https://orcid.org/0000-0002-1528-1795
Abstract
The objective of this study was to examine the role of pain catastrophizing and pain self-efficacy as possible mediators of race-based differences in pain intensity and to evaluate the possible moderating role of race on the relationship between pain catastrophizing and pain self-efficacy with pain outcomes among persons with chronic spinal pain receiving physical therapy.
This study was a secondary analysis of a cluster-randomized trial. Participants were persons with chronic spinal pain in outpatient physical therapy clinics who consented to complete assessments at baseline and after 2 weeks and 12 weeks. Assessments included pain intensity, physical function, pain catastrophizing, and self-efficacy. Baseline comparisons between Black and non-Hispanic White participants were made. Mediation analyses used a regression-based framework to examine whether baseline pain catastrophizing and self-efficacy mediated the association between race and pain intensity. Moderation analyses used multiple linear regression to evaluate the role of race in the relationship of baseline pain catastrophizing and self-efficacy with 12-week pain intensity outcomes.
A total of 274 participants were included (51 [18.6%] Black and 223 [81.4%] non-Hispanic White; mean age = 51.6 years [standard deviation = 14.9]; 180 [65.7%] female). At baseline, Black participants had higher pain intensity scores (mean difference = 0.80; 95% CI =1.5 to 0.12). Both pain catastrophizing and self-efficacy mediated the relationship between race and baseline pain intensity. Race moderated the relationships between baseline pain catastrophizing and self-efficacy and 12-week pain intensity scores.
Pain catastrophizing and self-efficacy had differential impacts on pain intensity based on race for both cross-sectional and longitudinal analyses among persons receiving physical therapy for chronic spinal pain.
Improved understanding of the differences in pain experience based on factors such as race, ethnicity, cultural background, and experience with the health care system may help reduce disparities in pain management.
Introduction
Chronic pain, specifically back and neck pain, is a leading cause of disability globally.1 In the United States, spinal pain conditions are a leading cause for health care visits and the most common conditions encountered in outpatient physical therapy settings.2,3 The medical field now understands that an individual’s experience with chronic spinal pain is rarely attributable to, or predicted by, identifiable pathology, but rather is impacted by a wide array of biopsychosocial influences.4,5 Intervention strategies used by physical therapists and other providers are increasingly focused on understanding and treating psychosocial factors impacting the pain experience.6,7
A common theoretical model for understanding the role of psychosocial factors in persons with chronic pain is the fear-avoidance model (FAM).8,9 Pain catastrophizing, defined as a state of anxiety toward pain characterized by negative cognitive and emotional processes such as helplessness, pessimism, and rumination about pain-related symptoms,7 is a key component of the FAM and a predictor of poor pain outcomes in persons with chronic pain.10,11 More recently, the important role of pain resiliency factors in an individual’s response to chronic pain has been recognized. Pain self-efficacy, or an individual’s perceived ability to function in spite of pain,12 is one resilience factor that has been found to enhance the predictive value of the traditional FAM.13 Self-efficacy has been explored as a mediator between psychosocial factors in the FAM model and disability. While this is emerging work, it does support self-efficacy’s role in outcomes of physical therapy.14 Interventions designed to improve pain self-efficacy may be particularly beneficial for persons with chronic pain.15
The experience of chronic pain also differs based on race. Specifically, several studies have found that Black persons report higher levels of pain intensity for many clinical pain conditions and in experimental pain studies when compared to non-Hispanic White persons.16–21 Black persons with chronic pain also report higher levels of pain catastrophizing,20,22 and a few studies report that pain catastrophizing mediates differences in pain intensity by race.23,24 Race-based differences in resilience factors in persons with chronic pain have also been reported,25,26 but few studies have examined resilience factors or evaluated them as potential mediators in the relationship between race and pain intensity in persons with chronic pain. We are not aware of prior studies explicitly examining the FAM framework and race, but several studies highlight racialized differences in pain coping.20 Additionally, race and racism have been shown to impact treatment and outcomes for anxiety-related disorders including obsessive–compulsive disorder and posttraumatic stress disorder, thus it would not be surprising if these factors also impact physical therapy for chronic pain conditions.27–29
Prior studies examining relationships among race, psychosocial factors, and pain intensity have primarily used cross-sectional designs and nonclinical populations.16–18,20,24 No prior studies have examined patients with chronic pain receiving physical therapy. Patients with chronic pain receiving physical therapy are a clinical population for which the longitudinal impacts of pain catastrophizing or self-efficacy are particularly relevant because these psychosocial factors are known to influence outcomes in this setting,30 and physical therapists often use intervention strategies for persons with chronic pain that are intended to positively influence pain catastrophizing and self-efficacy.31,32 Evaluating the possible moderating effect of race on the relationship between baseline levels of pain catastrophizing or self-efficacy and longitudinal outcomes for persons with chronic pain receiving physical therapy has not been examined. Considering the importance of pain catastrophizing and self-efficacy as outcome predictors and concerns about race-based disparities in health outcomes, examining the role of these factors in explaining race-based differences in the pain experience, and prediction of outcomes in a physical therapy setting is warranted.
The purpose of this secondary analysis of a cluster-randomized clinical trial33 was to examine the role of pain catastrophizing and pain self-efficacy as possible mediators of race-based differences in self-reported pain intensity and to evaluate the possible moderating role of race on the relationship between pain catastrophizing and pain self-efficacy with pain outcomes among persons with chronic spinal pain receiving physical therapy. We hypothesized that self-reported pain intensity at baseline would differ between Black participants and non-Hispanic White participants, and pain catastrophizing and pain self-efficacy would each mediate this difference. We further hypothesized that race would moderate the relationship between baseline pain catastrophizing and self-efficacy with pain intensity outcomes assessed after 12 weeks.
Methods
Study Design
This investigation is a secondary analysis of a cluster-randomized trial investigating the influence of training physical therapists in pain neuroscience education (PNE) on patient-reported outcomes for patients with chronic neck or back pain receiving physical therapy.33 For the parent study, regions of physical therapy clinics were randomly assigned to either train physical therapists in providing PNE or receive no additional training. Patients beginning care in a participating clinic for chronic spinal pain were then consented to provide additional baseline data and outcome measures. Other than the PNE training, there were no attempts to direct the content or amount of physical therapist care in any way as a result of study participation. The study protocol is published34 and was approved by the University of Utah Institutional Review Board. The trial was registered prior to initiating enrollment (NCT03168165 on Clinicaltrials.gov), and the results are published.33 There was no treatment effect for the outcome of pain intensity as a result of treatment group in the parent study.33
Participants
Patient participants were recruited from 42 outpatient physical therapy clinics of BenchMark Rehab Partners located in the Atlanta, Georgia, or Birmingham, Alabama area from May 2017 to February 2019. Eligible participants were between 18 and 75 years of age at the time of their first physical therapy session and primarily seeking care for chronic low back pain and/or neck pain. Chronic pain was defined using the NIH Task Force definition (ie, neck pain and/or low back pain has been present for at least 3 months and on at least half the days in the past 6 months).35 Patients were excluded if they had physical therapy for spinal pain within the previous 6 months, spinal surgery within the previous 12 months, evidence of a red flag condition (eg, cauda equine syndrome, cancer, fracture, infection, or systemic disease) that required immediate referral to medical care, or were knowingly pregnant at the time of recruitment. For these analyses, we included participants from the clinical trial whose race was self-reported as either Black/African American or White/Caucasian and whose ethnicity was not reported as Hispanic/Latino, and who provided a baseline rating of pain intensity. We excluded Hispanic White/Caucasian participants from this secondary analysis as pain outcomes between Hispanic and non-Hispanic Whites has been shown to differ.36
Assessments
Participants were recruited and provided baseline assessment data via a RedCap (Vanderbilt University, Nashville, TN, USA) survey emailed to the participant prior to their first physical therapy session. Baseline data included self-reported spinal pain history, demographic information, and comorbid health conditions. Follow-up assessments were completed using RedCap at 2 and 12 weeks after the baseline assessment. The primary outcome for the parent trial, the PROMIS Physical Function computer adaptive test (PF-CAT),37 was evaluated at each assessment. The PF-CAT is reported as a t score with a population mean of 50 and standard deviation of 10. Higher scores indicate greater physical function.
