater pain interference. Model 2 explained an additional 36% of variance in self-reported pain interference (F = 18.15, p < .001), such that older age (β = .14, p = .01), identifying as Black (β = .18, p = .001), no college experience (β = .11, p = .03), and taking medications (β = .11, p = .04) were associated with greater pain interference. Moreover, greater depressive symptoms (β = .33, p < .001), and insomnia symptoms (β = .32, p < .001) were associated with greater pain interference. These findings remained significant in Model 3 (F = 16.42, p < .001). All regression coefficients will be shown in Table 5.

As seen in Figure 1, there was a significant indirect effect of discrimination on pain severity through psychosocial risk factors, with a point estimate of.015 (95% CI: .003 to .030). Specifically, experiences of discrimination were associated with greater perceived stress (β = .39, p < .001), greater perceived stress was associated with greater depressive symptoms (β = .22, p = .001), greater depressive symptoms were associated with greater insomnia symptoms (β = .61, p < .001), and greater insomnia symptoms were associated with greater pain severity (β = .29, p < .001). Full descriptions and coefficients for each individual pathway will be shown in Figure 1.

There was also a significant indirect effect of discriminatory experiences on pain interference via psychosocial risk factors, with a point estimate of .015 (95% CI: .004 to .031). Specifically, greater experiences of discrimination were associated with greater perceived stress (β = .34, p < .001), greater perceived stress were associated with greater depressive symptoms (β = .22, p < .001), greater depressive symptoms were associated with greater insomnia symptoms (β = .62, p < .001), and greater insomnia symptoms were associated with greater pain interference (β = .32, p < .001). Full descriptions and coefficients for each individual pathway will be shown in Figure 2.

We conducted exploratory analyses to assess the indirect effects of pain severity and interference on experiences of discrimination. There was a significant indirect effect of pain severity on discriminatory experiences via psychosocial risk factors with a point estimate of .014, 95% CI: .001 to .029). Specifically, greater pain severity was associated with greater insomnia symptoms (β = .52, p < .001), greater insomnia symptoms was associated with greater depressive symptoms (β = .54, p < .001), greater depressive symptoms was associated with greater perceived stress (β = .20, p = .02) and greater perceived stress was associated with greater discriminatory experiences (β = .23, p < .001). Full descriptions and coefficients for each individual pathway will be shown in Figure 3.

There was also a significant indirect effect of pain interference on daily discrimination via psychosocial risk factors (point estimate = .012, 95% CI: .000 to .029). Specifically, greater pain interference was associated with greater insomnia symptoms (β = .62, p < .001), greater insomnia symptoms were associated with greater depressive symptoms (β = .44, p < .001), greater depressive symptoms were associated with greater perceived stress (β = 21, p = .02), and greater perceived stress was associated with greater discriminatory experiences (β = .20, p = .004). Full descriptions and coefficients for each individual pathway will be shown in Figure 4.

Based on our research, we found significant associations between discrimination, psychosocial risk factors (e.g., sleep, stress and mood), and pain outcomes in people with cLBP. Additionally, the indirect effect of discrimination on pain severity and interference was significant. These findings align with prior research that suggests an association between discrimination and negative pain-related outcomes, ultimately affecting quality of life, co-morbidity burden, and accelerated mortality (16, 83, 84). Moreover, our regressions showed that certain demographic (e.g., Black race, lower education) and behavioral (e.g., taking medications, depression and insomnia symptoms) characteristics were associated with worse pain outcomes, which aligns with previous research (48, 53, 80, 81, 85). Lastly, mediation analyses showed that the associations between daily discrimination and pain outcomes could be partially explained by other psychosocial risk factors (stress, depression and insomnia)—a finding that hasn't been explored and sets the foundation for future intervention research and health initiatives. Taken together, this manuscript highlights that having cLBP introduces discrimination, which may contribute to negative physical and emotional aspects associated with worse pain outcomes (i.e., pain severity, mood, physical functioning). As seen above, discrimination impacts perceptions of stress, depression and insomnia, and these effects lead to greater perceptions of pain severity and interference (44, 84, 86, 87). In turn, experiencing chronic pain may present as a precursor for co-morbid symptoms like insomnia and depression, thus introducing a vulnerability to stigma and related discrimination (53, 83). In our study, these relationships persisted even after controlling for demographics (i.e., age, gender, race, income and education, medications). This suggests that in addition to demographic disparities in pain and mental health, there is an association between discrimination, pain, and quality of life that deserves a further look (12, 88, 89). Together, these findings demonstrate the robust and negative impact of discrimination and pain on multiple levels of health and functioning (84, 89–91). Overall, we provide novel evidence of potential therapeutic targets (e.g., stress-reduction, positive psychology, trauma informed care) for intervention, while also contributing to prior literature aiming to improve pain and physical functioning in people with cLBP, which we will elaborate on further (92, 93).

