Patients 18 years of age and older presenting for care at the University of Michigan Back & Pain Center between June 2018 and March 2020 were included in the present analyses. The Back & Pain Center is a tertiary-care, outpatient clinic servicing chronic pain patients across a range of conditions. All new patients are provided with a packet of questionnaires to complete prior to their first visit to be used for clinical care and research purposes (29). University of Michigan Medical School Institutional Review Board (HUM00041820) approval was obtained. A waiver of informed consent was obtained due to the extensive use of the survey in a clinical setting. Remuneration was not provided. Prior studies have utilized the patient data obtained through the Back & Pain Center; however, our specific sample and research question are unique compared to these previous studies (4, 30).

Patients were excluded if they did not provide information regarding abuse history. Additionally, if a participant indicated “no” to one or two forms of abuse, but left any missing, they were treated as missing in the final analyses. Sex, marital status, and childhood abuse were dummy coded, such that male sex, being nonmarried, and no childhood abuse were reference categories. Descriptive statistics were obtained for the variables of interest. Differences in the variables of interest by history of childhood abuse were assessed using t-tests and contingency analyses with associated χ² tests. Bivariate correlations were also assessed. Variance inflation factor did not exceed 3.0 for any variable; thus, multicollinearity was not a concern for the path model.

Path analysis was used to assess whether perceived emotional and instrumental support mediated the association between childhood abuse and widespread pain, pain severity, and negative affect. Mediators were allowed to covary as were the outcome variables. Childhood abuse was allowed to associate with outcomes conditional on the mediators. Thus, the model evaluated the presence of partial mediation using a saturated model and model fit statistics were not obtained. Sex, age, and marital status were included as covariates for each mediator and outcome. Full information maximum likelihood method of estimation and robust standard errors were used to account for missing values and non-normal data. Squared multiple correlation coefficients (R²) were used to determine the variance explained in each dependent variable.

A bootstrap procedure with 10,000 replicates was used to create bias-corrected and accelerated (BCa) 95% confidence intervals (95% CI) for the indirect effects. Complete case analysis was used to evaluate the significance of the indirect effects. Confidence intervals not including zero were considered significant. Total indirect effect coefficients were divided by total effect coefficients for each outcome to determine how much of the total effect was accounted for by the indirect effects through perceived emotional and instrumental support. Additionally, specific indirect path coefficients were divided by the total indirect effect coefficients to determine the relative importance of each pathway. The sum of the absolute values was used for the total indirect effect. Analyses were run using Stata version 15 (College Station, TX).

In the present study, patients living with chronic pain who have a history of childhood abuse reported lower perceptions of social support. These findings are in line with previous research. Childhood abuse may mold negative perceptions of others and impact how social information is interpreted (9, 11, 35–37), which may be reflected in less positive perceptions of and less satisfaction with social support (10–12). Childhood abuse may also fuel unhealthy relationship interactions marked by more negativity and less positive behavior that, in turn, results in less support being provided by others (38–40). How accurate these perceptions are and how well they fit the reality of support provision may not be particularly important; previous research suggests that perceptions of support are more important for health-related outcomes than actual support received (41). Additionally, perceptions of social support may be primarily dependent upon observer characteristics (41), including schema and expectations driven by prior experiences like abuse. Similar findings have been reported for the impact of childhood abuse on marital satisfaction in adulthood, which suggest that neither positive nor negative partner characteristics moderated the association (37).

Our findings also suggest that childhood abuse was associated with more widespread pain, higher pain severity, and higher anxiety and depressive symptoms. Indeed, previous research and theory suggests that childhood abuse is associated with poor physical and psychological health outcomes in adulthood (2–5, 42). This extends to a more complex clinical pain phenotype, including more widespread pain, higher pain severity, and higher anxiety and depressive symptoms (4).

Although the association between childhood abuse and psychological and physical health outcomes has been well-established in the literature, researchers have highlighted the necessity of understanding the mechanisms of this association. As considered in the current study, perceptions of social relationships may be important to consider. Indeed, social factors (e.g., interpersonal style and social cognition) have been proposed as potential pathways conveying the impact of childhood abuse on health outcomes (8). Yet, other pathways are important as well, including the influence of childhood abuse on health risk and promotive behavior and posttraumatic stress symptoms (8). Thus, it is unsurprising that significant, direct associations were found and maintained even in the present mediation model.

Our findings, however, suggest that the association between childhood abuse and sensory and affective pain-related characteristics, including widespread pain, anxiety symptoms, and depressive symptoms, could be partially explained by perceptions of social support. Similar to our findings, previous research has found that perceptions of social support and perceived relationship quality were important mechanisms linking childhood abuse with depressive symptoms (27), allostatic load in middle adulthood (12), as well as worse health perceptions, worse mental health, and more somatic complaints (13). Furthermore, recent research found that marital quality, composed of perceptions of support, strain, and disagreement, conveyed the association between childhood abuse and negative affect among older adults (43). Thus, perceptions of social support are a potentially modifiable mechanism that should be further investigated. Of note, although variance in widespread pain was better accounted for by the model than variance in pain severity, both outcomes exhibited low variance explained suggesting that other predictors are important to consider for sensory pain-related characteristics. Instead, the present model appears to better predict affective pain-related characteristics.