The measures specific to this investigation assessed at baseline and follow-up assessments included pain intensity evaluated using a 0 to 10 numeric pain rating scale.38 Specifically, we asked participants to report the worst pain intensity experienced in the past 24 hours. Pain self-efficacy and pain catastrophizing were assessed with the Pain Self-Efficacy Questionnaire (PSEQ)12 and Pain Catastrophizing Scale (PCS),39 respectively. The PSEQ measures the respondent’s belief in their ability to perform activities despite pain using 10 items with responses rated on a 7-point scale ranging from 0 (“not confident at all”) to 6 (“completely confident”) summed to provide a score from 0 to 60, with higher scores indicating greater self-efficacy.12 The PSEQ demonstrates excellent validity, reliability (intraclass correlation coefficient [ICC] = 0.89), and responsiveness (effect size = 0.53, standardized response mean = 0.63).40 The PCS is a 13-item scale assessing the extent to which people catastrophize in response to pain, capturing subdomains of helplessness, rumination, and magnification. Each item is scored from 0 to 4 and summed to a total score, with higher scores indicating greater catastrophizing.39 The PCS demonstrates good internal reliability (α = .92) and test–retest reliability scores (Spearman ρ = 0.88).41
Data Analysis
Baseline data were examined for all participants using mean (standard deviation) for continuous measures and frequency (percentages) for categorical measures. We compared baseline values between non-Hispanic White participants and Black participants using independent group t tests or nonparametric equivalent for continuous variables and χ square tests of association for categorical variables. Pearson correlation coefficients were used to examine bivariate relationships among baseline pain catastrophizing, pain self-efficacy, and pain intensity at baseline and after 12 weeks. Correlations were examined for all participants and separately by race.
Mediation analyses were conducted using a regression-based framework42 to examine whether baseline PCS or PSEQ scores mediated the association between race and baseline pain intensity. In this mediation analysis, race is used as an exposure. In the field of epidemiology, the term exposure has a wide application and refers to any factor that could potentially be linked to an outcome.43 A complete case analysis was conducted, including only participants with complete data on the mediator and outcome. Covariates included in each model were age and sex as well as any baseline variable that differed between Black participants and non-Hispanic White participants. Sex, living situation, comorbid hypertension, and anxiety were added as dichotomous variables. Age and body mass index were added as continuous variables. These covariates were informed by literature identifying factors impacting outcomes for low back pain and by our evaluation of baseline differences to avoid potential confounding. Mediation analyses were conducted using the Hayes PROCESS macro (model 4, version 3.5) in IBM SPSS for Windows, version 26.0 (IBM Corporation, Armonk, NY, USA). We estimated the direct, indirect, and total effects of race and pain intensity, with 5000 bootstrap resamples used to calculate 95% CI around these parameter estimates. The total effect (Fig. 1) represents the unstandardized regression coefficient or path (c) from race to pain intensity after adjusting for covariates. The total effect is the sum of the direct and indirect effects. The direct effect is the unstandardized path (c′) from race to pain intensity once the putative mediator is added to the model. The total indirect effect (c – c′) represents the difference between the effect of race on pain intensity when the mediator is controlled versus when it is not. Inference about the presence of mediation is based on the indirect effect. If the 95% CI does not include zero, the null hypothesis of no mediation can be rejected.44
Path diagram of the mediation model for pain catastrophizing on race and baseline pain intensity. Covariates in the model include age, sex, body mass index (BMI), living situation, comorbid anxiety, and comorbid hypertension. Path a represents the unstandardized path coefficient of pain catastrophizing regressed on race adjusted for covariates. Path b represents the unstandardized path coefficient of pain intensity regressed on pain catastrophizing, adjusted for covariates. Path c represents the unstandardized path coefficient of race regressed on pain intensity, adjusted for covariates without controlling for pain catastrophizing. Path c’ represents the unstandardized path coefficient of race regressed on pain intensity, after controlling for covariates and pain catastrophizing. The total indirect effect (c – c’) indicates the difference between the effect of race on pain intensity when pain catastrophizing is controlled versus when it is not. Pain catastrophizing is supported as a mediator because the 95% CI for the total indirect effect does not include zero. M = mediator; X = independent variable; Y = dependent variable.
We examined the potential moderating role of race in the relationship of baseline PCS and PSEQ scores and pain intensity outcome using multiple linear regression. A treatment moderator is an initial or preexisting variable that identifies specific patient subgroups in the population with varying magnitudes of treatment effects.45 Twelve-week pain intensity was the dependent variable. The base model included the previous covariates and baseline pain intensity, race, number of physical therapy sessions attended, randomized treatment group, and baseline PCS and PSEQ scores. In the second step, 2 interaction terms (race × PCS and race × PSEQ) were added to the base model. Adjusted R2 values from each step were used to determine the degree of explained variance in 12-week pain intensity. Assumptions for linear regression were assessed for the final model through inspection of the predicted probability (P–P) plot of residuals and scatterplot of residuals versus predicted values. Cook’s distance was used to identify outlier cases if the distance was ≥1.0. The variance inflation factor (VIF) of variables in the final model were evaluated for any evidence of multicollinearity if VIF ≥ 10.0. Unstandardized regression coefficient with 95% CIs was examined to identify variables in the final model that contributed significantly to explaining variance in the 12-week pain intensity outcome. Significant interactions were further explored through plotting the unstandardized predicted values for 12-week pain intensity by race.
Results
A total of 319 participants enrolled in the clinical trial, of which 274 (85.9%) met the criteria for inclusion in these analyses. Of these participants, 51 (18.6%) were Black and 223 (81.4%) were non-Hispanic White. Overall, included participants included had a mean age of 51.6 years (standard deviation = 14.9) and 180 (65.7%) were female. Thirteen participants provided height and weight data that resulted in missing or out of range body mass index information. For these individuals, body mass index was imputed using predicted values from regression model with age, sex, race, comorbid diabetes, and hypertension in the model.
