Fax: (651) 690-7774
Version of Record online: 15 NOV 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 11, pages 2268–2275, 1 December 2002
How to Cite
Prehn, A. W., Topol, B., Stewart, S., Glaser, S. L., O'Connor, L. and West, D. W. (2002), Differences in treatment patterns for localized breast carcinoma among Asian/Pacific islander women. Cancer, 95: 2268–2275. doi: 10.1002/cncr.10965
See editorial on pages 2257–9, this issue.
The content of this publication does not necessarily reflect the views or policies of the U.S. Department of Health and Human Services or the California Department of Health Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. government or state of California.
- Issue online: 19 NOV 2002
- Version of Record online: 15 NOV 2002
- Manuscript Accepted: 26 JUN 2002
- Manuscript Revised: 21 JUN 2002
- Manuscript Received: 10 APR 2002
- National Cancer Institute. Grant Number: N01-CN-65107
- National Institutes of Health
- California Cancer Registry, a project of the Cancer Surveillance Section, California Department of Health Services
- Public Health Institute. Grant Numbers: 050M-8701/8-S1522, 453A-8704-S1474
- Department of the Army. Grant Number: DAMD17-94-J-4508
- Pacific Islanders;
- breast neoplasms;
- breast-conserving surgery (BCS);
- radiation therapy;
- hormone therapy;
- End Results (SEER)
Many studies have examined racial/ethnic differences in treatment for localized breast carcinoma, but to the authors' knowledge few have included Asian/Pacific Islander (API) women.
The population-based study included API and non-Hispanic white women diagnosed with localized invasive breast carcinoma in the Greater San Francisco Bay Area during 1994 (n = 1772). Multiple logistic regression was used to assess the association between race/ethnicity and type of surgery, radiation therapy following breast-conserving surgery (BCS), and hormone therapy for estrogen receptor-positive tumors while adjusting for demographic, medical, and census block-group socioeconomic characteristics.
API women were significantly more likely to undergo mastectomies than white women (58% vs. 42%). This difference remained for Chinese and Filipino women after multivariate adjustment (odds ratio vs. whites [OR] = 2.4, 95% confidence interval [95% CI] = 1.4–4.2; OR [95%CI] = 1.8[1.0–3.1], respectively). Chinese women were also more likely than white women to not receive adjuvant therapy, be it radiation after BCS or hormone therapy for estrogen receptor-positive disease. Other API women did not differ from white women in adjuvant therapy use.
This population-based study identified differences in treatment for localized breast carcinoma by race/ethnicity that were not explained by differences in demographic, medical, or socioeconomic characteristics. These results underscore the importance of looking at treatment patterns separately for API subgroups and support the need for research into cultural differences that may influence breast carcinoma treatment choices. Cancer 2002;95:2268–75. © 2002 American Cancer Society.
In 1991, a National Institutes of Health consensus statement on the treatment of early-stage breast carcinoma made recommendations that included the use of breast-conserving surgery (BCS) with adjuvant radiation therapy instead of mastectomy, whenever possible.1 Since that time, many studies have evaluated factors related to compliance with the treatment recommendations, including race/ethnicity.2–12 To our knowledge, very few of these studies, however, have included Asian/Pacific Islanders (APIs), a population that is growing rapidly in the United States.13 As breast carcinoma is the most common malignancy diagnosed among API women, and as a significant proportion of these women are diagnosed with early-stage disease,14, 15 it is important to understand whether this group is receiving the recommended treatment. Morris et al.9 found that API women in California were more likely to undergo a mastectomy than white, black, or Hispanic women. However, APIs are a diverse group, representing many ethnic and cultural backgrounds. Disaggregating the ethnic subgroups in studies of treatment differences may help to better understand how and why these differences occur. Although the use of adjuvant therapy, such as radiation and hormone therapy, is an important part of the treatment protocol, racial/ethnic differences in this aspect of treatment have not been well studied. To address these gaps in the literature, we undertook the following study to 1) evaluate surgical treatment patterns for early-stage breast carcinoma among Chinese, Japanese, Filipino, and other API women compared with white women; 2) explore their respective uses of adjuvant radiation and hormone therapy; and 3) investigate possible explanations for any racial/ethnic differences found.
