The effect of ethnicity on immediate reconstruction rates after mastectomy for breast cancer

Authors


  • Presented in part at the 57th annual meeting of the Society of Surgical Oncology, New York, New York, March 18–21, 2004.

Abstract

BACKGROUND

Multiple factors may influence whether patients undergo immediate breast reconstruction along with mastectomy for breast cancer. The authors investigated whether ethnicity was an independent predictor of immediate breast reconstruction.

METHODS

The authors identified 1004 patients who underwent mastectomy for breast cancer during the period 2001–2002. The rates of immediate reconstruction among different ethnicities were evaluated using the chi-square test. Logistic regression was used to adjust for covariates, including age and disease stage. Medical records were analyzed to identify factors that influenced each patient's decision for or against immediate breast reconstruction.

RESULTS

Three hundred seventy-six women (37.5%) underwent immediate breast reconstruction: This included 20.2% of African-American women, compared with 40.0% of white women, 42.0% of Hispanic women, 42.2% of Asian women, and 10.0% of Middle Eastern women (P < 0.001). The unadjusted odds ratio (OR) for immediate reconstruction for African-Americans versus whites was 0.38 (95% confidence interval [95% CI], 0.23–0.63; P < 0.001). After multivariate analysis, this disparity persisted, with an adjusted OR of 0.34 (95% CI, 0.18–0.62; P = 0.001). Asian women had lower rates of immediate reconstruction compared with white women (adjusted OR, 0.50; 95% CI, 0.24–1.04; P = 0.06). Hispanic women did not have immediate reconstruction rates that differed significantly from white women. Middle Eastern women had lower rates of immediate reconstruction compared with white women (adjusted OR, 0.08; 95% CI, 0.02–0.38; P = 0.002), but they had a corresponding increase in the rate of delayed reconstruction. In a stepwise analysis of the decision pathway to immediate reconstruction, it was found that African-American women were less likely to be offered referrals for reconstruction, were less likely to accept offered referrals, were less likely to be offered reconstruction, and were less likely to elect reconstruction if it was offered.

CONCLUSIONS

African-American women underwent immediate breast reconstruction at significantly lower rates compared with white women, Hispanic women, and Asian women. After adjusting for covariates, including age and disease stage, African-American women and Asian women had lower rates of reconstruction compared with white women. The factors that contribute to these differences warrant further study. Cancer 2004. © 2004 American Cancer Society.

Patients who are diagnosed with breast cancer are confronted with an increasing array of information. Within a limited time, patients are required to make important decisions regarding their treatment. Once they have decided to undergo mastectomy, patients must determine whether to undergo immediate reconstruction.

The primary benefits of immediate reconstruction at the time of mastectomy include improved psychological well being and improved cosmetic results. It has been found that patients who undergo mastectomy for breast cancer experience more postoperative psychosocial distress compared with patients who undergo other surgical procedures.1 Investigators have found that various types of immediate breast reconstruction pose a minimal risk of delaying treatment.2, 3 In addition, it has been shown that breast reconstruction does not interfere with follow-up for recurrent cancer.4 In a randomized clinical trial, patients who underwent immediate reconstruction experienced higher physical self-esteem compared with patients who were offered delayed reconstruction 12 months after mastectomy.5 However, not all patients experience psychological benefits after reconstruction.6

The role of race/ethnicity in breast cancer represents a growing concern.7 Race and related patient factors can affect health care in a complex fashion.8 Low socioeconomic status, which variably is linked to race, is associated with worse outcomes in many diseases, including breast cancer.9 Racial disparities have been described in medical areas as diverse as childhood mortality,10 lung cancer,11 prostate cancer,12 and heart disease.13–15 Disparities in breast cancer care and survival have been documented for decades.16–19 Breast cancer appears to have a lower overall incidence in older African-American women in the United States compared with non-Hispanic white women, but it has an increased incidence in women age ≤ 50 years.20 Breast cancer is diagnosed at higher American Joint Committee on Cancer (AJCC) stages in African-American women compared with in white women.21, 22 Breast cancer mortality is higher in African Americans than in other ethnic groups, and the mortality rate decreased only 1% per year for African-American women in the 1990s, compared with a 2.5% yearly decrease for white women.7, 23 Differences in treatment patterns also have been documented between African-American women and women in other ethnic groups.24, 25 Less information is available currently about differences in breast cancer treatment and outcomes in Hispanic, Asian, and Middle Eastern patients who receive care in the United States.26, 27

