Pain, depressive symptoms, and self-efficacy for pain management: examination in African-American women with breast cancer

Abstract

Purpose

African-American women with breast cancer face significant disparities, including high levels of pain. Depressive symptoms and self-efficacy for pain management impact how women with breast cancer manage pain, yet little is known about how these variables relate to pain specifically for African-American women with breast cancer.

Methods

Baseline linear regression analyses were conducted using a sample of women (n = 98) with stage I–III breast cancer identifying as Black or African-American who were part of a larger intervention trial. Linear regressions explored the effect of depressive symptoms on pain (i.e., severity and interference), and the effect of self-efficacy for pain management on pain. Covariates were age (M = 57.22, SD = 10.76), cancer stage (50% = stage 1), and education level (36% = some college).

Results

Participants reported moderate levels of pain severity and interference. Higher depressive symptoms were related to both higher pain severity and interference; (B = 0.06, p < 0.01, 95% CI [0.02,0.09]) and (B = 0.13, p < 0.001, 95% CI [0.09, 0.17]) respectively. Likewise, lower self-efficacy for pain management was also related to both higher pain severity and interference; (B = − 0.04, p < 0.001, 95% CI [− 0.05, − 0.02]) and (B = − 0.06, p < 0.001, 95% CI [− 0.08, − 0.04]) respectively. Women reporting less than a high school diploma endorsed significantly higher pain severity and interference than women reporting some college. Age and cancer stage were not significantly related to pain.

Conclusion

Pain for African-American women with breast cancer may be influenced by depressive symptoms and self-efficacy for pain management, in addition to other important variables. Attending to better assessment and treatment of depressive symptoms and self-efficacy for pain management may improve outcomes.

Introduction

Breast cancer is one of the most common cancers for women, affecting approximately 4 million women in the USA [1]. African-American women with breast cancer have higher rates of mortality and worse survivorship outcomes [2]. African-American women diagnosed with breast cancer are also more likely to experience higher pain levels compared to other groups [3]. Several factors contribute to these persistent and significant disparities.

Pain related to cancer or its treatment affects more than half of women with breast cancer [4] and is consistently related to poor outcomes for survivors [5]. Cancer pain disparities are found across cancers, with African-American adults reporting higher levels of pain, pain-related distress, and lower functioning than their White counterparts [6, 7]. African-American women may carry additional risk for pain above that of non-Hispanic Whites due to breast cancer–specific systemic and health risk factors [3, 8]. Systemic factors such as inequitable access to cancer screenings and timely cancer care often mean that African-American women may be diagnosed at later breast cancer stages [9] or require more aggressive treatments [10], contributing to greater pain [5]. Research highlights that African-American women are more likely than White women to have health risk factors such as being diagnosed at a younger age [9] and with more aggressive cancers (e.g., triple-negative breast cancer [TNBC] [11]) and these factors are associated with high pain levels [12]. African-American women with breast cancer may also experience suboptimal pain management due to complex and interacting provider, patient, and structural-level variables. Providers are more likely to record lower pain ratings and underprescribe pain medications [13]. These prescribing behaviors have been linked to false beliefs about pain tolerance among African-Americans [14], and prejudicial attitudes about addiction or medication-seeking [15]; factors that in turn may influence patients’ concerns about medication use and fears of addiction [16]. Structurally, patients have less access to adequate screening, less access to certain prescribed medications, and lack access to behavioral pain treatments [7, 13, 16]. Despite these known disparities, research is still limited on the pain experience among African-Americans adults with cancer even when examined more broadly. As such, it is critical to add to this literature and better understand the experience of pain for African-American women with breast cancer to optimize treatment and improve outcomes.