MATERIALS AND METHODS
Breast carcinoma cases were identified through the Northern California Cancer Center's Greater Bay Area Cancer Registry (GBACR), a participant in both the National Cancer Institute's (NCI) Surveillance, Epidemiology, and End Results (SEER) program and the statewide California Cancer Registry. Women included in the study were all API and non-Hispanic white females diagnosed with localized first primary invasive breast carcinoma in 1994 while resident in the greater San Francisco Bay Area (Alameda, Contra Costa, Marin, Monterey, San Benito, San Francisco, San Mateo, Santa Clara, and Santa Cruz counties). SEER historic stage of disease at diagnosis (local, regional and distant) was used to identify women with localized disease, defined as invasive carcinoma that had not spread beyond the breast16 (n = 1780). This equates roughly to the American Joint Committee on Cancer (AJCC) TNM classifications of any TN0M0 (tumor of any size, no regional lymph node metastasis, and no distant metastasis).17 Women without surgical treatment were excluded from the study (n = 8), leaving a final sample of 1772 women.
Demographic and Tumor Characteristics
Covariates were chosen based on associations with race/ethnicity and/or treatment as demonstrated in previous research. The GBACR provided patient and tumor information routinely abstracted from medical records as part of state-mandated cancer surveillance. Demographic variables included age at diagnosis (<65, 65–74, 75+ years), race/ethnicity (white, Chinese, Japanese, Filipino, other API), marital status (never married, married, separated/divorced, widowed/unknown), and whether patients were treated at facilities run by a local health maintenance organization (HMO). This last variable was used as a proxy for HMO membership (yes/no).
Tumor characteristics included size (millimeters), number of lymph nodes evaluated surgically, histologic subtype and grade, and estrogen and progesterone receptor status. Tumors were categorized histologically as ductal carcinoma (International Classification of Diseases-oncology, version 2 [ICD-O-2]18 morphology code 8500), lobular carcinoma (ICD-O-2 code 8520), or other (all others). Histologic grade was categorized as well differentiated, moderately differentiated, poorly or undifferentiated, or unknown. Estrogen and progesterone receptor status were coded as positive, negative, or unknown.
1990 Census Data
To estimate the socioeconomic status (SES) of patients, we used 1990 Census data summarized at the block-group level (census tract subdivisions of approximately 1000 persons) acquired from summary tape file 3A.19 Data for each block-group included median age, median household income, per capita income, percent of households living below poverty level, percent of persons 25 and older with a college education, percent of employed persons in a working class occupation, the total number of children ever born to women 15 years or older, and the total number of women 15 years or older. The last two variables were combined to determine the average number of children ever born per woman.
Treatment data had been obtained by a larger project to update and verify treatment information in the GBACR database for all 1994 breast carcinoma cases. This larger project used algorithms based on the NCI's Physician's Data Query Breast Cancer recommended treatment guidelines20 to identify breast carcinoma cases in the registry who lacked complete information on surgery, radiation therapy, chemotherapy, or hormone therapy. Missing treatment data were sought from a mailed query to the women's primary physician or obtained directly from physicians' office records. Of the 1772 women included in the current study, 23 (1%) had had complete treatment information in the GBACR database according to the algorithm. Updated information had been received for 1454 (83%) of the remaining women, with the additional treatment information merely verifying GBACR data for 1188 (82%) and new information being obtained for 266 (18%). None of these proportions differed by race/ethnicity.
These updated records provided information on type of surgery and whether the patient received adjuvant radiation and/or hormone therapy. For patients whose records were not updated, treatment information was obtained from available data in the registry. Surgery was defined as mastectomy or BCS, which included segmental mastectomy, lumpectomy, tylectomy, wedge resection, quadrantectomy, nipple resection, or excisional biopsy. For these analyses, radiation and hormone therapy were dichotomized (any/none).