Participants’ baseline characteristics by race are outlined in Table 1. As anticipated, significant differences were observed between Black participants and non-Hispanic White participants for baseline pain intensity and the PCS and PSEQ. Differences in living situation, body mass index, comorbid anxiety, and hypertension by race were identified. The Supplementary Table provides variables at the 12-week assessment by race for participants with available 12-week pain intensity scores (n = 220; 80.3%). No differences were noted in pain intensity or treatment variables by race. Zero-order correlations between pain intensity and potential mediators are shown in Table 2.
Baseline Characteristics of the Patient Samplea
| Baseline Variable | All Patients (n = 274) | Black (n = 51) | Non-Hispanic White (n = 223) | P | Mean Difference (95% CI) |
|---|---|---|---|---|---|
| Age, mean y (SD) | 51.6 (14.9) | 48.5 (15.4) | 52.3 (14.8) | .10 | 3.8 (−0.73 to 8.3) |
| Sex, n female (%) | 180 (65.7.2%) | 36 (70.6%) | 144 (64.6%) | .41 | |
| Living situation, n married or living with significant other (%) | 187 (68.2%) | 26 (51.0%) | 161 (72.2%) | .003b | |
| Education, n with college degree (%) | 150 (54.9%) (n = 273) | 24 (47.1%) | 126 (56.8%) (n = 222) | .21 | |
| Currently working full-time, n (%) | 129 (47.6%) (n = 271) | 28 (56.0%) (n = 50) | 101 (45.7%) (n = 221) | .19 | |
| Body mass index, mean kg/m2 (SD) | 29.7 (7.9) | 32.3 (10.2) | 29.1 (7.2) | .009b | −3.5 (−6.1 to −0.82) |
| Current smoker, n (%) | 15 (5.5%) | 2 (3.9%) | 13 (5.8%) | .12 | |
| Diabetes, n (%) | 43 (16.0%) (n = 268) | 11 (21.6%) | 32 (14.7%) (n = 217) | .23 | |
| Depression, n (%) | 58 (21.6%) (n = 268) | 9 (18.0%) (n = 50) | 49 (22.5%) (n = 218) | .48 | |
| Anxiety, n (%) | 64 (24.0%) (n = 264) | 6 (12.2%) (n = 49) | 58 (26.2%) (n = 218) | .033b | |
| Hypertension, n (%) | 89 (33.0%) (n = 270) | 23 (45.1%) | 66 (30.1%) (n = 219) | .041b | |
| Prior physical therapy for spine pain, n (%) | 121 (44.2%) | 22 (43.1%) | 99 (4.4%) | .87 | |
| Prior spine surgery, n (%) | 42 (15.3%) | 6 (11.8%) | 36 (16.1%) | .43 | |
| Pain duration >6 mo, n (%) | 207 (77.0%) (n = 269) | 35 (72.9%) (n = 48) | 172 (77.8%) (n = 221) | .43 | |
| Worst pain intensity in past 24 h, mean (SD) | 6.7 (2.2) | 7.3 (1.9) | 6.5 (2.3) | .021b | −0.80 (−1.5 to −0.12) |
| Pain catastrophizing, mean (SD) | 20.0 (12.8) (n = 260) | 25.9 (13.6) (n = 50) | 18.6 (12.3) (n = 210) | <.001b | −7.3 (−11.9 to −3.4) |
| Pain self-efficacy, mean (SD) | 35.5 (15.2) (n = 260) | 31.2 (17.1) (n = 50) | 36.5 (14.6) (n = 210) | .025b | 5.4 (0.67 to 10.0) |
| PROMIS physical function, mean t score (SD) | 41.0 (7.5) | 41.1 (8.7) | 41.0 (7.6) | .88 | −0.17 (−.25 to 2.2) |
| Baseline Variable | All Patients (n = 274) | Black (n = 51) | Non-Hispanic White (n = 223) | P | Mean Difference (95% CI) |
|---|---|---|---|---|---|
| Age, mean y (SD) | 51.6 (14.9) | 48.5 (15.4) | 52.3 (14.8) | .10 | 3.8 (−0.73 to 8.3) |
| Sex, n female (%) | 180 (65.7.2%) | 36 (70.6%) | 144 (64.6%) | .41 | |
| Living situation, n married or living with significant other (%) | 187 (68.2%) | 26 (51.0%) | 161 (72.2%) | .003b | |
| Education, n with college degree (%) | 150 (54.9%) (n = 273) | 24 (47.1%) | 126 (56.8%) (n = 222) | .21 | |
| Currently working full-time, n (%) | 129 (47.6%) (n = 271) | 28 (56.0%) (n = 50) | 101 (45.7%) (n = 221) | .19 | |
| Body mass index, mean kg/m2 (SD) | 29.7 (7.9) | 32.3 (10.2) | 29.1 (7.2) | .009b | −3.5 (−6.1 to −0.82) |
| Current smoker, n (%) | 15 (5.5%) | 2 (3.9%) | 13 (5.8%) | .12 | |
| Diabetes, n (%) | 43 (16.0%) (n = 268) | 11 (21.6%) | 32 (14.7%) (n = 217) | .23 | |
| Depression, n (%) | 58 (21.6%) (n = 268) | 9 (18.0%) (n = 50) | 49 (22.5%) (n = 218) | .48 | |
| Anxiety, n (%) | 64 (24.0%) (n = 264) | 6 (12.2%) (n = 49) | 58 (26.2%) (n = 218) | .033b | |
| Hypertension, n (%) | 89 (33.0%) (n = 270) | 23 (45.1%) | 66 (30.1%) (n = 219) | .041b | |
| Prior physical therapy for spine pain, n (%) | 121 (44.2%) | 22 (43.1%) | 99 (4.4%) | .87 | |
| Prior spine surgery, n (%) | 42 (15.3%) | 6 (11.8%) | 36 (16.1%) | .43 | |
| Pain duration >6 mo, n (%) | 207 (77.0%) (n = 269) | 35 (72.9%) (n = 48) | 172 (77.8%) (n = 221) | .43 | |
| Worst pain intensity in past 24 h, mean (SD) | 6.7 (2.2) | 7.3 (1.9) | 6.5 (2.3) | .021b | −0.80 (−1.5 to −0.12) |
| Pain catastrophizing, mean (SD) | 20.0 (12.8) (n = 260) | 25.9 (13.6) (n = 50) | 18.6 (12.3) (n = 210) | <.001b | −7.3 (−11.9 to −3.4) |
| Pain self-efficacy, mean (SD) | 35.5 (15.2) (n = 260) | 31.2 (17.1) (n = 50) | 36.5 (14.6) (n = 210) | .025b | 5.4 (0.67 to 10.0) |
| PROMIS physical function, mean t score (SD) | 41.0 (7.5) | 41.1 (8.7) | 41.0 (7.6) | .88 | −0.17 (−.25 to 2.2) |
Comparisons were made using χ square tests for categorical variables and independent group t tests for continuous variables. Mean differences with 95% CI are provided for comparisons of continuous variables. n = number; PROMIS = Patient-Reported Outcomes Measurement Information System.
P < .05.