Depressive symptoms are also commonly reported among women with breast cancer [17]. Greater depression symptoms predict cancer mortality [18] and are linked with reduced engagement with cancer healthcare [19] which can exacerbate cancer survivorship disparities for African-Americans. While studies show that approximately 22% of African-American breast cancer patients experience some depressive symptoms [20], in the context of pain these rates are increased such that 38% of African-American cancer patients receiving pain care endorse at least moderate depressive symptoms [21]. Cancer pain has been shown to be a consistent predictor of depressive symptoms for African-American cancer patients more broadly [21, 22], but the specific experience of African-American women with breast cancer remains understudied. Yet African-American women with breast cancer may have additional risk for depressive symptoms due to younger age and later stage at time of breast cancer diagnosis [21], in addition to systemic stressors such as mistrust in the medical system [23]. Understanding the particular burden of depressive symptoms for African-American women with breast cancer is thus an important part of cancer pain management [24, 25].

Self-efficacy for pain management, or one’s confidence in her ability to enact particular pain coping strategies, has implications for how women with breast cancer manage pain [26]. In heterogenous samples of cancer patients, self-efficacy for pain management has been shown to help reduce pain and other important pain-related variables (e.g., depression, disability) [27]. African-American women with breast cancer may be particularly likely to have lower levels of self-efficacy for managing cancer-related pa in given the unique stressors (e.g., younger age and later stage at diagnosis, pain-related distress) they face [6]. Emerging evidence suggests that self-efficacy is also potentially important for African-American women with breast cancer. Sheppard and colleagues [28] reported that low self-efficacy for symptom management was associated with increased anxiety and depression for African-American women with breast cancer. Likewise, Watkins and colleagues [29] found lower capacity to cope to be associated with higher psychological distress when African-American women with breast cancer were receiving chemotherapy. Yet, despite the potential for African-American women to carry a higher burden of pain, no study to date has critically examined the role of self-efficacy for pain management in this group.

This study aimed to examine the relationships between pain (i.e., severity, interference), depressive symptoms, and self-efficacy for pain management in a sample of women with breast cancer who identified as Black or African-American (n = 98). This was an exploratory, secondary analysis of baseline data from a larger trial (N = 327) that assessed a cognitive behavioral pain management intervention for women with breast cancer and pain [30]. The first hypothesis was that African-American women who reported higher levels of depressive symptoms would report higher levels of pain (i.e., severity, interference). The second hypothesis was that African-American women who reported lower levels of self-efficacy for pain management would report higher levels of pain (i.e., severity, interference). These relationships were expected to be evident even after controlling for demographic and medical variables (i.e., age, cancer stage, education level) known to be related to pain [31].

Methods

Participants

Women at least 18 years of age were recruited from 2017 to 2020 at Duke University Medical Center and affiliated cancer clinics. Inclusion and exclusion criteria for the parent trial were set to ensure the cognitive behavioral pain management intervention would be appropriate for participants based on previous research [30, 27]. Inclusion criteria included the following: (1) diagnosis of stage I–IIIC breast cancer (initial or recurrence) within the past 2 years; (2) life expectancy ≥ 12 months; and (3) pain severity rating ≥ 5 out of 10 at screening, consistent with moderate-to-severe pain [30]. Exclusion criteria included the following: (1) cognitive impairment; (2) brain metastases; (3) presence of a severe psychiatric condition (e.g., psychotic disorder or episode, suicidal intent) that would contraindicate safe participation; or (4) current or past (< 6 months) engagement in cognitive-behavioral coping skills treatment for cancer pain that would overlap with skills taught in the trial. The parent study was a randomized trial with ethical approval by the Duke University Institutional Review Board (IRB #: Pro00070823) and registered on ClinicalTrials.gov (NCT02791646).

Procedures

Potential study participants were assessed for eligibility by study staff using electronic medical record review and mailed recruitment letters and interested participants engaged in informed consent procedures. Written informed consent for participation in study procedures as well as permission for data publication was obtained from enrolled participants. After enrollment in the parent trial, participants completed an online baseline assessment consisting of self-report questionnaires measuring pain severity, pain interference, depressive symptoms, and self-efficacy for pain management. The current analysis is a secondary analysis of baseline data from the larger clinical trial. Additional trial information and another baseline analysis have been published elsewhere [30, 31].