Few studies have elucidated the effect of ethnicity on specific physician-patient interactions that may affect treatment and outcome. These include physicians' recommendations to patients and patients' perceptions of the treatment options presented to them.26, 27 For all ethnic groups, decision-making by patients and their clinicians may be affected by both overt and imperceptible cultural differences.28 In this retrospective study, we investigated whether patient race was an independent predictor of immediate breast reconstruction after mastectomy in a single, large-volume cancer center with a diverse patient population and cohesive surgical oncology and plastic surgery departmental practices.

MATERIALS AND METHODS

Database

From a prospective database of patients undergoing breast surgery at our institution, we identified 1023 consecutive women who underwent mastectomy for cancer between January 1, 2001, and December 31, 2002. Nineteen patients had undergone prophylactic mastectomies and were excluded from further analysis, leaving a total of 1004 patients who had undergone mastectomy for in situ or invasive breast carcinoma. Our study was approved by The University of Texas M. D. Anderson Cancer Center Institutional Review Board.

Medical Records

The medical records of all identified patients were reviewed. Discrete patient information was abstracted and recorded in a confidential database for analysis. Body mass index (BMI) was calculated using the standard formula: BMI = weight in kilograms/height in meters.2 Self-identified race was determined from patient intake data and was confirmed from an analysis of the medical record. To assist with evaluation of socioeconomic status, both patient ZIP code and insurance class information were used. The median household income associated with each patient's five-digit ZIP code was obtained from the United States Census.29 Patients who had undergone immediate breast reconstruction were identified. Patients who had undergone plastic surgical referral, evaluation, and/or operation solely for chest wall coverage without breast mound reconstruction were excluded from the reconstruction group. Delayed plastics referral, consultation, and/or operation were identified and noted separately. We analyzed dictated physician notes to assess the patient-physician interaction, paying specific attention to whether the surgical oncologist discussed breast reconstruction and offered a referral for plastic/reconstructive surgery, whether the patient accepted such a referral, whether the plastic surgeon offered breast reconstruction to the patient, and whether the patient underwent breast reconstruction.

Statistical Analysis

The unadjusted rates of breast reconstruction among different self-reported races were evaluated using the chi-square test. Multivariate logistic regression was used to construct a model to predict the odds of immediate reconstruction that included the biologic covariates of AJCC stage, patient age, race, and BMI. Adjusted odds ratios (ORs), 95% confidence intervals (95% CIs), and P values were calculated. An identical model was constructed to predict the odds of immediate or delayed reconstruction within the follow-up period of the study. Secondary regression modeling was done that further incorporated covariates related to socioeconomic status, i.e., insurance type and median income for ZIP code. Age (per year), median income for ZIP code (per $1000), and BMI (per kg/m2) were considered continuous variables for the purposes of the models. A commercially available software package (SPSS version 11.5; SPSS Inc., Chicago, IL) was used to perform statistical analyses.

To analyze the decision-making process involved in immediate breast reconstruction, we divided the pathway toward reconstruction into four simplified steps: 1) whether a surgical oncologist's offer of referral for immediate breast reconstruction was documented, 2) whether such a referral was accepted by the patient, 3) whether a plastic surgeon offered immediate reconstruction to the patient, and 4) whether the patient opted to undergo immediate breast reconstruction. We calculated sequential ORs for African-American, Hispanic, Asian, and Middle Eastern women compared with non-Hispanic white women for a yes at each of the four branch points. These sequential ORs were adjusted for the biologic covariates of age, disease stage, and BMI.

RESULTS

Demographics

The 1004 patients included 718 non-Hispanic white women (72%), 99 African-American/black women (10%), 112 Latina/Hispanic women (11%), 45 Asian women (4%), and 30 Middle Eastern women (3%). Nine hundred twenty patients had addresses in the United States. The 45 Asian women included women of Chinese, Japanese, Filipino, Indian, and Pakistani descent. The majority of the 30 Middle Eastern women were international patients from Arab countries, Iran, and Israel.