Although both mediations were significant, the indirect effects of discrimination on pain severity and interference were slightly stronger. This finding aligns with prior work suggesting that consistent experiences of ostracism heighten the severity of pain, as well as how pain interferes with daily life (i.e., pain perception, disability) (33, 94). Importantly, the associations between discrimination and pain outcomes were fully explained by psychosocial mediators, such as stress, depressive symptoms, and poor sleep. To elaborate, experiences of discrimination can often be deemed stressful, and theories such as the Social Pain theory, and the Transactional Model of Stress and Coping support our findings (95–97). Specifically, the social pain theory suggests that experiences of ostracism ignite some of the same negative emotions that are induced during physical threat or injury—thus contributing to a heightened pain perception and experience (96, 98). On the contrary, the transactional model of stress and coping suggests that stress negatively impacts health and coping behaviors when a person perceives that their individual resources and capacities are consistently outweighed by their daily demands (95). For people with cLBP, this experience can be common due to the taxing demands of having a debilitating health condition, amongst other intersectional identities and pressures (i.e., Black race, lower socioeconomic status, taking pain or related medications) (99). Taken together, the experience of discrimination has underlying health effects that warrants further attention for the development of stress-informed interventions and care in people with cLBP.

In addition to appraisals of stress, prior literature suggests that discriminatory experiences can introduce feelings of helplessness, and rumination—both symptoms of depression that can be detrimental (44–46, 100). Importantly, nearly half of our sample met diagnostic criteria for clinically significant depression. Although not measured directly, prior work has shown that experiences of stigma and discrimination can be internalized, leaving people to consistently cope and appraise the situation, and throwing off daily patterns (i.e., rumination, poor sleep) (44, 84, 101, 102). Consequently, symptoms of insomnia can arise, and this is heightened in people with pain conditions for a variety of reasons (e.g., shared neural circuitry, coping patterns, demographics) (47, 48, 89, 103). Future research could involve using the following strategies to address these areas: (1) cognitive or brief behavioral therapy for insomnia (CBT-I, BBT-I) to improve sleep and mood, (2) resilience enhancing interventions (positive psychology, exercise) to improve mental health among minority groups who experience discrimination, (3) talk-therapy and mindfulness-based stress reduction (MBSR) interventions for symptoms of stress and depression, and (4) yoga and spirituality as holistic measures for targeting all factors (93, 104–106). More specifically, a recent study in people with HIV and pain found that the use of Brief Behavioral Therapy for Insomnia (BBT-i) decreased insomnia symptoms and pain interference from pre- to post-intervention (104). This trend has also been replicated in football players, with decreases in both pain severity and interference (107). Positive psychology and music analgesia have also been implicated as successful interventions to reduce stress, pain interference and severity (92, 93, 108). Lastly, mindful meditation and physical activity interventions have been shown to decrease both depression and pain (109, 110). While interventions addressing stigma and discrimination are scarce, future work could involve bias trainings and policy changes that build health equity and equal access to resources (i.e., health education, science communication, pop-up health screenings, food and transportation assistance) with hopes to lessen the burden of systemic and interpersonal marginalization (12).

A series of regressions showed that identifying as Black, having a lower education, and taking daily medications were also associated with greater pain-related interference. Moreover, identifying as Black was also associated with a greater severity of pain. While previously noted, these findings shed light upon the economic, physical, and mental risk that minoritized groups with pain face daily (99). For example, prior work states that living with one or more stigmatizing identities (e.g., intersectionality) can influence the trajectory of your health—especially for people with chronic pain (12, 100). Moreover, a vast amount of literature has shown that those with lower socioeconomic statuses have an additive burden with pain management (access to medication and resources, health insurance) that often exacerbates their pain experience (82, 111, 112). Coupled with the daily pressures of having chronic pain, these additive layers of complexity can introduce a range of daily complications, and could be intervened with by introducing more resources, health education interventions, and mutual aid initiatives (93, 113). For example, if a person with lower income is taking pain medications, a health education session could determine the difference between adaptive or maladaptive coping patterns. In addition to this, mutual aid initiatives could bridge the gap between pain-related disability and pain management in those with less tools and disadvantaged backgrounds. In summary, future research will extend beyond psychological risk factors to enhance pain management and treatment among those from disadvantaged backgrounds.

