Prospective Study of Pain Outcomes Associated with Breast Surgery in Women with Nonhereditary Breast Cancer

Abstract

Objective

To assess pain severity and interference with life in women after different types of breast cancer surgery as well as the demographic, treatment-related and psychosocial variables associated with these pain outcomes.

Summary Background Data

Data are conflicting regarding pain outcomes and quality of life among women who undergo different types of breast surgery.

Methods

Women with nonhereditary breast cancer completed the Brief Pain Inventory before surgery and at 1, 6, 12, and 18 months post-surgery. We assessed associations between pain outcomes and contralateral prophylactic mastectomy (CPM) status and mastectomy status using multivariable repeated measures models. We assessed associations between pain outcome and quality of life (QOL) and decision satisfaction.

Results

Of 288 women (mean age 56 years, 58% non-Hispanic white), 50 had CPM, 163 had unilateral mastectomy, and 75 had breast conserving surgery (BCS). Mean pain severity scores were higher at one (2.78 versus 1.9, p=.016) and 6 months (2.79 versus 1.96, p=.031) post-surgery in women who had CPM versus those who did not, but there was no difference at 12 and 18 months. Comparing mastectomy versus BCS, pain severity was higher at 1 and 12 months. There was a significant interaction between pain severity and time point for CPM (p=.006), but not mastectomy status (p=.069). Regardless of surgery type, Black women had higher pain severity (p=.004) than white women. Higher pain interference was associated with lower QOL (p<.001) and lower decision satisfaction (p=.034).

Conclusions

Providers should counsel women considering mastectomy about the potential for greater acute pain and its impact on overall well-being. Racial/ethnic disparities in pain exist and influence pain management in breast surgical patients.

Mini Abstract:

We examined pain severity and interference with life in women after breast cancer surgery as well as demographic, treatment-related and psychosocial variables associated with these pain outcomes. Women who had contralateral prophylactic mastectomy (CPM) reported worse pain outcomes compared to those who did not have CPM at 1 and 6 months after surgery. Both CPM and unilateral mastectomy were associated with worse pain outcomes compared to breast conserving surgery at 1 and 12 months after surgery. Black women had worse pain severity than white women regardless of surgical type.

INTRODUCTION

The American Cancer Society estimates 281,550 women will be diagnosed with breast cancer in 2021, making it the most common cancer among women in the US.¹ While the majority of women with early stage breast cancer receive breast-conserving surgery (BCS) or unilateral mastectomy (ULM),² an increasing number of patients elect to have a contralateral prophylactic mastectomy (CPM) to reduce cancer worry³ and minimize contralateral breast cancer risk.⁴ The rate of CPM in the US has increased by 15% every year between 1998 and 2008, and is even higher among women with in situ or stage I disease who are often ideal candidates for BCS.^3–5

Breast surgery has known risks that include acute pain that resolves with surgical wound healing and chronic pain⁶ that may persist for up to five years after surgery.^{7, 8} The prevailing notion is that more extensive surgery begets more pain, however, there are conflicting data whether patients who undergo ULM report more pain than those who undergo BCS.^{6, 9, 10} Pain after breast surgery has also been shown to limit a patient’s ability to perform activities of daily living (ADLs) and to negatively impact quality of life (QOL).^{7, 9, 10} We previously demonstrated that patients who have CPM have worse QOL up to 18 months after surgery compared to women who have ULM or BCS.¹¹ However, the impact of pain outcomes among women undergoing breast surgery on QOL and decision satisfaction is not known.

Studies show pain severity and duration are influenced by certain patient characteristics that are relevant to women undergoing breast surgery. For example, Black patients report higher pain intensity¹² and duration¹³ compared to patients of other races.¹⁴ Other characteristics including younger age,⁶ higher body mass index (BMI),⁶ depression, and pre-operative anxiety¹⁵ correlate with higher pain severity among women undergoing breast surgery.

The objective of this study is to prospectively examine the associations between type of breast surgery and pain severity and interference among an ethnically diverse cohort of women with nonhereditary breast cancer. We also examined the impact of these pain outcomes on QOL and satisfaction with the surgical treatment decision.