The mean patient age (± standard deviation) was 54.6 years ± 12.2 years. White women were the oldest group (55.9 years ± 12.3 years), and Middle Eastern women were the youngest (43.1 years ± 8.8 years). Asians were most likely to present at earlier AJCC stages; 91.1% had Stage 0–IIB disease, compared with 81.2% of white women, 70.7% of African-American women, 76.8% of Hispanic women, and 60.0% of Middle Eastern women). The mean (± standard deviation) ZIP code household income averaged $48,600 ± 20,600 for the entire cohort. The ZIP code income averaged $50,250 for whites, $38,080 for African Americans, $38,530 for Latinas, $64,410 for Asians, and $57,400 for Middle Eastern patients. Demographic information is shown in Table 1.

Table 1. Baseline Patient Demographics
CharacteristicAll patientsRace/ethnicity
WhiteAfrican AmericanHispanicAsianMiddle Eastern
  • SD: standard deviation; AJCC: American Joint Committee on Cancer; BMI: body mass index; HMO: health maintenance organization; PPO: preferred provider organization; VA: Veterans Administration.

  • a

    I.e., patient is responsible for payment.

No. of patients1004718991124530
Mean age ± SD (yrs)54.6 ± 12.255.9 ± 12.353.8 ± 12.152.2 ± 11.548.3 ± 8.643.1 ± 8.8
AJCC stage (%)      
 In situ (Stage 0)17.817.820.216.124.410.0
 Stage I29.731.920.230.426.710.0
 Stage IIA20.320.817.219.624.416.7
 Stage IIB11.610.713.110.715.623.3
 Stage IIIA6.86.011.18.02.213.3
 Stage IIIB9.38.911.110.70.020.0
 Stage IV4.53.97.14.56.76.7
Mean BMI ± SD (kg/m2)27.9 ± 11.027.7 ± 12.331.8 ± 6.927.4 ± 6.523.6 ± 3.327.2 ± 5.1
Mean ± SD income by Zip code (in $1000s)48.6 ± 20.650.3 ± 20.438.1 ± 16.538.5 ± 15.464.4 ± 24.857.4 ± 20.3
Insurance (%)      
 Medicare A/B23.226.625.314.32.20.0
 PPO31.635.021.223.235.610.0
 Blue Cross/Blue Shield19.320.916.215.222.23.3
 HMO9.78.814.19.815.66.7
 Self-paya7.13.52.016.115.663.3
 Free care (indigent)2.81.77.17.12.20.0
 Commercial insurance2.31.93.04.52.20.0
 Bad debt (collections)2.21.06.12.72.216.7
 Medicaid0.90.32.04.50.00.0
 Harris County (indigent)0.60.12.02.70.00.0
 Champus/VA0.40.31.00.02.20.0

Immediate Reconstruction

In the entire cohort, 376 women (37.5%) underwent immediate breast reconstruction: This included 20.2% of African American women, compared with 40.0% of white women, 42.0% of Hispanic women, 42.2% of Asian women, and 10.0% of Middle Eastern women. On chi-square analysis, significant differences were found between African-American and white women (P = 0.001) and between Middle Eastern and white women (P = 0.003), (Fig. 1). These differences remained significant after adjusting for multiple comparisons. The rates of immediate reconstruction for each ethnic group were stratified by AJCC stage (Table 2). Stage for stage, African American women underwent immediate breast reconstruction at lower rates compared with white women. The possibility of postoperative radiation therapy was documented in preoperative dictations for 24.5% of white women, 15.2% of African-American women (chi-square vs. whites: P = 0.04), 27.7% of Hispanic women (P = 0.48), 26.7% of Asian women (P = 0.74), and 63.3% of Middle Eastern women (P < 0.0001). The unadjusted OR for immediate reconstruction for African American versus whites was 0.38 (95% CI, 0.23–0.63; P < 0.001). Unadjusted ORs for immediate breast reconstruction for Hispanic, Asian, and Middle Eastern women were 1.04 (95% CI, 0.69–1.55), 1.09 (95% CI, 0.59–2.0), and 0.17 (95% CI, 0.05–0.55), respectively (Table 3).

Figure 1.

This chart illustrates the rates of immediate breast reconstruction by ethnicity. African-American and Middle Eastern women had significantly lower unadjusted rates of reconstruction compared with white women (chi-square analysis; P values as shown).