Baseline Characteristics of the Patient Samplea
| Baseline Variable | All Patients (n = 274) | Black (n = 51) | Non-Hispanic White (n = 223) | P | Mean Difference (95% CI) |
|---|---|---|---|---|---|
| Age, mean y (SD) | 51.6 (14.9) | 48.5 (15.4) | 52.3 (14.8) | .10 | 3.8 (−0.73 to 8.3) |
| Sex, n female (%) | 180 (65.7.2%) | 36 (70.6%) | 144 (64.6%) | .41 | |
| Living situation, n married or living with significant other (%) | 187 (68.2%) | 26 (51.0%) | 161 (72.2%) | .003b | |
| Education, n with college degree (%) | 150 (54.9%) (n = 273) | 24 (47.1%) | 126 (56.8%) (n = 222) | .21 | |
| Currently working full-time, n (%) | 129 (47.6%) (n = 271) | 28 (56.0%) (n = 50) | 101 (45.7%) (n = 221) | .19 | |
| Body mass index, mean kg/m2 (SD) | 29.7 (7.9) | 32.3 (10.2) | 29.1 (7.2) | .009b | −3.5 (−6.1 to −0.82) |
| Current smoker, n (%) | 15 (5.5%) | 2 (3.9%) | 13 (5.8%) | .12 | |
| Diabetes, n (%) | 43 (16.0%) (n = 268) | 11 (21.6%) | 32 (14.7%) (n = 217) | .23 | |
| Depression, n (%) | 58 (21.6%) (n = 268) | 9 (18.0%) (n = 50) | 49 (22.5%) (n = 218) | .48 | |
| Anxiety, n (%) | 64 (24.0%) (n = 264) | 6 (12.2%) (n = 49) | 58 (26.2%) (n = 218) | .033b | |
| Hypertension, n (%) | 89 (33.0%) (n = 270) | 23 (45.1%) | 66 (30.1%) (n = 219) | .041b | |
| Prior physical therapy for spine pain, n (%) | 121 (44.2%) | 22 (43.1%) | 99 (4.4%) | .87 | |
| Prior spine surgery, n (%) | 42 (15.3%) | 6 (11.8%) | 36 (16.1%) | .43 | |
| Pain duration >6 mo, n (%) | 207 (77.0%) (n = 269) | 35 (72.9%) (n = 48) | 172 (77.8%) (n = 221) | .43 | |
| Worst pain intensity in past 24 h, mean (SD) | 6.7 (2.2) | 7.3 (1.9) | 6.5 (2.3) | .021b | −0.80 (−1.5 to −0.12) |
| Pain catastrophizing, mean (SD) | 20.0 (12.8) (n = 260) | 25.9 (13.6) (n = 50) | 18.6 (12.3) (n = 210) | <.001b | −7.3 (−11.9 to −3.4) |
| Pain self-efficacy, mean (SD) | 35.5 (15.2) (n = 260) | 31.2 (17.1) (n = 50) | 36.5 (14.6) (n = 210) | .025b | 5.4 (0.67 to 10.0) |
| PROMIS physical function, mean t score (SD) | 41.0 (7.5) | 41.1 (8.7) | 41.0 (7.6) | .88 | −0.17 (−.25 to 2.2) |
| Baseline Variable | All Patients (n = 274) | Black (n = 51) | Non-Hispanic White (n = 223) | P | Mean Difference (95% CI) |
|---|---|---|---|---|---|
| Age, mean y (SD) | 51.6 (14.9) | 48.5 (15.4) | 52.3 (14.8) | .10 | 3.8 (−0.73 to 8.3) |
| Sex, n female (%) | 180 (65.7.2%) | 36 (70.6%) | 144 (64.6%) | .41 | |
| Living situation, n married or living with significant other (%) | 187 (68.2%) | 26 (51.0%) | 161 (72.2%) | .003b | |
| Education, n with college degree (%) | 150 (54.9%) (n = 273) | 24 (47.1%) | 126 (56.8%) (n = 222) | .21 | |
| Currently working full-time, n (%) | 129 (47.6%) (n = 271) | 28 (56.0%) (n = 50) | 101 (45.7%) (n = 221) | .19 | |
| Body mass index, mean kg/m2 (SD) | 29.7 (7.9) | 32.3 (10.2) | 29.1 (7.2) | .009b | −3.5 (−6.1 to −0.82) |
| Current smoker, n (%) | 15 (5.5%) | 2 (3.9%) | 13 (5.8%) | .12 | |
| Diabetes, n (%) | 43 (16.0%) (n = 268) | 11 (21.6%) | 32 (14.7%) (n = 217) | .23 | |
| Depression, n (%) | 58 (21.6%) (n = 268) | 9 (18.0%) (n = 50) | 49 (22.5%) (n = 218) | .48 | |
| Anxiety, n (%) | 64 (24.0%) (n = 264) | 6 (12.2%) (n = 49) | 58 (26.2%) (n = 218) | .033b | |
| Hypertension, n (%) | 89 (33.0%) (n = 270) | 23 (45.1%) | 66 (30.1%) (n = 219) | .041b | |
| Prior physical therapy for spine pain, n (%) | 121 (44.2%) | 22 (43.1%) | 99 (4.4%) | .87 | |
| Prior spine surgery, n (%) | 42 (15.3%) | 6 (11.8%) | 36 (16.1%) | .43 | |
| Pain duration >6 mo, n (%) | 207 (77.0%) (n = 269) | 35 (72.9%) (n = 48) | 172 (77.8%) (n = 221) | .43 | |
| Worst pain intensity in past 24 h, mean (SD) | 6.7 (2.2) | 7.3 (1.9) | 6.5 (2.3) | .021b | −0.80 (−1.5 to −0.12) |
| Pain catastrophizing, mean (SD) | 20.0 (12.8) (n = 260) | 25.9 (13.6) (n = 50) | 18.6 (12.3) (n = 210) | <.001b | −7.3 (−11.9 to −3.4) |
| Pain self-efficacy, mean (SD) | 35.5 (15.2) (n = 260) | 31.2 (17.1) (n = 50) | 36.5 (14.6) (n = 210) | .025b | 5.4 (0.67 to 10.0) |
| PROMIS physical function, mean t score (SD) | 41.0 (7.5) | 41.1 (8.7) | 41.0 (7.6) | .88 | −0.17 (−.25 to 2.2) |
Comparisons were made using χ square tests for categorical variables and independent group t tests for continuous variables. Mean differences with 95% CI are provided for comparisons of continuous variables. n = number; PROMIS = Patient-Reported Outcomes Measurement Information System.
P < .05.