The associations of primary interest were between race/ethnicity and three types of breast carcinoma treatment: type of surgery (mastectomy vs. BCS), radiation therapy following BCS (none vs. any), and hormone therapy for women 50 years and older with estrogen receptor-positive tumors, irrespective of surgery type (none vs. any). For the surgery analysis, only women with tumors less than or equal to 5 cm (AJCC classifications T1/2) were included, so that all women would be potential candidates for BCS. Chi-square statistics and analysis of variance were used to assess differences in demographic and tumor characteristics by type of treatment. Multiple logistic regression was used to calculate adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) for the association between race/ethnicity and breast carcinoma treatment while adjusting for possible confounders. Variables that either contributed significantly to the model (Wald P < 0.10) or acted as strong confounders of the association between race/ethnicity and treatment type (changed the OR for any racial/ethnic group by ≥10%) were retained in the final models. Only women without missing values for the variables of interest were included in the multivariate models (see Tables 1–3 for respective numbers). All analyses were performed using SAS 6.12 for UNIX21 and Intercooled Stata for Windows, version 5.0.22
|White||Chinese||Japanese||Filipino||All other API|
|Age at diagnosis (%) (yrs)||<0.01|
|Marital status (%)||<0.01|
|Member of local HMO (%)||26.4||27.4||34.4||37.7||23.8||0.25|
|Histologic subtype18 (%)||0.16|
|Ductal carcinoma (8500)||72.2||71.2||71.9||81.2||76.2|
|Lobular carcinoma (8520)||10.3||2.7||12.5||4.4||4.8|
|Histologic grade (%)||0.10|
|Estrogen receptor status (%)||0.35|
|Progesterone receptor status (%)||0.62|
|Mean (SD)||F P value|
|Tumor size (mm) (n = 1636)||17.3 (16.5)||21.1 (15.3)||15.5 (9.5)||19.5 (13.4)||20.1 (15.7)||0.19|
|No. of lymph nodes evaluated (n = 1751)||11.9 (7.9)||12.8 (8.7)||14.8 (8.4)||13.5 (8.0)||13.0 (7.9)||0.11|
|Census characteristics (n = 1534)|
|Median household income, $ in thousands||53.5 (21.3)||45.0 (21.1)||54.3 (25.1)||43.3 (12.6)||43.8 (15.2)||<0.01|
|Per capita income, $ in thousands||30.9 (13.5)||24.6 (12.0)||32.3 (16.9)||20.3 (5.8)||21.9 (7.3)||<0.01|
|Median age (yrs)||38.4 (7.6)||38.4 (8.6)||36.4 (4.6)||32.4 (3.9)||33.0 (5.4)||<0.01|
|Percent households below poverty level||5.5 (5.4)||9.2 (10.3)||5.7 (5.5)||9.0 (8.0)||8.8 (7.6)||<0.01|
|Percent in working class occupations||53.3 (14.2)||60.3 (15.7)||51.6 (15.4)||66.7 (12.9)||62.0 (15.0)||<0.01|
|Percent 25+ years with college degree||38.0 (18.3)||32.9 (17.2)||40.6 (19.0)||24.5 (14.3)||29.4 (15.8)||<0.01|
|Average no. of children/women 15+||1.55 (0.38)||1.66 (0.48)||1.47 (0.45)||1.69 (0.41)||1.69 (0.47)||<0.01|
|OR||95% CI||OR||95% CI|
|Other Asian/Pacific Islander||1.6||0.9–3.2||1.6||0.8–3.1|
|No radiation after BCS (N = 849)a||No hormone therapy (N = 798)b|
|AOR||95% CI||AOR||95% CI|
|Other Asian/Pacific Islander||0.5||0.1–3.7||0.9||0.3–2.7|
Characteristics of the study population by race/ethnicity are shown in Table 1. In general, API women were younger and more likely to be married than whites. Distributions of histologic subtype and grade were similar for white and Japanese women and differed from the other racial/ethnic groups, although these differences were not statistically significant. There were no statistically significant racial/ethnic differences in estrogen and progesterone receptor status, tumor size, or the number of lymph nodes evaluated. White and Japanese women were significantly more likely than Chinese, Filipino, and women of other API groups to live in census block-groups with a higher SES indicated by factors such as higher income, lower fertility rate, and higher average level of education.
Overall, 56% of women received BCS and the remaining 44% underwent a mastectomy. However, surgical treatment varied by race/ethnicity: 63% of Chinese, 59% of Japanese, 55% of Filipinas, and 55% of women from other API groups had mastectomies compared with only 42% of white women. Even after adjustment for demographic and tumor characteristics, the odds of having a mastectomy remained twice as great for Chinese and Filipino women compared with white women (Table 2). The odds of having a mastectomy were also greater for Japanese women and women of other API groups compared with white women, although these results were not statistically significant.