Table 2. Reconstruction by American Joint Committee on Cancer Stage for Each Racial/Ethnic Group
StagePercentage undergoing immediate reconstruction
All patientsWhiteAfrican AmericanHispanicAsianMiddle Eastern
  • N/A: not applicable.

  • a

    There were no Asian patients with Stage IIIB disease in the study cohort.

All stages37.540.020.242.042.210.0
In situ (Stage 0)68.971.950.077.863.633.3
Stage I47.348.525.050.058.333.3
Stage IIA35.836.211.850.045.520.0
Stage IIB22.427.315.425.00.00.0
Stage IIIA1.52.30.00.00.00.0
Stage IIIB8.67.89.116.7N/Aa0.0
Stage IV6.710.70.00.00.00.0
Table 3. Predictors of Immediate Breast Reconstruction
CovariateNo. of patientsUnivariate analysisMultivariate analysis
OR95% CIP valueOR95% CIP value
  1. OR: odds ratio; 95% CI: 95% confidence interval; AJCC: American Joint Committee on Cancer; SD: standard deviation.

AJCC stage1004
 Stage 0 (reference)1801.01.0
 Stage I2980.410.28–0.60< 0.0010.450.29–0.70< 0.001
 Stage IIA2040.250.16–0.39< 0.0010.220.14–0.35< 0.001
 Stage IIB1160.140.08–0.23< 0.0010.110.06–0.20< 0.001
 Stage IIIA680.0070.001–0.05< 0.0010.0050.001–0.03< 0.001
 Stage IIIB930.040.02–0.09< 0.0010.030.01–0.07< 0.001
 Stage IV450.030.01–0.11< 0.0010.030.008–0.10< 0.001
Age (per yr)0.960.94–0.97< 0.0010.960.93–0.97< 0.001
Age (per SD [12.2 yrs])0.570.50–0.66< 0.0010.580.45–0.74< 0.001
Body mass index (per kg/m2)0.960.94–0.98< 0.0010.990.97–1.00.13
Race1004 < 0.001   
 White (reference)7181.01.0
 African American990.380.23–0.63< 0.0010.340.18–0.620.001
 Hispanic1121.040.69–1.550.870.930.56–1.530.77
 Asian451.090.59–2.00.790.500.24–1.040.06
 Middle Eastern300.170.05–0.550.0030.080.02–0.380.002
Secondary model       
 Income (per $1,000 increase in Zip code median income)1.011.01–1.02< 0.0011.011.001–1.020.02
 Income (per SD [$20,620])1.341.17–1.56< 0.0011.261.03–1.530.02

Table 3 summarizes the predictors of immediate reconstruction that were used to construct the multivariate model. After adjustment for the baseline characteristics of age, AJCC stage, and BMI, African Americans still were significantly less likely to undergo immediate breast reconstruction compared with whites, with an adjusted OR of 0.34 (95% CI, 0.18–0.62; P = 0.001). Middle Eastern women also had a lower rate of immediate breast reconstruction than white women, with an adjusted OR of 0.08 (95% CI, 0.02–0.38; P = 0.002). Asian women showed a trend toward lower rates of immediate breast reconstruction, with an adjusted OR of 0.50 (95% CI, 0.24–.1.04; P = 0.06). Hispanic women did not have significantly different rates of immediate reconstruction compared with white women (adjusted OR, 0.93; 95% CI, 0.56–1.53; P = 0.77).

To assess the effects of socioeconomic status, a secondary multivariate model was constructed incorporating the economic covariates of median household income for ZIP code and insurance type and the biologic covariates of age, AJCC stage, and BMI. The results were similar to those generated in the primary model. African-American women had a lower adjusted rate of immediate breast reconstruction compared with white women (adjusted OR, 0.36; 95% CI, 0.18–0.71; P = 0.003), as did Asian women (adjusted OR, 0.31; 95% CI, 0.14–0.70; P = 0.05). With the addition of these covariates related to socioeconomic status, Middle Eastern women (adjusted OR, 0.67; 95% CI, 0.09–5.15; P = 0.698) and Hispanic women (adjusted OR, 1.013; 95% CI, 0.53–1.92; P = 0.968) did not show significant differences compared with white women.