Measures

Demographic and medical characteristics

Data were collected through electronic medical record review and self-report. Demographic information included the following: age, sex, race and ethnicity, education, and household income. Medical characteristics included breast cancer stage and other cancer specific variables (e.g., initial or recurrent diagnosis, surgeries).

Pain severity and interference

Pain was assessed with the Brief Pain Inventory (BPI) [32]. Four pain severity items asked participants to rate their pain at its worst, least, and average over the past week as well as their current pain. Response options ranged from 0 (no pain) to 10 (worst pain imaginable). Severity items were averaged for a composite score (Cronbach’s α = 0.86). Seven pain interference items asked participants to rate the degree to which, over the past week, pain interfered with daily activities (i.e., general activity, mood, walking ability, normal work [including housework], relations with others, sleep, and enjoyment of life). Response options ranged from 0 (does not interfere) to 10 (completely interferes). Pain interference items were averaged for a composite score (Cronbach’s α = 0.91).

Depressive symptoms

Depressive symptoms were assessed with the 20-item Center for Epidemiological Studies Depression Scale (CES-D) [33]. Participants were asked to rate how frequently they experienced various depressive symptoms (e.g., low mood, anhedonia, lack of appetite, difficulty concentrating) over the past week. Response options ranged from 0 (rarely or none of the time, less than one day) to 3 (all of the time, five to 7 days). Items were summed for a total score (Cronbach’s α = 0.91).

Self-efficacy for pain management

Self-efficacy for pain management was assessed with the five-item Chronic Pain Self-Efficacy Scale [34]. Participants were asked to rate their confidence in their ability to decrease their pain, continue their daily activities, keep pain from interfering with their sleep, make a small-to-moderate reduction in pain using methods other than taking extra medication, and make a large reduction in pain using methods other than taking extra medication. Response options ranged from 10 (very uncertain) to 100 (very certain). The five items were averaged for a composite score (Cronbach’s α = 0.86).

Statistical methods

Analyses were conducted using SPSS (version 27). Preliminary descriptive analyses included identification of outliers and examination of main study variable distributions for kurtosis, skewness, and assumptions of normality for multivariate data. No outliers were identified in boxplot analyses where outliers were defined as observations greater than 1.5 inter-quartile range (IQR) but less than 3 IQR. All main study variables exhibited normal distributions, with kurtosis and skewness values within − 2 and + 2 [35]. Assumptions for linear regressions were met [27]: (1) variables were measured at the continuous level; (2) variables exhibited a linear relationship; (3) variable distributions did not have outliers; (4) independence of observations was observed based on Durbin-Watson values between 1.5 and 2.5 for all linear regressions; (5) homoscedasticity was demonstrated via scatterplots of standardized predicted values; and (6) residuals of the regression line were normally distributed per normal P-P plot of regression standardized residuals.

Bivariate correlations were run for all main study variables: pain severity, pain interference, depressive symptoms, and self-efficacy for pain management. Two linear regression analyses were conducted to examine the association between depressive symptoms and (1) pain severity and (2) pain interference. Likewise, two linear regression analyses were conducted to examine the association between self-efficacy for pain management and (1) pain severity and (2) pain interference. Regressions were run both unadjusted and adjusted for age, cancer stage, and education level. Cancer stage was categorized to set stage 1 disease as the referent group. Education level was also categorized to set “some college” as the referent group. Referent groups were determined based on largest represented cancer stage and education level group among our sample. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Results

Participant characteristics

From the parent sample of 327 women with breast cancer, 98 participants (30.0%) self-identified as Black or African-American. The number of women who identified as Black or African-American enrolled in the parent study was representative of the population served by the treating cancer clinics based on clinic catchment area statistics. The current study used this sub-sample of women identifying as Black or African-American for the analyses. The average age of women in this sub-sample was 57.22 (SD = 10.87) years (Table 1). Approximately one-third (36%) of participants reported “at least some college”; nearly half (44%) of participants reported household income of up to $39,000/year. For nearly all women (97%), this was their first breast cancer diagnosis (Table 2). Forty-nine participants (50%) had stage 1 disease. Additional demographic and medical characteristics are reported in Tables 1 and 2, respectively. Participant reported pain severity (M = 4.59; SD = 1.90) and interference (M = 4.52; SD = 2.59) were in the moderate range (Table 3). Additional variable means and correlations are reported in Table 3.