Table 3 presents the results of the multivariate regression analyses examining the association between race/ethnicity and the likelihood of not receiving adjuvant therapy. The odds of not receiving radiation therapy after BCS were three times greater for Chinese women than for white women. Similarly, the odds of not receiving hormonal therapy were twice as great for Chinese women compared with white women. For the other racial/ethnic groups, ORs (vs. whites) ranged widely, from 0.5 (radiation therapy for other API women) to 2.6 (radiation therapy for Japanese), but none were statistically significant.
This study found that treatment patterns for localized breast carcinoma differed between API and white women. Overall, API women were less likely than white women to receive recommended treatment, especially BCS. In addition, there were large differences across API subgroups regarding the use of adjuvant therapy. Even after multivariate adjustment, Chinese women were significantly more likely than white women not to receive recommended adjuvant therapy, be it radiation after BCS or hormone therapy for estrogen receptor-positive disease. Women in the other API subgroups did not differ from white women in their use of adjuvant therapy.
These findings are consistent with the results on breast carcinoma treatment patterns in API women of Morris et al.9 In addition, Kagawa-Singer et al.23 found significantly lower rates of BCS plus adjuvant therapy among Asian-American women compared with Anglo Americans and Lin et al.24 reported a higher risk of mastectomy among Asian American women (Chinese, Japanese, and Filipino) compared with non-Hispanic whites.24 Our study expands these results by disaggregating the API category into component ethnic groups and identifying different treatment patterns within these groups compared with white women, most notably for Chinese and Filipino women. A recently published study using Hawaiian tumor registry data linked to insurance claims also found that Filipino women were less likely to receive BCS than women from other ethnic groups, although their results were not statistically significant and their study population was limited to mostly younger women who belonged to a particular health plan.25 Lee similarly noted higher mastectomy rates in Filipino women from the study of Lin et al.24, 26 In addition, our findings that Chinese women were more likely to have a mastectomy and less likely to use hormone therapy than white women are consistent with a study by Lee et al.27 However, their study did not focus on localized breast carcinoma and did not include other API groups.
Previous studies have found that racial/ethnic differences in breast carcinoma treatment disappeared after adjustment for tumor characteristics,10 whereas others have found that SES was the most important factor in determining breast carcinoma treatment.3, 8, 12, 28 Morrow et al.29 reported that contraindications to BCS leading nonwhites to have more mastectomies were social rather than medical (e.g., difficulty completing a radiation therapy course because of time and/or transportation). However, in the current analysis, differences in breast carcinoma treatment remained even after adjustment for both tumor characteristics and the socioeconomic characteristics of the patient's neighborhood.
There are several reasons why our results may differ from previous research. Previous studies were focused mainly on black/white differences and often did not include API populations. Therefore, those results may not apply directly to our study population. We used the socioeconomic characteristics of the patient's neighborhood as a proxy for the patient's SES in our analysis. As the census variables were not race specific, it is possible that they were not a good indicator of an individual's SES in our API populations. Some of the association between SES and treatment differences may be unaccounted for in our analysis. However, it has been shown that census-level data are effective proxies for individual-level data in similar studies.8, 9, 12, 25, 30–32 It is unlikely that any residual effect of SES not accounted for in our study is responsible for all of the racial/ethnic differences in breast carcinoma treatment found in our analysis.