Zero-Order Correlations With 95% CI Among Mediators and Pain Intensity Scores at Baseline and After 12-Weeks
| Pain Self-Efficacy | Pain Catastrophizing | Pain Intensity—Baseline | Pain Intensity—12 Weeks | ||
|---|---|---|---|---|---|
| Pain self-efficacy | All participants | - - - - - - - | −0.60a (−0.51 to −0.67) (n = 259) | −0.39a (−0.28 to −0.49) (n = 260) | −0.32a (−0.19 to −0.44) (n = 212) |
| Black | - - - - - - - | −0.64a (−0.44 to −0.78) (n = 49) | −0.42a (−0.16 to −0.62) (n = 50) | −0.29 (0.03 to −0.56) (n = 38) | |
| Non-Hispanic White | - - - - - - - | −0.57a (−0.47 to −0.66) (n = 210) | −0.37a (−0.25 to −0.48) (n = 210) | −0.32a (−0.18 to −0.45) (n = 174) | |
| Pain catastrophizing | All participants | - - - - - - - | - - - - - - - | 0.39a (0.28 to 0.49) (n = 260) | 0.34a (0.21 to 0.46) (n = 212) |
| Black | - - - - - - - | - - - - - - - | 0.47a (0.22 to 0.66) (n = 50) | 0.44b (0.14 to 0.67) (n = 38) | |
| Non-Hispanic White | - - - - - - - | - - - - - - - | 0.35a (0.22 to 0.46) (n = 210) | 0.30a (0.15 to 0.43) (n = 174) | |
| Pain intensity—baseline | All participants | - - - - - - - | - - - - - - - | - - - - - - - | 0.41a (0.30 to 0.52) (n = 220) |
| Black | - - - - - - - | - - - - - - - | - - - - - - - | 0.59a (0.34 to 0.77) (n = 39) | |
| Non-Hispanic White | - - - - - - - | - - - - - - - | - - - - - - - | 0.38a (0.25 to 0.50) (n = 181) | |
| Pain Self-Efficacy | Pain Catastrophizing | Pain Intensity—Baseline | Pain Intensity—12 Weeks | ||
|---|---|---|---|---|---|
| Pain self-efficacy | All participants | - - - - - - - | −0.60a (−0.51 to −0.67) (n = 259) | −0.39a (−0.28 to −0.49) (n = 260) | −0.32a (−0.19 to −0.44) (n = 212) |
| Black | - - - - - - - | −0.64a (−0.44 to −0.78) (n = 49) | −0.42a (−0.16 to −0.62) (n = 50) | −0.29 (0.03 to −0.56) (n = 38) | |
| Non-Hispanic White | - - - - - - - | −0.57a (−0.47 to −0.66) (n = 210) | −0.37a (−0.25 to −0.48) (n = 210) | −0.32a (−0.18 to −0.45) (n = 174) | |
| Pain catastrophizing | All participants | - - - - - - - | - - - - - - - | 0.39a (0.28 to 0.49) (n = 260) | 0.34a (0.21 to 0.46) (n = 212) |
| Black | - - - - - - - | - - - - - - - | 0.47a (0.22 to 0.66) (n = 50) | 0.44b (0.14 to 0.67) (n = 38) | |
| Non-Hispanic White | - - - - - - - | - - - - - - - | 0.35a (0.22 to 0.46) (n = 210) | 0.30a (0.15 to 0.43) (n = 174) | |
| Pain intensity—baseline | All participants | - - - - - - - | - - - - - - - | - - - - - - - | 0.41a (0.30 to 0.52) (n = 220) |
| Black | - - - - - - - | - - - - - - - | - - - - - - - | 0.59a (0.34 to 0.77) (n = 39) | |
| Non-Hispanic White | - - - - - - - | - - - - - - - | - - - - - - - | 0.38a (0.25 to 0.50) (n = 181) | |
P < .01.
P < .05.
Zero-Order Correlations With 95% CI Among Mediators and Pain Intensity Scores at Baseline and After 12-Weeks
| Pain Self-Efficacy | Pain Catastrophizing | Pain Intensity—Baseline | Pain Intensity—12 Weeks | ||
|---|---|---|---|---|---|
| Pain self-efficacy | All participants | - - - - - - - | −0.60a (−0.51 to −0.67) (n = 259) | −0.39a (−0.28 to −0.49) (n = 260) | −0.32a (−0.19 to −0.44) (n = 212) |
| Black | - - - - - - - | −0.64a (−0.44 to −0.78) (n = 49) | −0.42a (−0.16 to −0.62) (n = 50) | −0.29 (0.03 to −0.56) (n = 38) | |
| Non-Hispanic White | - - - - - - - | −0.57a (−0.47 to −0.66) (n = 210) | −0.37a (−0.25 to −0.48) (n = 210) | −0.32a (−0.18 to −0.45) (n = 174) | |
| Pain catastrophizing | All participants | - - - - - - - | - - - - - - - | 0.39a (0.28 to 0.49) (n = 260) | 0.34a (0.21 to 0.46) (n = 212) |
| Black | - - - - - - - | - - - - - - - | 0.47a (0.22 to 0.66) (n = 50) | 0.44b (0.14 to 0.67) (n = 38) | |
| Non-Hispanic White | - - - - - - - | - - - - - - - | 0.35a (0.22 to 0.46) (n = 210) | 0.30a (0.15 to 0.43) (n = 174) | |
| Pain intensity—baseline | All participants | - - - - - - - | - - - - - - - | - - - - - - - | 0.41a (0.30 to 0.52) (n = 220) |
| Black | - - - - - - - | - - - - - - - | - - - - - - - | 0.59a (0.34 to 0.77) (n = 39) | |
| Non-Hispanic White | - - - - - - - | - - - - - - - | - - - - - - - | 0.38a (0.25 to 0.50) (n = 181) | |
| Pain Self-Efficacy | Pain Catastrophizing | Pain Intensity—Baseline | Pain Intensity—12 Weeks | ||
|---|---|---|---|---|---|
| Pain self-efficacy | All participants | - - - - - - - | −0.60a (−0.51 to −0.67) (n = 259) | −0.39a (−0.28 to −0.49) (n = 260) | −0.32a (−0.19 to −0.44) (n = 212) |
| Black | - - - - - - - | −0.64a (−0.44 to −0.78) (n = 49) | −0.42a (−0.16 to −0.62) (n = 50) | −0.29 (0.03 to −0.56) (n = 38) | |
| Non-Hispanic White | - - - - - - - | −0.57a (−0.47 to −0.66) (n = 210) | −0.37a (−0.25 to −0.48) (n = 210) | −0.32a (−0.18 to −0.45) (n = 174) | |
| Pain catastrophizing | All participants | - - - - - - - | - - - - - - - | 0.39a (0.28 to 0.49) (n = 260) | 0.34a (0.21 to 0.46) (n = 212) |
| Black | - - - - - - - | - - - - - - - | 0.47a (0.22 to 0.66) (n = 50) | 0.44b (0.14 to 0.67) (n = 38) | |
| Non-Hispanic White | - - - - - - - | - - - - - - - | 0.35a (0.22 to 0.46) (n = 210) | 0.30a (0.15 to 0.43) (n = 174) | |
| Pain intensity—baseline | All participants | - - - - - - - | - - - - - - - | - - - - - - - | 0.41a (0.30 to 0.52) (n = 220) |
| Black | - - - - - - - | - - - - - - - | - - - - - - - | 0.59a (0.34 to 0.77) (n = 39) | |
| Non-Hispanic White | - - - - - - - | - - - - - - - | - - - - - - - | 0.38a (0.25 to 0.50) (n = 181) | |
P < .01.
P < .05.