Table 1.

Demographic characteristics (N = 98)

	N (%)	M (SD)
Age (years)		57.22 (10.78)
Race
Black or African-American	98 (100%)
Ethnicity
Non-Hispanic	98 (100%)
Education
Less than high school diploma	5 (5.1%)
High school diploma	16 (16.3%)
Some college	35 (35.7%)
Bachelor’s degree	24 (24.5%)
Graduate degree	18 (18.4%)
Income
Less than $10,000	7 (7.2%)
$10,000 to $19,999	14 (14.4%)
$20,000 to $39,999	22 (22.7%)
$40,000 to $59,999	28 (28.9%)
$60,000 to $100,000	18 (18.6%)
More than $100,000	8 (8.2%)

M, mean; SD, standard deviation

Table 2.

Medical characteristics (N = 98)

	N (%)	M (SD)
Cancer diagnosis
First/Initial	95 (96.9%)
Recurrence	3 (3.1%)
Months since diagnosis		10.39 (6.42)
Stage
I	49 (50.0%)
II	39 (39.8%)
III	10 (10.2%)
Mastectomy (one breast only)
Yes	18 (18.8%)
No	78 (81.3%)
Mastectomy (both breasts)
Yes	12 (12.5%)
No	84 (87.5%)
Breast-conserving surgery
Yes	57 (59.4%)
No	39 (40.6%)
Lymph node removal
Yes	15 (15.6%)
No	81 (84.4%)
Reconstructive surgery
Yes	13 (13.5%)
No	83 (86.5%)
Use of antidepressant medication
Yes	27 (27.6%)
No	71 (72.4%)

M, mean; SD, standard deviation; Breast-conserving surgery—lumpectomy, quadrantectomy, partial mastectomy, segmental mastectomy

Table 3.

Means (M), standard deviations (SD), and correlation matrix for main study variables (N = 98)

Variable	Pain severity	Pain interference	Depressive symptoms	Pain self-efficacy
M (SD)	4.59 (1.90)	4.52 (2.59)	17.20 (10.91)	60.14 (22.24)
Pain severity	1	-		-
Pain interference	0.73**	1		-
Depressive symptoms	0.36**	0.59**	1
Pain self-efficacy	− 0.50**	− 0.55*	− 0.47**	1

M, mean; SD, standard deviation;

^*p < .05;

^**p < .01

Relationship between depressive symptoms and pain severity and pain interference

Depressive symptoms were significantly related to both pain severity (B = 0.06, p < 0.001, 95% CI [0.03, 1.00]) and pain interference (B = 0.14, p < 0.001, 95% CI [0.10, 0.18]), such that higher depressive symptoms were associated with higher pain severity and higher pain interference.

Depressive symptoms remained significantly related to both pain severity (B = 0.06, p < 0.01, 95% CI [0.02, 0.09]) and pain interference (B = 0.13, p < 0.001, 95% CI [0.09, 0.17]) in the linear regression models adjusted for age, cancer stage, and education. After adjusting for the covariates, depressive symptoms explained an additional 9.0% (p < 0.01) and 27.6% (p < 0.001) of the variance in pain severity and pain interference, respectively. The control variable comparing “less than high school diploma” to “some college” was also a significant predictor of both pain severity (B = 2.27, p < 0.01, 95% CI [0.64, 3.90]) and pain interference (B = 2.34, p < 0.05, 95% CI [0.39, 4.29]) within this adjusted model. These results indicate that, on average, women reporting “less than high school diploma” endorsed pain severity and pain interference 2.27 and 2.34 points higher than those reporting “some college,” respectively. Age and cancer stage were not significantly related to pain. Significant results for the adjusted models only are presented in Table 4.