At each of the four branch points in the immediate breast reconstruction decision pathway, differences were observed between African-American women and white women. The ORs were calculated for a yes answer at each branch point, adjusting for the biologic factors of age, AJCC stage, and BMI. Differences were seen between African-American women and white women at each step: 1) whether the surgical oncologist offered a referral for breast reconstruction (adjusted OR, 0.52; 95% CI, 0.30–0.89; P = 0.02), 2) whether the patient accepted such a referral (adjusted OR, 0.33; 95% CI, 0.16–0.71; P = 0.003), 3) whether the plastic surgeon recommended reconstruction (adjusted OR, 0.35; 95% CI, 0.12–.1.03; P = 0.06), and 4) whether reconstruction was chosen by the patient (adjusted OR, 0.50; 95% CI, 0.15–1.64; P = 0.25). The steps on the pathway to immediate breast reconstruction with adjusted ORs for African-American women and white women are shown in Figure 2. The univariate (unadjusted) and multivariate (adjusted) ORs for patients from each ethnic group at each branch point are shown in Table 4.

Figure 2.

This is an illustration of the pathway to immediate breast reconstruction for African-American women and white women. Data are represented graphically as odds ratios (ORs) (point estimates adjusted for the covariates of age, stage, and body mass index) with 95% confidence intervals (95% CI). The open diamond represents the overall adjusted OR for immediate breast reconstruction in African-American women compared with white women.

Table 4. Pathway To Immediate Breast Reconstruction by Race/Ethnicity (Relative to White Women)
Race/ethnicityUnivariate analysisMultivariate analysis
OR95% CIP valueOR95% CIP value
  1. OR: odds ratio; 95% CI: 95% confidence interval.

White
African American      
 Step 1: Surgeon referral0.560.37–0.860.0080.520.30–0.890.02
 Step 2: Referral accepted0.460.25–0.870.020.330.16–0.710.003
 Step 3: Plastics offered0.340.13–0.900.030.350.12–1.030.06
 Step 4: Reconstruction0.580.21–1.630.300.500.15–1.640.25
Hispanic      
 Step 1: Surgeon referral1.00.66–1.510.990.960.58–1.610.89
 Step 2: Referral accepted0.730.41–1.320.300.630.33–1.210.17
 Step 3: Plastics offered4.160.55–31.260.174.30.55–33.710.16
 Step 4: Reconstruction1.340.50–3.560.561.150.41–3.20.79
Asian      
 Step 1: Surgeon referral0.990.53–1.840.970.500.24–1.080.08
 Step 2: Referral accepted1.170.43–3.160.760.570.19–1.640.31
 Step 3: Plastics offered0.860.19–3.850.840.580.12–2.910.51
 Step 4: Reconstruction1.380.31–6.140.670.850.18–3.980.84
Middle Eastern      
 Step 1: Surgeon referral0.350.16–0.740.0060.250.09–0.700.008
 Step 2: Referral accepted1.140.24–5.390.870.340.06–1.920.22
 Step 3: Plastics offered0.160.04–0.690.010.230.03–1.860.17
 Step 4: Reconstruction0.150.03–0.740.020.090.01–0.740.03

Delayed Reconstruction

Eighty-one patients underwent consultation for delayed breast reconstruction at The University of Texas M. D. Anderson Cancer Center during the follow-up period of this analysis (range, 15–27 months postmastectomy), comprising 8.1% of the original cohort and 13.0% of the 623 patients who did not undergo immediate breast reconstruction. Fifty of those women underwent delayed reconstruction during the follow-up period. Thirty-four white women (4.7%), 2 African-American women (2.0%), 3 Hispanic women (2.7%), and no Asian women underwent delayed breast reconstruction. Middle Eastern women had a higher rate of delayed breast reconstruction (11 patients; 36.7%) compared with the other ethnic groups.

Any Reconstruction

The unadjusted rates of any breast reconstruction (immediate or delayed) by ethnic group are displayed in Figure 3. The crude rate of any reconstruction for whites was 44.7%, compared with 22.2% for African-Americans, 44.6% for Hispanics, 42.2% for Asians, and 46.7% for Middle Eastern women.

Figure 3.

This bar chart illustrates any breast reconstruction (immediate or delayed) by ethnicity. African-American women had a significantly lower unadjusted rate of any reconstruction compared with white women (chi-square analysis; P values as shown).