METHODS

Study Population

The study population was previously described in detail by Parker, et al.¹¹ Study participants were women diagnosed with breast cancer and treated at The University of Texas MD Anderson Cancer Center and Kelsey-Seybold Clinic between March 2014 and December 2015. Inclusion criteria were: 1) newly diagnosed ductal carcinoma in situ or stage I to III unilateral breast cancer, 2) age 18 years or older, 3) ability to speak, read, and write in English. Exclusion criteria were: 1) prior history of breast cancer or prophylactic mastectomy, 2) positive test for a germline mutation that increased their risk of hereditary breast cancer (e.g. BRCA1, BRCA2), 3) referral to genetic counseling by their oncology provider for the presence of a strong family history of cancer.

Recruitment and Study Procedures

Participants were recruited after a confirmed breast cancer at their initial surgery appointment or shortly thereafter, and informed consent was obtained. They completed patient-reported outcome measures including the Brief Pain Inventory (BPI),¹⁶ the Functional Assessment of Cancer Therapy-Breast (FACT-B) (version 4),^{17, 18} and the Satisfaction With Decision Scale¹⁹ at the time of enrollment and approximately 1, 6, 12, and 18 months after surgery.¹¹ Study participants were compensated $20 for finishing each survey. This study was approved by the institutional review board at each institution.

Measures

Demographic information for each participant was collected including age, race, ethnicity, marital status, and education. Additionally, clinical data was collected including BMI, stage, receipt of chemotherapy and endocrine therapy, surgery type, whether reconstruction was performed, and estrogen receptor and progesterone receptor status. The BPI assesses pain severity, impact of pain on daily function and type of pain medications used. Patients rate their pain on a scale of 0–10 (10 being most severe) when at its “worst”, “least”, “average” and “now” (current pain).¹⁶ An average of the “worst”, “least”, “average”, and “now” pain scores was used to represent severity. Pain interference was assessed by asking patients to rate how much their pain interfered with seven daily activities: general activity, walking, work, mood, enjoyment of life, relations with others, and sleep. Interference is scored as a mean of all the interference items, or at least 4 of the seven items if not all of them were completed.²⁰ Higher scores indicate more pain severity or interference. Patients were asked to list all medications they are currently taking for pain. We used the FACT-B to assess quality of life (QOL). FACT-B is a multidimensional scale with possible scores ranging from 0 to 123 that assesses physical, social, emotional, and functional well-being and breast cancer-specific concerns. Higher total scores on the FACT-B indicate better QOL.^{17, 18} The Satisfaction With Decision Scale¹⁹ was used to assess treatment satisfaction. Possible scores range from 5 to 30, and higher scores indicate higher satisfaction with treatment decision.

Data Analysis

We reported frequencies and percentages for categorical variables, and we provided summary statistics (e.g. mean and standard deviation [SD]) for continuous data. We used the chi-squared test and Fisher’s exact test to evaluate the association between categorical variables and surgery type (i.e. CPM, ULM and BCS). We used Wilcoxon’s rank sum test to compare the distributions of continuous variables (such as psychosocial scores at each time point) between the surgery type groups. We categorized the continuous BMI data into 4 groups: < 18.5 (underweight), 18.5 to < 25 (normal), 25 to < 30 (overweight), and ≥ 30 (obese). Univariate and multivariable repeated measures models were fitted using PROC GENMOD to assess the association between pain scores and CPM status (CPM versus no CPM) as well as pain scores and mastectomy status (ULM and CPM versus BCS) over different time points with and without adjusting for the other covariates (age, race, chemotherapy, adjuvant hormone therapy, radiation, BMI, and breast reconstruction status). Pain scores were measured at multiple time points which were treated as a categorical variable in the model. The interaction between CPM status or mastectomy status and time was included in the multivariable model if the interaction was statistically significantly associated with the pain scores. The repeated measures models were based on compound symmetry (CS) covariance structure, since Akaike information Criterion (AIC) and Bayesian information criterion (BIC) values using CS were smaller than those using unstructured covariance structure and auto-regressive covariance structures. Mean and standard error (SE) plots over time for pain severity and pain interference scores were estimated for CPM. All tests are two-sided. P-values without adjusting for multiple comparisons less than 0.05 are considered statistically significant. The BPI severity scores at baseline were significantly different by race, therefore multivariable longitudinal models were performed where the baseline pain score was considered as a covariate instead of using it as part of the dependent data.

We performed descriptive analyses of types of pain medications prescribed by race using chi-squared or Fisher’s exact testing as appropriate. We categorized pain medications self-reported by patients as opioids, non-opioid pain medications, non-pharmacologic pain management, or no pain medications listed. All analyses were conducted using SAS 9.4 (SAS, Cary, NC) and S-Plus 8.0 (TIBCO Software Inc., Palo Alto, CA) software.