Table 4.

Predictors of pain severity and pain interference

	Pain severity			Pain interference
	B	P	95% CI (lower, upper bounds)	B	P	95% CI (lower, upper bounds)
Adjusted model:
Depressive symptoms	0.06	< .001	0.02, 0.09	0.13	< 0.001	0.09, 0.17
Education level	2.27	< 0.01	0.64, 3.90	2.34	< 0.05	0.39, 4.29
Adjusted model:
Self-efficacy	− 0.04	< 0.001	− 0.05, − 0.02	− 0.06	< 0.001	− 0.08, − 0.04
Education level	2.24	< 0.01	0.69, 3.79	2.30	< 0.05	0.26, 4.34

Regression analyses shown here were adjusted for age, cancer stage, and education level. Only significant predictors shown. Education level—less than high school diploma vs. some college

Relationship between self-efficacy for pain management and pain severity and pain interference

Self-efficacy for pain management was significantly related to both pain severity (B = − 0.04, p < 0.001, 95% CI [− 0.06, − 0.03]) and pain interference (B = − 0.06, p < 0.001, 95% CI [− 0.08, − 0.04]), such that lower self-efficacy for pain management was associated with higher pain severity and higher pain interference.

Self-efficacy for pain management remained significantly related to both pain severity (B = − 0.04, p < 0.001, 95% CI [− 0.05, − 0.02]) and pain interference (B = − 0.06, p < 0.001, 95% CI [− 0.08, − 0.04]) in the linear regression models adjusted for age, cancer stage, and education. After adjusting for the covariates, pain self-efficacy explained an additional 15.6% (p < 0.001) and 21.1% (p < 0.001) of the variance in pain severity and pain interference, respectively. The control variable comparing “less than high school diploma” to “some college” was also a significant predictor of both pain severity (B = 2.24, p < 0.01, 95% CI [0.69, 3.79]) and pain interference (B = 2.30, p < 0.05, 95% CI [0.26, 4.34]) within this adjusted model. These results indicate that, on average, women reporting “less than high school diploma” endorsed pain severity and pain interference 2.24 and 2.30 points higher than those reporting “some college,” respectively. Age and cancer stage were not significantly related to pain. Significant results for the adjusted models only are presented in Table 4.

Discussion

The purpose of this study was to examine the relationships between pain, depressive symptoms, and self-efficacy for pain management among women who identify as Black or African-American. Higher levels of depressive symptoms were significantly associated with both greater pain severity and pain interference. Lower self-efficacy for pain management was significantly associated with higher pain severity and pain interference. Lower education significantly predicted both greater pain severity and interference in all regression models; age and cancer stage were not independently related to pain. Depressive symptoms and self-efficacy for pain management remained significant predictors of both pain severity and interference after adjusting for the contribution of education level.

The current study is the first to demonstrate that depressive symptoms and self-efficacy for pain management significantly predict pain severity and pain interference, after adjusting for relevant covariates of age, cancer stage, and education level, in a sample of Black or African-American women with breast cancer. Literature suggests that this population is particularly vulnerable to high levels of pain due to the influence of numerous health risk and systemic factors. However, this is one of the first studies to examine how psychosocial variables (e.g., depressive symptoms and self-efficacy for pain management) affect the pain experience in this population.