Multivariate analysis adjusting for the covariates of age, stage, and BMI was used to determine whether race was an independent predictor of any reconstruction. The results paralleled the results found for immediate breast reconstruction. African Americans and Asians were less likely to have undergone immediate or delayed reconstruction: The adjusted OR for African Americans was 0.30 (95% CI, 0.17–0.54; P < 0.001), and the adjusted OR for Asians was 0.36 (95% CI, 0.18–0.74; P = 0.005). The rates of immediate or delayed breast reconstruction for Hispanic women (OR, 0.75; 95% CI, 0.46–1.21; P = 0.24) and Middle Eastern women (OR, 0.75; 95% CI, 0.29–1.98; P = 0.56) did not differ significantly from the rates among white women.

DISCUSSION

In our analysis, we found that in a single, large cancer center with a diverse patient population, African-American women underwent immediate breast reconstruction after mastectomy for breast cancer at significantly lower rates compared with white, Hispanic, and Asian women. In a multivariate analysis adjusting for age, stage, and BMI, African-American and Asian women had lower adjusted rates of immediate breast reconstruction compared with whites. Hispanic women did not have different rates of overall reconstruction compared with white women. Middle Eastern women had higher rates of delayed breast reconstruction, and their overall reconstructive rates were congruent with those of white women. The rate of delayed breast reconstruction was lower in African-American and Asian women compared with their white counterparts and did not affect the overall gap in reconstruction between those two groups and white women. Preoperative assessment of the need for postoperative radiation therapy did not explain these differences. Analysis of the steps to immediate breast reconstruction demonstrated that the surgical oncologist, the patient, and the plastic surgeon all may contribute to the overall difference.

Issues of access and parity in health care in general and breast cancer in particular are of increasing relevance in our multicultural society. Direct comparisons of outcomes or decisions based on patient factors, such as race and ethnicity, can be problematic. Limited studies may not be relevant to the general population. Larger studies based on registries necessarily are limited by their multicentricity, the heterogeneity of patients and practices by region, and the inability to identify individual patient characteristics and associate them with outcomes. Previous studies that have included analyses of breast reconstruction in different racial groups have produced varying results. A study of postmastectomy breast reconstruction among 4688 patients with breast cancer in Connecticut from 1992 to 1996 found that reconstruction was not related to the patient's race (black vs. white) or tumor size but declined with increasing age at diagnosis and with poverty.30 Data from the Surveillance, Epidemiology, and End Results (SEER) Program indicate that 15% of registered patients who underwent mastectomy in 1998 had reconstruction within the first 4 months after mastectomy. Compared with white women, African-American, Hispanic, and Asian women were significantly less likely to undergo reconstruction than white women (OR, 0.48, 0.45, and 0.29, respectively; all P values < 0.001).27 The effects of disease stage, geography, and practice pattern heterogeneity on conclusions drawn from large pooled data sets like the SEER registry remain unclear.

Race and ethnicity are constructed categories that may reflect history, geographic origin, cultural identity, and socioeconomic status as well as genetics and biology, all in varying degrees. From a purely biologic point of view, human geneticists and anthropologists generally agree that human races do not exist.31 However, races do exist as social constructs and are sustained by internal and external racial identification.32 Whatever their basis, different racial groups have substantially different rates of diagnosis, treatment, and outcome in a variety of diseases, including cancer.33, 34 Thus, explicating these differences has inherent value, in that the understanding and treatment of patients with identified diseases such as breast cancer may be improved.

The University of Texas M. D. Anderson Cancer Center may be suited particularly to explore these issues. The patient population is diverse in terms of race, ethnicity, and financial background. Clinical departments are made up of medical staffs that treat patients on the basis of departmental protocols, practice guidelines, and multidisciplinary evaluation. Reimbursement of faculty is based solely on academic salary with no fee for service. All registered patients of The University of Texas M. D. Anderson Cancer Center are eligible for treatment, including chemotherapy, radiation therapy, oncologic surgery, and reconstructive surgery.