RESULTS

Patient demographic and clinical characteristics are described in Table 1. We enrolled 288 women, and the mean age of the patients enrolled was 56 years. The participants were 58% non-Hispanic white and 17% Non-Hispanic black. Fifty women had CPM, 163 had ULM and 75 had BCS. Eighty-six women had breast reconstruction. Response rates at baseline and each post-surgical time point were: 87.5%, 90.4%, 96.7%, 98.9%, and 100%, respectively. Further details regarding study enrollment and response can be found in Supplemental Figure 1.

Table 1.

Demographic and Clinical Characteristics (n=288)

Variable	Total	Surgery Type			P
		CPM	Unilateral Mastectomy	Breast-Conserving Surgery
	N=288 (%)	N=50 (%)	N=75 (%)	N=163(%)
Mean age at diagnosis, years (SD)	56.0 (11.9)	49.8 (11.4)	54.7 (12.7)	58.5 (10.8)	<.001
Race					.040
Non-Hispanic White	167 (58.0)	22 (44.0)	45 (60.0)	100 (61.3)
Non-Hispanic Black	49 (17.0)	7 (14.0)	11 (14.7)	31 (19.0)
Hispanic	49 (17.0)	16 (32.0)	10 (13.3)	23 (14.1)
Other	23 (8.0)	5 (10.0)	9 (12.0)	9 (5.5)
Education
Less than High School	12 (4.2)	1 (2.0)	5 (6.7)	6 (3.7)	.072
High school/some college	131 (45.5)	25 (50.0)	24 (32.0)	82 (50.3)
College graduate/post college	145 (50.3)	24 (48.0)	46 (61.3)	75 (46)
Employment					.123
Employed	165 (57.3)	33 (66.0)	47 (62.7)	85 (52.1)
Unemployed	61 (21.2)	11 (22.0)	16 (21.3)	34 (20.9)
Retired/Disabled	62 (21.5)	6 (12.0)	12 (16.0)	44 (27.0)
Marital Status					.505
Married/Living with partner	216 (75.0)	39 (78.0)	58 (77.3)	119 (73.3)
Single/Divorced	71 (24.7)	11 (22.0)	16 (21.3)	44 (27.0)
Missing	1 (0.3)	-	1 (1.3)	-
Annual Income					.601
≤30,000	54 (18.8)	12 (24.0)	12 (16.0)	30 (18.4)
30,000–75,000	96 (33.3)	14 (28.0)	27 (36.0)	55 (33.7)
>75,000	119 (41.3)	18 (36.0)	32 (42.7)	69 (42.3)
Missing	19 (6.6)	6 (12.0)	4 (5.3)	9 (5.5)
Family history of cancer					.283
Yes	72 (25.0)	14 (28.0)	21 (28.0)	37 (22.7)
No	215 (74.7)	35 (70.0)	54 (72.0)	126 (77.3)
Missing	1 (0.3)	1 (2.0)	-	-
Stage					<.001
0	55 (19.1)	12 (24.0)	14 (18.7)	29 (17.8)
I	107 (37.2)	19 (38.0)	13 (17.3)	75 (46.0)
II	106 (36.8)	14 (28.0)	39 (52.0)	53 (32.5)
III	20 (6.9)	5 (10.0)	9 (12.0)	6 (3.7)
Hormone receptor status					.183
ER and PR negative	52 (18.1)	13 (26.0)	15 (20.0)	24 (14.7)
ER and/or PR positive	234 (81.3)	36 (72.0)	60 (80.0)	138 (84.7)
Missing	2 (0.7)	1 (2.0)	-	1 (0.6)
Chemotherapy					.008
Yes	142 (49.3)	27 (54.0)	47 (62.7)	68 (41.7)
No	146 (50.7)	23 (46.0)	28 (37.3)	95 (58.3)
Breast reconstruction					<.001
Yes	86 (29.9)	41 (82.0)	41 (54.7)	4 (2.5)
No	202 (70.1)	9 (18.0)	34 (45.3)	159 (97.5)

Abbreviations: CPM, contralateral prophylactic mastectomy; ER, estrogen receptor; PR, progesterone receptor; SD, standard deviation.