Greater depressive symptoms predicted both higher pain severity and pain interference. Notably, 27% of the variance in pain interference was explained by depressive symptoms, as opposed to only 9% of the variance in pain severity. This builds upon previous study findings that showed greater depressive symptoms were predictive of greater pain interference for African-American cancer patients [21]. Depressive symptoms such as reduced motivation and persistent negative attitudes about the future may exacerbate systemic vulnerabilities of reduced access to and engagement in care for African-American women. When patients experience depressive symptoms alongside greater pain severity and even greater pain interference, the combined burden of these symptoms may reduce patients’ willingness to seek or persist in accessing adequate cancer care, ultimately worsening outcomes. Indeed, higher overall symptom burden is associated with reduced engagement with breast cancer treatments [36]. These data lend support to previous research calling for improved screening and treatment of depressive symptoms among African-American cancer patients [21], and extend it to women with breast cancer as a means of improving pain.

While previous studies have examined self-efficacy in relation to anxiety and depressive symptoms for African-American women with breast cancer, this is the first study to directly assess the impact of self-efficacy for pain management for this population. We found that lower pain self-efficacy predicted both higher pain severity and higher pain interference. This is consistent with previous work examining self-efficacy for managing pain in broader samples of breast cancer patients [27, 37, 38] and examinations of perceived control over pain for cancer patients identifying as African-American [6, 39], providing support for the relevance of self-efficacy for pain management for African-American women coping with breast cancer and pain. Breast cancer diagnosis and treatment can challenge many women’s ability to cope. African-American women may be particularly vulnerable to lower self-efficacy for pain management due to health risk and systemic factors. Lack of access to adequate financial and other resources during cancer care, structural racism leading to poor responsiveness from the medical community, and vulnerabilities for more aggressive cancers requiring more intensive treatment may further burden African-American women coping with breast cancer. This increased burden may reduce self-efficacy for pain management, leading to greater pain problems because women may not actively seek out and effectively trouble-shoot known ways to improve pain (e.g., through pleasant activities, relaxation, exercise).

Pain self-efficacy and depressive symptoms remaining significant predictors of pain after adjusting for the contribution of education level are informative. While education level cannot be changed, self-efficacy for pain management and depressive symptoms can be targeted in evidence-based behavioral interventions (e.g., Pain Coping Skills Training [40]) that teach effective pain coping strategies. In circumstances where breast cancer patients have lower levels of education, targeting pain self-efficacy and depressive symptoms may serve as an effective means of reducing pain. This is particularly important because in this study, participants with the lowest education (less than high school diploma) also reported pain severity and interference over 2 points higher than those with high levels of education (some college). Indeed, promising research shows that low-literacy adaptations of self-efficacy skills training can improve cancer pain outcomes in medically underserved communities [38].

Limitations and future directions

The current study has limitations. First, this was a cross-sectional study of baseline levels of theoretically important variables. Future research should include larger, longitudinal studies to best characterize these relationships over time. Additionally, this study was conducted in the context of a large academic medical center in a mostly-urban area. Future research assessing these variables should move beyond those served in a large academic medical center and incorporate those (i.e., in rural areas) who may not have access to such care. Another limitation was the fact that this study did not assess other possible contributors to pain disparities, such as patients’ experience of racism and medical mistrust [7, 28]. Future research should include measurement of these experiences, along with studies that assist healthcare systems in overcoming the institutional inequities contributing to cancer and pain disparities [41].

Conclusions

If not adequately addressed, cancer-related pain can become persistent pain, leading to poor overall functioning [42]. These results strengthen previous studies outlining the critical importance of adequate screening and treatment of depressive symptoms among African-American cancer patients. Furthermore, the results add to previous study on pain self-efficacy and specifically highlight its potential importance as a means of treating cancer-related pain and preventing persistent pain among African-American women.

Cognitive behavioral interventions promoting self-efficacy have been shown to reduce depressive symptoms, pain severity, and pain interference for breast cancer patients more broadly [27]; however, future research should specifically examine and potentially adapt these interventions to best support African-American women with breast cancer and pain. Spiritual practices [43] and interpersonal support [44] also show benefits for pain outcomes for African-American cancer patients, and could play important roles in self-efficacy-informed interventions aimed at reducing pain for African-American women with breast cancer.