Figures 1 and 2 display the mediation models examining pain catastrophizing and pain self-efficacy, respectively, and the relationship between race and baseline pain intensity. Covariates in each model were age, sex, body mass index, living situation, and comorbid hypertension and anxiety. Because zero was excluded from the 95% CI for the total indirect effect, both pain catastrophizing and pain self-efficacy were supported as mediators of the relationship between baseline pain intensity and race.
Path diagram of the mediation model for pain self-efficacy on race and baseline pain intensity. Covariates in the model include age, sex, body mass index (BMI), living situation, comorbid anxiety, and comorbid hypertension. Path a represents the unstandardized path coefficient of pain self-efficacy regressed on race adjusted for covariates. Path b represents the unstandardized path coefficient of pain intensity regressed on pain self-efficacy, adjusted for covariates. Path c represents the unstandardized path coefficient of race regressed on pain intensity, adjusted for covariates without controlling for pain self-efficacy. Path c’ represents the unstandardized path coefficient of race regressed on pain intensity, after controlling for covariates and pain self-efficacy. The total indirect effect (c – c’) indicates the difference between the effect of race on pain intensity when pain self-efficacy is controlled versus when it is not. Pain self-efficacy is supported as a mediator because the 95% CI for the total indirect effect does not include zero. M = mediator; X = independent variable; Y = dependent variable.
Outcomes of the regression model examining predictors of 12-week pain intensity are provided in Table 3. The base regression model explained a significant proportion of the variance in 12-week pain intensity (adjusted R2 = 0.23; P < .001). Addition of the interaction terms added significantly to the explained variance of the base model (adjusted R2 = 0.25; P < .042). The interaction between race × pain self-efficacy (unstandardized β = .087; 95% CI = 0.004; 0.17; P = .040) and between race × pain catastrophizing (unstandardized β = .14; 95% CI = 0.03 to 0.24; P = .013) both contributed significantly to the final model. The nature of these interactions was graphed (Figure 3A and B). The regression equations to predict 12-week pain intensity from baseline PCS score by race (Fig. 3A) indicates that at very low baseline PCS scores, non-Hispanic White participants were predicted to have higher 12-week pain intensity outcomes; however, for each 1-point increase in PCS, the predicted increase in 12-week pain intensity was greater for Black participants (0.14) versus non-Hispanic White participants (0.06) such that at high baseline PCS scores Black participants were predicted to have higher 12-week pain intensity outcomes. The regression equations to predict 12-week pain intensity from baseline PSEQ score by race (Fig. 3B) indicate that for very low PSEQ scores, both Black participants and non-Hispanic White participants were predicted to have high 12-week pain intensity outcomes. For each 1-point increase in PSEQ however, the predicted decrease in 12-week pain intensity outcomes was greater for non-Hispanic White participants (−0.07) than for Black participants (−0.04). Thus, at high baseline PSEQ score, Black participants were predicted to have higher 12-week pain intensity outcomes (Fig. 3B).
Linear Regression Model Evaluating the Impact of Baseline Pain Catastrophizing, Pain Self-Efficacy, Race, and Interactions Terms on the Prediction of 12-Week Pain Intensity Scores Adjusting for Covariates
| Variables | Unstandardized β (95% CI) | Standardized β | P |
|---|---|---|---|
| Base model: adjusted R2 (P) = .23 (≤.001) | |||
| Age | .022 (−0.004 to 0.049) | .12 | .091 |
| Sex | .58 (−0.18 to 1.35) | .098 | .13 |
| BMI | −.011 (−0.063 to 0.041) | −.027 | .69 |
| Living situation | .49 (−0.32 to 1.30) | .082 | .23 |
| Hypertension | .38 (−0.51 to 1.27) | .061 | .40 |
| Anxiety | −.94 (−1.83 to −0.043) | −.14 | .040 |
| Baseline pain intensity | .41 (0.24 to 0.58) | .34 | <.001 |
| Number of visits | −.023 (−0.075 to 0.030) | −.055 | .39 |
| Treatment group | −.70 (−1.45 to 0.052) | −.12 | .068 |
| Race | −.21 (−1.21 to 0.79) | −.029 | .68 |
| Baseline Pain-Self Efficacy Questionnaire | −.019 (−0.052 to 0.014) | −.10 | .25 |
| Baseline Pain Catastrophizing Scale | .030 (−0.010 to 0.071) | .13 | .14 |
| Final model: adjusted R2 (P change in R2) = .25 (.042) | |||
| Age | .021 (−0.005 to 0.47) | .11 | .11 |
| Sex | .67 (−0.091 to 1.44) | .11 | .084 |
| BMI | −.012 (−0.063 to 0.040) | −.030 | .65 |
| Living situation | .47 (−0.33 to 1.27) | .078 | .25 |
| Hypertension | .34 (−0.54 to 1.22) | .054 | .45 |
| Anxiety | −.90 (−1.78 to −0.010) | −.13 | .048 |
| Baseline pain intensity | .41 (0.25 to 0.58) | .34 | <.001 |
| Number of visits | −.032 (−0.084 to 0.021) | −.076 | .24 |
| Treatment group | −.73 (−1.47 to 0.010) | −.12 | .053 |
| Race | .11 (−0.92 to 1.13) | .015 | .83 |
| Baseline Pain-Self Efficacy Questionnaire | −.031 (−0.066 to 0.005) | −.16 | .091 |
| Baseline Pain Catastrophizing Scale | .012 (−0.030 to 0.055) | .054 | .57 |
| Pain self-efficacy × race | .087 (0.004 to 0.17) | .23 | .040 |
| Pain catastrophizing × race | .14 (0.030 to 0.24) | .28 | .013 |
| Variables | Unstandardized β (95% CI) | Standardized β | P |
|---|---|---|---|
| Base model: adjusted R2 (P) = .23 (≤.001) | |||
| Age | .022 (−0.004 to 0.049) | .12 | .091 |
| Sex | .58 (−0.18 to 1.35) | .098 | .13 |
| BMI | −.011 (−0.063 to 0.041) | −.027 | .69 |
| Living situation | .49 (−0.32 to 1.30) | .082 | .23 |
| Hypertension | .38 (−0.51 to 1.27) | .061 | .40 |