Pain Severity

The mean and median pain severity scores were higher at 1 month (mean 2.78, median 2.38 versus mean 1.90, median 1.25; p=.016 for median) and 6 months (mean 2.79, median 2.00 versus mean 1.96, median 1.00; p=.031 for median) after surgery in women who had CPM compared to those who did not. At 12 months following surgery, the mean and median pain severity scores were not significantly different between women in the two groups (Figure 1).

Figure 1.

Mean +/− Standard Error in Pain Severity by CPM Group

In the univariable longitudinal model there was a significant interaction between time point and CPM status for pain severity (p=.005). These results suggest that CPM patients had higher pain severity, but the severity decreased over time and did not differ from women who did not have CPM at 12 months and beyond. Women who had mastectomies reported higher mean and median pain severity compared to those who had BCS at one (p=.004 for median) and 12 months (p=.034 for median) after surgery (Supplemental Figure 2). BMI, adjuvant radiation therapy, and adjuvant hormone therapy were not significantly associated with pain severity in the univariate longitudinal analyses for CPM or mastectomy status.

In the multivariable longitudinal model, the interaction between time point and CPM status remained significant for pain severity after adjusting for other risk factors (p=.006) (Table 2). However, the interaction between time point and mastectomy status did not remain significant for pain severity (p=.069) (Table 3). This means that pain severity varied significantly over time by CPM status, though not by mastectomy status. Black women had higher pain severity (estimated coefficient of association 0.96, SE=0.33; p=.004) compared to white women regardless of the type of breast surgery performed.

Table 2.

Multivariable Longitudinal Analyses Using Mixed Repeated Measures Models to Evaluate the Associations Between CPM Status and Pain Outcomes

Variable	Pain Severity			Pain Interference
	Estimated Coefficient	Standard Error	P	Estimated Coefficient	Standard Error	P
CPM Yes vs. No	0.10	0.31	.759	0.28	0.34	.410
Time Point x CPM			.006			a
Pain Outcome Baseline	0.50	0.05	<.001	0.54	0.06	<.001
Assessment of Time Effect
T2 (vs. T1)	0.04	0.20	.842	−0.01	0.15	.925
T3 (vs. T1)	−0.48	0.17	.004	0.7	0.15	<.001
T4 (vs. T1)	−0.77	0.20	<.001	−0.67	0.15	<.001
Age at diagnosis (years)	0.01	0.01	.206	−0.01	0.01	.541
Race/Ethnicity
Black, non-Hispanic (vs. White, non-Hispanic)	0.96	0.33	.004	0.61	0.34	.077
Hispanic (vs. White, non-Hispanic)	0.13	0.26	.609	−−0.01	0.31	.966
Other/missing (vs. White, non-Hispanic)	−0.20	0.30	.516	−0.21	0.35	.550
Chemotherapy Yes vs. No	0.56	0.18	.002	0.42	0.20	.039
Reconstruction Yes vs. No	0.40	0.26	.128	0.24	0.25	.342

NOTES. T1=1 month post-surgery; T2=6 months post-surgery; T3=12 months post-surgery; T4=18 months post-surgery. Abbreviations: CPM, contralateral prophylactic mastectomy.

Body mass index, adjuvant radiation therapy, and adjuvant hormone therapy were not significantly associated with pain severity or interference in the longitudinal univariate analysis and were not included in the multivariable analysis.

^aThe interaction between CPM and time point was not statistically significantly associated with pain interference score in the univariate and multivariable models, so the interaction term was excluded from the model.

Table 3.

Multivariable Longitudinal Analyses Using Mixed Repeated Measures Models to Evaluate the Associations Between Mastectomy Status and Pain Outcomes

Variable	Pain Severity			Pain Interference
	Estimated Coefficient	Standard Error	P	Estimated Coefficient	Standard Error	P
Mastectomy Yes vs. No	0.36	0.23	.122	0.07	0.29	.804
Time Point x Mastectomy			.069			.030
Pain Outcome Baseline	0.50	0.05	<.001	0.54	0.06	<.001
Assessment of Time Effect
T2 (vs. T1)	0.04	0.13	.741	−0.04	0.15	.776
T3 (vs. T1)	−0.26	0.13	.038	−0.53	0.14	<.001
T4 (vs. T1)	−0.37	0.14	.007	−0.70	0.15	<.001
Age at diagnosis (years)	0.01	0.01	.204	−0.006	0.01	.531
Race/Ethnicity
Black, non-Hispanic (vs. White, non-Hispanic)	0.97	0.33	.003	0.63	0.34	.067
Hispanic (vs. White, non-Hispanic)	0.13	0.25	.605	0.03	0.30	.914
Other/missing (vs. White, non-Hispanic)	−0.21	0.29	.469	−0.18	0.35	.604
Chemotherapy Yes vs. No	0.48	0.19	.012	0.40	0.21	.058
Reconstruction Yes vs. No	0.20	0.26	.430	0.29	0.32	.363

NOTE. T1=1 month post-surgery; T2=6 months post-surgery; T3=12 months post-surgery; T4=18 months post-surgery.