The limitations of any studies based on race and ethnicity, including this analysis, are evident. Most African Americans have ancestors who were white or Caucasian; many white individuals have ancestors from Africa.35 Data from the 2000 United States Census show that almost 7 million people identified themselves as members of more than 1 race; ≈ 800,000 respondents said they were both white and black.36 Hispanics can be of any race. At The University of Texas M. D. Anderson Cancer Center, the overwhelming majority of Hispanic/Latino patients are from Mexico, Central America, and South America. In contrast, a cancer center with a different geographic location and referral pattern may see a very different mix of patients who identify themselves as Hispanic. Asians in our study were largely of Chinese, Japanese, and Filipino descent but included three women of Indian descent and one international patient (i.e. one non–United States resident) each from Korea and Pakistan. The Middle Eastern patients in the current study were largely international patients who, by definition, had the resources to travel to a United States cancer center for their care; they likely are not representative of the women in their native countries. No patients who identified themselves as Native American were in our study cohort; therefore, we are unable to draw any conclusions about breast reconstruction in this group.

Breast reconstruction encompasses a broad range of procedures of varying complexity. Options include autologous tissue flaps and manufactured tissue expanders and implants. Treatment factors, such as the possible need for chest wall irradiation; patient factors, such as medical comorbidities, body habitus, and smoking history; and surgeon, patient, and insurer preferences may influence the type of reconstruction offered and chosen. The objective of the current study, however, was to determine the factors that influence patients' decisions about whether to undergo immediate breast reconstruction. For the purposes of this study, the type of reconstruction was considered secondary to the decision to undergo breast reconstruction of any kind.

Our results lead inevitably to the question: Why do these differences occur? Is there a tendency on the part of surgical oncologists or plastic surgeons to perceive the reconstructive needs of different patients differently? Do patients of various ethnic groups seek treatment with altered expectations and needs? The first branch point in our decision pathway, whether the patient was offered a referral for immediate breast reconstruction from her surgical oncologist, can appear deceptively one-sided. Patients enter into patient-physician relationships with individual preferences and needs.37 Certain women may request or insist upon referral for reconstruction a priori, regardless of the surgical oncologist's recommendation. Our methods magnified the physician's side of the patient-physician relationship by relying on physician dictations to assess whether reconstructive options were discussed, recommended, or declined. Because patients were not contacted directly for this study, and because it has been found that patients and physicians perceive patient-physician interactions differently,38 our findings are incomplete. However, the fact that this study was conducted at a single center with a consistent group of surgeons suggests that any confounding based on incomplete reporting should have been consistent across racial groups. Differences in body image and self-esteem between groups of patients also may contribute to discrepancies. African-American women and girls have been found by some investigators to have higher self-esteem compared with their white counterparts.39–41 Asians have been found to have distinct cultural differences from other Americans and, in some cases, are more accepting of physician recommendations and less questioning of medical authority compared with women of other races.42–44 Patient social networks, perception of self in relation to the health care system, and factors as diverse as the level of Internet use may affect patient decision-making.45

Another important issue is that of breast conservation therapy (BCT) versus mastectomy. BCT is the current standard of care46, 47 for patients with breast cancer that is amenable to segmental resection and who are willing and able to undergo postsurgical breast irradiation. In our study, we analyzed the choices of women who had decided to undergo mastectomy. To understand more completely the decisions of patients with breast cancer, the earlier branch point of BCT versus mastectomy is salient. Additional studies currently are underway that will help us determine whether women from different racial and ethnic backgrounds, among other characteristics, have different rates of BCT and mastectomy. Recently, investigators reported that elderly women with early-stage breast cancer who consulted medical oncologists before surgery were more likely to receive guideline-recommended care, including higher rates of definitive surgery and axillary lymph node dissection and lower rates of BCT.48 The opinions of medical oncologists also may affect patients' decision-making regarding breast reconstruction after mastectomy. In the current study, not all patients were seen by medical oncologists preoperatively; and, because discussions regarding reconstruction seldom were documented in medical oncologists' notes, the effects of medical oncologists' advice on breast reconstruction rates are unknown.

Finally, although the data on patients in our cohort were gathered from a prospective database, our analysis of the medical records was retrospective in nature. Retrospective studies inherently are limited in their ability to identify cause and predict outcome. Prospective studies are needed to provide better answer to questions regarding the effect of patient background, including race and ethnicity, on decision-making in breast cancer treatment.

Acknowledgements

The authors thank Marc S. Sabatine, M.D., M.P.H., and Harry R. Gibbs, M.D., for their thoughtful review of the current article. The authors also thank Bertha Martinez for her meticulous administrative assistance.

Ancillary