| Anxiety | −.94 (−1.83 to −0.043) | −.14 | .040 |
| Baseline pain intensity | .41 (0.24 to 0.58) | .34 | <.001 |
| Number of visits | −.023 (−0.075 to 0.030) | −.055 | .39 |
| Treatment group | −.70 (−1.45 to 0.052) | −.12 | .068 |
| Race | −.21 (−1.21 to 0.79) | −.029 | .68 |
| Baseline Pain-Self Efficacy Questionnaire | −.019 (−0.052 to 0.014) | −.10 | .25 |
| Baseline Pain Catastrophizing Scale | .030 (−0.010 to 0.071) | .13 | .14 |
| Final model: adjusted R2 (P change in R2) = .25 (.042) | |||
| Age | .021 (−0.005 to 0.47) | .11 | .11 |
| Sex | .67 (−0.091 to 1.44) | .11 | .084 |
| BMI | −.012 (−0.063 to 0.040) | −.030 | .65 |
| Living situation | .47 (−0.33 to 1.27) | .078 | .25 |
| Hypertension | .34 (−0.54 to 1.22) | .054 | .45 |
| Anxiety | −.90 (−1.78 to −0.010) | −.13 | .048 |
| Baseline pain intensity | .41 (0.25 to 0.58) | .34 | <.001 |
| Number of visits | −.032 (−0.084 to 0.021) | −.076 | .24 |
| Treatment group | −.73 (−1.47 to 0.010) | −.12 | .053 |
| Race | .11 (−0.92 to 1.13) | .015 | .83 |
| Baseline Pain-Self Efficacy Questionnaire | −.031 (−0.066 to 0.005) | −.16 | .091 |
| Baseline Pain Catastrophizing Scale | .012 (−0.030 to 0.055) | .054 | .57 |
| Pain self-efficacy × race | .087 (0.004 to 0.17) | .23 | .040 |
| Pain catastrophizing × race | .14 (0.030 to 0.24) | .28 | .013 |
Linear Regression Model Evaluating the Impact of Baseline Pain Catastrophizing, Pain Self-Efficacy, Race, and Interactions Terms on the Prediction of 12-Week Pain Intensity Scores Adjusting for Covariates
| Variables | Unstandardized β (95% CI) | Standardized β | P |
|---|---|---|---|
| Base model: adjusted R2 (P) = .23 (≤.001) | |||
| Age | .022 (−0.004 to 0.049) | .12 | .091 |
| Sex | .58 (−0.18 to 1.35) | .098 | .13 |
| BMI | −.011 (−0.063 to 0.041) | −.027 | .69 |
| Living situation | .49 (−0.32 to 1.30) | .082 | .23 |
| Hypertension | .38 (−0.51 to 1.27) | .061 | .40 |
| Anxiety | −.94 (−1.83 to −0.043) | −.14 | .040 |
| Baseline pain intensity | .41 (0.24 to 0.58) | .34 | <.001 |
| Number of visits | −.023 (−0.075 to 0.030) | −.055 | .39 |
| Treatment group | −.70 (−1.45 to 0.052) | −.12 | .068 |
| Race | −.21 (−1.21 to 0.79) | −.029 | .68 |
| Baseline Pain-Self Efficacy Questionnaire | −.019 (−0.052 to 0.014) | −.10 | .25 |
| Baseline Pain Catastrophizing Scale | .030 (−0.010 to 0.071) | .13 | .14 |
| Final model: adjusted R2 (P change in R2) = .25 (.042) | |||
| Age | .021 (−0.005 to 0.47) | .11 | .11 |
| Sex | .67 (−0.091 to 1.44) | .11 | .084 |
| BMI | −.012 (−0.063 to 0.040) | −.030 | .65 |
| Living situation | .47 (−0.33 to 1.27) | .078 | .25 |
| Hypertension | .34 (−0.54 to 1.22) | .054 | .45 |
| Anxiety | −.90 (−1.78 to −0.010) | −.13 | .048 |
| Baseline pain intensity | .41 (0.25 to 0.58) | .34 | <.001 |
| Number of visits | −.032 (−0.084 to 0.021) | −.076 | .24 |
| Treatment group | −.73 (−1.47 to 0.010) | −.12 | .053 |
| Race | .11 (−0.92 to 1.13) | .015 | .83 |
| Baseline Pain-Self Efficacy Questionnaire | −.031 (−0.066 to 0.005) | −.16 | .091 |
| Baseline Pain Catastrophizing Scale | .012 (−0.030 to 0.055) | .054 | .57 |
| Pain self-efficacy × race | .087 (0.004 to 0.17) | .23 | .040 |
| Pain catastrophizing × race | .14 (0.030 to 0.24) | .28 | .013 |
| Variables | Unstandardized β (95% CI) | Standardized β | P |
|---|---|---|---|
| Base model: adjusted R2 (P) = .23 (≤.001) | |||
| Age | .022 (−0.004 to 0.049) | .12 | .091 |
| Sex | .58 (−0.18 to 1.35) | .098 | .13 |
| BMI | −.011 (−0.063 to 0.041) | −.027 | .69 |
| Living situation | .49 (−0.32 to 1.30) | .082 | .23 |
| Hypertension | .38 (−0.51 to 1.27) | .061 | .40 |
| Anxiety | −.94 (−1.83 to −0.043) | −.14 | .040 |
| Baseline pain intensity | .41 (0.24 to 0.58) | .34 | <.001 |
| Number of visits | −.023 (−0.075 to 0.030) | −.055 | .39 |
| Treatment group | −.70 (−1.45 to 0.052) | −.12 | .068 |
| Race | −.21 (−1.21 to 0.79) | −.029 | .68 |
| Baseline Pain-Self Efficacy Questionnaire | −.019 (−0.052 to 0.014) | −.10 | .25 |
| Baseline Pain Catastrophizing Scale | .030 (−0.010 to 0.071) | .13 | .14 |
| Final model: adjusted R2 (P change in R2) = .25 (.042) | |||
| Age | .021 (−0.005 to 0.47) | .11 | .11 |
| Sex | .67 (−0.091 to 1.44) | .11 | .084 |
| BMI | −.012 (−0.063 to 0.040) | −.030 | .65 |
| Living situation | .47 (−0.33 to 1.27) | .078 | .25 |
| Hypertension | .34 (−0.54 to 1.22) | .054 | .45 |
| Anxiety | −.90 (−1.78 to −0.010) | −.13 | .048 |
| Baseline pain intensity | .41 (0.25 to 0.58) | .34 | <.001 |
| Number of visits | −.032 (−0.084 to 0.021) | −.076 | .24 |
| Treatment group | −.73 (−1.47 to 0.010) | −.12 | .053 |
| Race | .11 (−0.92 to 1.13) | .015 | .83 |
| Baseline Pain-Self Efficacy Questionnaire | −.031 (−0.066 to 0.005) | −.16 | .091 |
| Baseline Pain Catastrophizing Scale | .012 (−0.030 to 0.055) | .054 | .57 |
| Pain self-efficacy × race | .087 (0.004 to 0.17) | .23 | .040 |
| Pain catastrophizing × race | .14 (0.030 to 0.24) | .28 | .013 |
Graphs of predicted values for 12-week pain intensity outcome based on race and Baseline Pain Catastrophizing Scale (PCS) score and on race and Baseline PSEQ score. The blue lines represent Black participants, and the red lines represents non-Hispanic White participants. (A) The regression equations to predict 12-week pain intensity from the baseline PCS were 1.80 + 0.14 × PCS for Black participants and 3.22 + 0.006 × PCS for non-Hispanic White participants. (B) The regression equations to predict 12-week pain intensity from the baseline PSEQ were 6.56–0.04 × PSEQ for Black participants and 6.76–0.07 × PSEQ for non-Hispanic White participants.
Role of the Funding Source
The funders played no role in the design, conduct, or reporting of this study.