Body mass index, adjuvant radiation therapy, and adjuvant hormone therapy were not significantly associated with pain severity or interference in the longitudinal univariate analysis and were not included in the multivariable analysis.

The higher mean scores in pain severity among black compared to white women were observed at baseline (2.34 versus 1.18, p=.030) and persisted up to 12 months after surgery. The receipt of chemotherapy was also associated with higher pain severity when stratified by CPM status (p=.002) and mastectomy status (p=.012). Multivariable analyses comparing CPM to no CPM and mastectomy to BCS did not show an association between age, reconstruction status, and pain severity.

Pain Interference

Mean and median pain interference scores at 1 month (mean 3.01, median 2.59 versus mean 1.77, median 1.00; p=.001 for median) and 6 months (mean 2.61, median 1.57 versus mean 1.72, median 0.57; p=.046 for median) after surgery were also significantly higher among women who had CPM compared to those who did not (Figure 2). Similar to pain severity scores, there was no significant difference in mean and median pain interference scores at 12 and 18 months after surgery. Women who had mastectomies reported higher mean and median pain interference compared to those who had BCS at one (p<.001 for median) and 12 months (p=.009 for median) after surgery (Supplemental Figure 3). In the univariable longitudinal model there was a not a significant interaction between time point and CPM status for pain interference (p=.127). BMI, adjuvant radiation therapy, and adjuvant hormone therapy were not significantly associated with pain interference in the univariate longitudinal analyses for CPM or mastectomy status.

Figure 2.

Mean +/− Standard Error in Pain Interference by CPM Group

In the multivariable longitudinal analyses, the interaction between time point and CPM status was not significant for pain interference after adjusting for other risk factors (Table 2). However, the interaction between time point and mastectomy status was significant (p=.030) (Table 3). This means that pain interference varied significantly over time by mastectomy status. Mean differences in pain interference were not statistically significantly different by race. The receipt of chemotherapy was associated with higher pain interference in the analysis by CPM status (p=.039). Multivariable analyses comparing CPM to no CPM and mastectomy to BCS did not show an association between age, reconstruction status and pain interference.

Quality of Life and Decisional Satisfaction by Surgery Type

Adjusting for CPM status and other risk factors in the multivariable analysis, higher pain interference scores were associated with lower QOL scores (p<.001) and lower decision satisfaction scores (p=.034). There was no significant association between pain severity and QOL (p=.979) or decision satisfaction (p=.571) (Table 4).

Table 4.

Multivariable Model to Evaluate the Associations Between Pain Outcomes and Psychosocial Outcomes^a

Variable	Quality of Life			Decisional Satisfaction
	Estimated Coefficient	Standard Error	P	Estimated Coefficient	Standard Error	P
Pain Severity	−0.008	0.31	.979	0.009	0.02	.571
Pain Interference	−3.51	0.33	< .001	−0.04	0.02	.034

Model was adjusted for contralateral prophylactic mastectomy status, time, age, race, stage, and chemotherapy.

Pain Medication Prescriptions by Race

Forty-nine black women and 167 white women provided pain medication information. Black women were more likely to report receiving any type of pain medication compared to white women (60.4% vs. 46.1%), including opioids (45.8% vs 22.4%; chi-squared p=0.02).

DISCUSSION

Advances in multidisciplinary treatment modalities have created more opportunities for BCS. However, North American women have lower rates of BCS and higher rates of CPM compared to Asian and European women.²¹ In our study, 56.5% of women had ULM, 29.9% had BCS, and 17.4% underwent CPM. The prevelence of mastectomy among our study participants is higher than the national rate of mastectomy.²² A retrospective cohort study of over 1.2 million women from the National Cancer Data Base found a number of factors associated with increased likelihood of mastectomy.²² Some of the factors may be present at a higher frequency in our study population compared to the general population, specifically young age, white race, and living in the South. These characteristics may have contributed to the high prevelence of mastectomy in our study population.