The association between childhood abuse and widespread pain is particularly interesting. Compared to pain severity, widespread pain may be more characteristic of nociplastic pain, in which central nervous system factors are believed to play a part in the spread or amplification of pain (25, 44). Stressors have been implicated as precipitating factors of nociplastic pain (25). Thus, low satisfaction with social support, negative social perceptions, and accompanying relationship distress may be sources of chronic stress that contribute to poor physical health including nociplastic pain (8). Alternatively, survivors of childhood abuse may not benefit from favorable perceptions of social support, which are shown to reduce perceptions of stress, improve coping strategies, and reduce the physiological stress response to pain (17).

The lack of findings with pain severity may also be associated with the time frame in which it is evaluated. The measure of widespread pain is indicative of chronicity, whereas the measure of pain severity considered the preceding week. Our model likely reflects long processes in which childhood abuse negatively impacts social relationships across the lifespan, and dissatisfying social perceptions and relationships influence perceptions of stress as well as present direct sources of distress that compound over time (45). On the other hand, pain severity may be a brief snapshot of sensory experiences. Future research should consider dynamic trajectories of pain severity and whether these instead are associated with social perceptions.

Additionally, instrumental support was associated with higher pain severity in the model, although this direct association was nonsignificant. This likely contributed to the significant indirect effect linking childhood abuse to lower pain severity through instrumental support. The model findings suggest that instrumental support may function as a solicitous form of social support that promotes and reinforces pain behavior like expressions of severity (46). Thus, the availability of instrumental support may encourage individuals to acknowledge the severity of their pain in order to receive much needed assistance with daily tasks.

The findings also suggest that, overall, perceived emotional support better predicts sensory and affective pain-related characteristics compared to perceived instrumental support. The benefit of each form of support may depend on what the individual requires in specific situations (14). Emotional support may be more beneficial for chronic health conditions (47), whereas instrumental support may be more valuable for specific needs.

The present study had many strengths. We evaluated these associations in a large sample of patients with pain using measures of both emotional and instrumental social support, and also broadly conceptualized pain-related characteristics as encompassing both sensory and affective components. Numerous limitations should also be acknowledged. We tested a mediation model which imposed a specific directional effect and our results reflect our statistical approach. However, the data were cross-sectional which limits our ability to make any causal claims. There are likely bidirectional relationships between social support perceptions and pain-related characteristics. Our goal, however, was not to propose a specific direction of effects to the exclusion of alternative models, but to evaluate a model that may have implications for chronic pain interventions. Our measure of abuse was limited due to the data being collected in the context of clinical care and the necessity to keep measures brief to ease participant burden. Type and timing of abuse should be investigated in future research. Additionally, many victims of abuse do not acknowledge or label the experience as abuse. Measures that provide behavioral descriptions of what the abuse would entail would likely result in higher prevalence rates. The participants also had various pain diagnoses and mixed etiology. This may contribute to important, yet indistinguishable, heterogeneity in the current sample. Future research should consider these factors. Additionally, due to the secondary nature of the present analyses, we did not have measures related to other forms of social integration or negative social perceptions such as hostility and perceived rejection. Future research should consider the role of social integration, which may be more likely to exhibit direct effects on health outcomes (14), as well as the impact of negative social perceptions on health outcomes, which may be more strongly associated with poor health outcomes compared to conceptually positive experiences of emotional and instrumental support (48).

Despite these limitations, understanding the impact of childhood abuse on perceptions of social support and, in turn, social support on pain-related characteristics is important to understand as chronic pain management techniques and interventions that promote the value of social support (e.g., support groups, engaging spouses or loved ones in care) are often recommended. Yet, patients with a history of childhood abuse victimization may perceive support differently which would have important implications for these pain management techniques and interventions. Thus, it is critical for healthcare providers to consider the interpersonal history of patients. Previous studies have demonstrated the importance of screening all chronic pain patients for a history of abuse (49, 50). Doing so can help guide intervention administration and modifications. Indeed, patients with a history of childhood abuse may especially benefit from interventions that improve the sense of safety and care in their relationships.

The findings also extend to the patient-provider relationship. Patients' trust in the recommendations of their provider is predicated on trust in the relationship they have with that provider. There is evidence that childhood abuse results in patients having less trust in their providers, as well as providers rating these patients as more difficult (50–53). This bidirectional challenge can undermine the very safety and trust needed to be able to bring up and discuss the role that childhood abuse could have or be having on the overall management of their pain, thus making long term success and improvement even harder to achieve. Further research into the impact of trauma-informed care and abuse screening in the setting of chronic pain management is needed to explore ways to make the seeking of care safe and supportive.

Survivors of childhood abuse exhibit a worse clinical pain phenotype, highlighting the critical need to understand the various mechanisms driving this association. Researchers and clinicians should consider the role of social perceptions within this complex relationship, including their impact on the patient-provider relationship. Interventions that improve patients' sense that others love and care for them may contribute to symptom improvement.