Discussion
Studies consistently find Black persons report higher levels of pain intensity than White persons with similar clinical conditions.18,19,23,46 The findings of this secondary analysis of persons receiving physical therapy for chronic spinal pain also identified a significant baseline difference in pain intensity between Black participants and non-Hispanic White participants. This difference was not evident in baseline physical function. Baseline differences by race were also present for 2 psychosocial variables known to impact the pain experience: pain catastrophizing and pain self-efficacy. Black participants reported higher baseline levels of catastrophizing and lower pain self-efficacy than non-Hispanic White participants. As hypothesized, race-based differences in pain intensity at baseline were mediated by pain catastrophizing and self-efficacy. We also found that race moderated the relationship between baseline pain catastrophizing and self-efficacy and pain intensity outcomes after 12 weeks. These results suggest that the impact of pain catastrophizing and self-efficacy varies by race among persons receiving physical therapy for chronic spinal pain.
Race-based differences in pain intensity have been identified across a range of pain conditions,47,48 but psychosocial risk and resiliency factors outlined in the FAM of chronic pain have been examined less frequently. Our finding of higher baseline pain catastrophizing as measured by the PCS among Black participants replicates results from studies involving persons with chronic knee pain23 and low back pain.48 Although pain catastrophizing is predictive of a number of pain outcomes,7 precisely what is measured by the PCS and how it ought to inform our understanding of the development and persistence of chronic pain is a matter of some controversy.49 Traditionally, pain catastrophizing is conceived as a negative cognitive process characterized by rumination, magnification, and helplessness around the pain experience.39 From this perspective, pain catastrophizing has been described as a passive coping strategy that exaggerates the threat value of actual or anticipated pain50 and persons who catastrophize may be stigmatized by providers and viewed as less credible in describing their personal pain experience.51
An alternative perspective views pain catastrophizing as a communication strategy to convey distress and solicit support which can vary based on an individual’s background or culture.52,53 The recognition of pain catastrophizing as a communication strategy may be especially helpful when considering racialized differences. A recent study by Ziadni and colleagues54 reported strong correlations between pain catastrophizing and perceptions of injustice and discrimination in a racially mixed cohort of participants with chronic low back pain. The authors further reported that while pain catastrophizing was correlated with pain intensity scores and pain-related outcomes; injustice appraisals, but not pain catastrophizing, mediated race-based differences in models including all predictors.54 Similarly, Losin and colleagues55 reported that higher pain ratings in response to experimental pain in Black participants was mediated by experience of discrimination. These findings suggest that the nature and impact of pain catastrophizing differs for Black persons and non-Hispanic White persons with chronic pain,23,55 and for Black persons may reflect systemic disparities in the management of pain in health care systems, which have historically undervalued and undertreated pain in Black patients.56,57
Few studies have examined racialized differences in factors associated with resiliency in persons with chronic pain, but race-based differences in how resilience factors impact the pain experience have been identified. Studies comparing resilience factors including gratitude in daily life, optimism, and positive affect have reported significant moderating effects on physical function and pain with movement in patients with chronic low back or knee pain.25,26 For example, higher levels of gratitude were found to be associated with greater performance on a physical function test in non-Hispanic White persons with chronic back pain, but not so for Black participants.25 In a group of persons with chronic knee pain, Black participants reporting low general optimism had much higher levels of self-reported pain with movement relative to Black participants with high optimism, while optimism had no effect on self-reported pain among non-Hispanic White participants.26
Our results examining pain self-efficacy as a resiliency factor are similar to prior studies in identifying race as a moderating factor. Pain self-efficacy reflects a person’s confidence in their ability to persist in daily activities despite pain12 and is associated with better pain outcomes for persons with chronic pain.58–60 We found the expected protective effect of higher baseline pain self-efficacy on pain outcomes for non-Hispanic White participants in our sample, but the protective effect was lessened for Black participants (Fig. 3B). The particular nature of the moderating effect of race in studies examining resilience factors has varied. In our study, the anticipated protective effect for the resiliency factor (pain self-efficacy) was dampened for Black participants; while prior studies have been mixed on whether protection was more or less based on the presence of a resilience factor.25,26 Differences across studies may reflect the distinctive nature of the particular resiliency factor being examined (pain self-efficacy versus optimism, gratitude, positive affect, etc.). Our design also differed from previous studies by evaluating the moderating effects of race on longitudinal pain outcomes instead of using a cross-sectional design.
The reason for our finding of a reduced protective effect for higher baseline pain self-efficacy among Black participants is unclear. Prior studies suggest that Black persons with chronic pain are more likely than non-Hispanic White persons to have an external locus of control and use passive strategies for pain management (eg, denial, hoping, and praying).20,61 It is possible that the concept of pain self-efficacy, or confidence in one’s ability to persist in activities despite pain, is perceived differently by Black respondents and non-Hispanic White respondents. Further research is needed to better understand possible difference by race and other factors in the meaning and role of pain self-efficacy for persons with chronic pain.
Overall, the results of this study are consistent with a body of literature finding racialized differences in pain intensity reporting. Many factors have been proposed to explain higher pain intensity reports by Black respondents in clinical studies.48 While older explanations suggested differences in nociceptive pain processing by race,62,63 more recent evidence supports a more comprehensive biopsychosocial-cultural perspective understanding of pain intensity ratings as reflecting an intrapersonal communicative response informed by prior experiences with bias, discrimination, injustice, and the consequences of these experiences.54,55 Our study supports this more comprehensive perspective in a group of participants receiving physical therapy for chronic spinal pain in typical outpatient clinical settings. While additional research is needed to understand the complex relationships among race, pain, and physical therapy outcomes, our findings support the need to consider culturally tailored pain treatments.
There are several important limitations to consider in interpreting the findings from this study. We report secondary analyses from a clinical trial that were not preplanned. Participants were recruited from one geographic area of the United States and may not generalize to all regions. Our sample included <19% Black participants recruited from a geographical area with a substantially larger Black population. However, the proportion of Black participants is similar to that of the patient population of the clinics included. Grouping Black persons and non-Hispanic White persons into a single category may obscure important intraracial differences based on generational status, country of origin, recent immigration status, etc.64 We did not collect measures of constructs that may be important in understanding race-based differences in pain experience such as experiences of bias, injustice, or discrimination in health care settings or distrust of health care providers.54,65
Conclusion
The findings of this study contribute to a growing body of research identifying racialized differences in the pain experience and specifically on the role of pain catastrophizing and self-efficacy based on race. Our study highlights the need for additional consideration and research on how factors identified in the literature as predictive of outcomes for persons with spinal pain vary based on race, ethnicity, cultural background, and experiences with the health care system.
Improving the understanding of racialized differences in the pain experience may ultimately lead to strategies that can reduce well-established disparities in pain management.
Author Contributions
Elizabeth Lane (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing), Christopher Barnes (Conceptualization, Writing—original draft, Writing—review & editing), and Julie Fritz (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing).
Ethics Approval
This study was approved by the University of Utah Institutional Review Board.
Funding
The original trial was funded by a grant from the Orthopedic Academy of the American Physical Therapy Association (51005577).
Clinical Trial Registration
The trial was registered prior to initiating enrollment (Clinicaltrials.gov NCT03168165).
Disclosures and Presentations
The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest. Elizabeth Lane received a consulting fee or honorarium by MedBridge Institute of Advanced Manual Therapy for continuing education course instruction and creation.
Comments