In addition to novel advances in research, our study has implications that can be applied clinically. For example, knowledge of the co-occurrence among stress, depression and insomnia symptoms can inform more holistic approaches to clinical care. Therefore, instead of focusing on the physical representation of the patient, physicians can move towards individualized, person-centered care (114). Non-pharmacological techniques such as breathing exercises, journaling, and physical activity have all shown to decrease symptoms of stress, pain and depression (115–118). Moreover, spending additional time assessing the patient's daily routines and habits could really inform their treatment recommendations. For example, if a patient with cLBP presents with severe pain, doctors can begin to assess the upstream or underlying reasons (e.g., asking how they sleep on a regular basis, what are their daily stress levels, how is their mood?) instead of providing prescription medications alone. Treating the underlying root of the condition could help lessen the burden and severity of pain, while also giving patients tools that could be translated into their daily lives (e.g., breathing activities for stress reduction, improving sleep hygiene). Most importantly, experiences of healthcare and interpersonal discrimination could be addressed by emphasizing patient advocacy and resilience (10, 119). While discriminatory experiences are hard to control, leveraging individual resources, building barriers of resilience, and community (i.e., social support) can circumvent the internalized thoughts that often affect the perception and experience of pain in minorities.

This study possesses several limitations that warrant discussion. First, this study was cross-sectional and hypothesis generating in nature. Specifically, our sequential mediation models were designed to test the possibility of an indirect association between daily discrimination and pain outcomes through psychosocial risk factors. Because sessions were only separated by one week, we are limited in determining directionality and inference. Thus, results should be interpreted with caution due to the limitations of time constraints on directionality. To circumvent this, we conducted backwards mediations to assess indirect effects from both discrimination and pain outcomes. Future research will incorporate a longitudinal design to assess the true directionality (i.e., causal models) and strength of associations. Secondly, our study used single-site recruitment and involved a high level of compensation (e.g., $400) that could introduce possible selection bias. Future work could incorporate these measures on a larger scale, using multi-site recruitment strategies to further generalize our findings. Third, we examined cumulative experiences of discrimination; however, discriminatory experiences are often shaped by historical and cultural contexts, due to power differentials and oppression affecting some groups more than others; thus, future studies should include larger and diverse samples to examine how these relationships vary based on sub-groups. We did not do group comparisons based on sociodemographic characteristics, and we also didn't have measures of stress reactivity despite the mention in our introduction (i.e., HPA axis dysregulation). However, like previous studies, we wanted to first examine the associations between stressful experiences, pain, and mental health that haven't been shown in people with cLBP. Future research will incorporate objective biomarkers of stress, and measures of differences based on demographic characteristics to get a fuller picture of discrimination and pain outcomes. Lastly, the possibility of pain outcomes being explained by other confounders (i.e., BMI, pain duration) wasn't included in this study, and our sample size was relatively small. Future work will incorporate a more extensive approach to understanding pain and pain management, as well as using a larger sample size to promote generalizability to the larger cLBP population.

As the first study testing these sequential models in a cLBP sample, our results highlight potential avenues that explain, in part, the associations between daily discrimination and pain outcomes. These pathways highlight the need to evaluate how discrimination influences additional psychological variables (i.e., stress, depression) and sleep functioning, which often co-exist with and exacerbate the pain experience. Our findings also provide evidence for vulnerability to discriminatory experiences in people with cLBP, and how these experiences may be internalized through depressive symptoms and stress. Moreover, our findings represent the importance of recognizing sociodemographic characteristics in clinical interventions and study design (e.g., Black race, lower SES, taking medications). This is the first study to elaborate on these indirect effects, and future studies should target these mechanistic pathways for intervention and treatment strategies (e.g., longitudinal design, larger sample, MBSR, BBTi, mutual aid and bias trainings), as well as narrow down on minority groups (e.g., racial minorities, LGBTQ+, older adults, people who use substances, women).