Given the substantial numbers of women undergoing CPM, research is needed on their pain experience and impact on their daily lives to better inform the decision making process. We found that women who had CPM experienced higher pain severity in the first year after surgery compared to women who did not have CPM, but the severity decreased over time and did not differ between groups at 12 months and beyond. After adjusting for reconstruction and other clinical factors, women who underwent ULM or CPM experienced higher pain severity than women who had BCS and it appeared to persist for a longer period of time.

The downstream consequences of pain outcomes among breast surgical patients, particularly those with CPM, have not been well investigated. A cross-sectional descriptive study of 23 women who reported chronic pain that began after having ULM or BCS found that patients with widespread pain had lower QOL scores compared with patients with regional pain only.²³ A prospective cohort study of women with non-metastatic breast cancer collected pain outcomes at baseline and 3, 6, 9 and 12 months after ULM or BCS and found that women with musculoskeletal pain had worse QOL compared to those without musculoskeletal pain.²⁴ We examined this question and showed that regardless of surgery type, women who reported higher levels of pain interference were more likely to experience lower QOL and lower decision satisfaction. Physicians should use this information to counseling women experiencing pain interference about its impact on other aspects of the life including overall well-being and surgical decision satisfaction.

Our study also demonstrated that black women experience higher pain severity than white women regardless of surgery type. This is consistent with a systematic review of pain perception and management in black adults which concluded that black patients often report higher levels of acute and chronic pain compared to white patients.¹⁴ A cross-sectional descriptive study of patients receiving treatment for cancer found that black patients had higher pain intensity, pain interference, and pain-related distress than white patients.²⁵ A retrospective study of 119 women who had surgery for breast cancer found that black women and Latinas reported increased rates of pain compared to white women.²⁶ Racial disparities in pain may to be due to differences in the perception, assessment, communication, and management of pain.¹⁹ It is also known that pain expression can be affected by factors beyond physical injury such as psychological or spiritual distress²⁷ and in our study, the race/ethnic disparity in pain was present prior to surgery.

Evidence indicates that physicians may be less likely to prescribe opioids for their black patients,²⁸ and that black patients may be reluctant to use prescription analgesics, though data are conflicting.^{29, 30} Our analysis of the frequency of opioid prescriptions by race was not consistent with this trend and suggests that additional factors such as patient-provider communication or perceptions of pain may have influenced pain management optimization.

Our study has some limitations. It is known that additional surgical procedures such as axillary lymph node dissection is associated with increased morbidity compared to sentinel lymph node biopsy.³¹ Unfortunately, we did not have information regarding the type of lymph node sampling patients or the type of breast reconstruction received. A survey of 445 patients in the United States evaluating chronic pain and opioid use found an association between low health literacy and higher pain intensity.³² We did not collect information on health literacy, and other variables that may affect pain outcomes, such as comorbidities and surgical complications, were not assessed. While a number of studies have noted increased pain severity among Hispanic patients compared to non-Hispanic white patients,³³ we did not see a significant difference in pain outcomes. This may be due to the exclusion of non-English-speaking participants as language barriers have been shown to impede physician-patient communication and impact care. The BPI investigates general pain experience rather than specific sites or the character of pain experienced. Therefore, future studies using additional instruments that capture specific types of pain may provide further insights into the natural history and etiology of pain experienced among women undergoing breast surgery. While some of differences in mean pain and interference scores appear small, prior studies that have used the BPI have reported clinically significant differences in pain outcomes when scores differ by as little as one point or 0.5 times the standard deviation of the mean score.^{34, 35}

In conclusion, this study shows that women considering ULM or elective CPM should be counseled about the potential risk for pain over several months. Interventions that target surgical decision-making should account for the impact of pain interference on decision satisfaction after surgery. Additionally, racial and ethnic disparities in pain outcomes were observed among this cohort of women undergoing breast surgery and further investigation is needed to better understand patient and provider-related factors and healthcare system barriers that may contribute to the inequity.

Supplementary Material

Supplemental Figure 1

Supplemental Figure 1. CONSORT Diagram

Supplemental Figure 2

Supplemental Figure 2. Mean +/− Standard Error in Pain Severity by Surgery Type

Supplemental Figure 3

Supplemental Figure 3. Mean +/− Standard Error in Pain Interference by Surgery Type