Another explanation for the racial/ethnic differences in treatment may be culture, such as cultural beliefs about body image, illness and death, approaches to medical decision-making, and issues with patient/provider communication. In general, Asian cultures view illnesses as inevitable and unchangeable and individuals may be more likely to avoid treatment.33, 34 Modesty and embarrassment may prevent many Asian women from participating fully in the treatment decision.35, 36 In addition, the breast is not as important in the perception of body image in Asian cultures as in Western cultures.23 As a result, concerns about the cosmetic results of surgery may not play a prominent role in influencing treatment choices. There are also cultural differences in medical decision-making. For example, Chinese culture holds a family-centered model of medical decision-making, where the family is of primary importance in determining the course of treatment.34, 36–38 It has been hypothesized that Chinese women may not choose BCS and radiation therapy because of the inconvenience that the radiation therapy would have to the family.22, 33 In addition, medical authorities are treated with deference and it is considered disrespectful to question their opinions.22 Given the evidence that physicians are still more likely to recommend mastectomies than BCS as the treatment of choice,28 it is possible that Asian women's selection of mastectomy is based solely on their physician's recommendation. Problems with patient/provider communication also may have influenced the course of breast carcinoma treatment in these women.33, 39 Patients may not understand their full range of treatment options if they are not fluent in English or may have difficulty communicating concerns about the treatment options offered to them. It would be informative to obtain information regarding whether treatment patterns differ when providers of the same race/ethnicity as the patient and/or those who speak the same language as the patient are involved in the treatment decisions.
In addition to cultural factors, the degree of acculturation may also influence breast carcinoma treatment choices. The ethnic groups in this study vary with regards to their migration history and degree of acculturation. For example, the Japanese are a more acculturated (Westernized) group than the more recently immigrated Chinese and Filipinos40 and may be more likely to have treatment patterns similar to white women than other API groups. However, our data do not show any clear pattern in this regard. Chinese women had treatment patterns that differed significantly from whites, although Filipinas and women from other API groups did not. Understanding the barriers to recommended treatment is complicated and most likely involves a constellation of factors.
This study was conducted using a large population-based dataset with updated and complete information on the first course of treatment. The data were obtained from a SEER cancer registry, in which quality control is rigorous and case ascertainment is virtually complete.15 However, these data do not include information on several contraindications to BCS, such as current pregnancy status, previous irradiation, and tumor-to-breast size ratio.6, 27, 41 Although pregnancy status is not likely to be a large confounder of the study results given the age of the population (78% age 50 or older at diagnosis), the tumor-to-breast size ratio is of concern. API women have a smaller average breast size than women of other racial/ethnic groups,42–44 which could be a contraindication for BCS. However, analyses for the 1100 women with small tumors (≤ 20 mm), which should be eligible for BCS even in small-breasted women,42 did not produce appreciably different results (data not shown). We also found racial/ethnic differences in the use of adjuvant therapy as well as surgery type, indicating that there were differences in breast carcinoma treatment not completely explainable by surgical contraindications alone.
This study did not include information regarding treatment patterns at particular hospitals and/or clinics in which the women were treated. If the majority of the Chinese and Filipino women were treated by physicians or at hospitals with a bias toward performing mastectomies, that could explain the increased prevalence of mastectomies in these populations. This does not explain, however, why these two groups had very different patterns of adjuvant therapy use. Another possible influence on these treatment differences is the use of alternative therapies, such as traditional Chinese medicine. It may be that Chinese women choose to use traditional medicines as adjuvant therapies instead of radiation or hormonal therapies. Although we had no information on the use of alternative therapies in our dataset, a study by Lee et al.27 also conducted in the Bay Area suggested that a substantial percentage of women diagnosed with breast carcinoma between 1990 and 1992 used alternative therapies, regardless of race/ethnicity. However, the study did not include information on the relationship between the use of conventional and alternative therapies and how that might differ by racial/ethnic group. It might be that Chinese women use alternative therapies instead of standard adjuvant therapies whereas women of other races/ethnicities use a combination of the two therapies. Further research on the use of alternative therapies is needed to fully understand how they might explain part of the racial/ethnic differences in breast carcinoma treatment.
The treatment differences found among racial/ethnic groups in this study necessitate a close look at the current treatment recommendations and whether efforts should be made to rectify the differences found. There are no survival differences between women who undergo BCS and those who have a mastectomy.41, 45, 46 Therefore, the racial/ethnic differences in surgical treatment seen in this study would not likely affect the women's prognosis. However, the use of radiation therapy after BCS reduced local disease recurrence rates in several clinical trials, as has the use of tamoxifen for estrogen receptor-positive tumors.41, 45, 47, 48 Consequently, our finding that Chinese women are less likely to receive adjuvant therapy after their initial surgery is of concern, because this choice potentially affects both short and long-term survival. Barriers to receiving recommended adjuvant therapy, including inaccessibility to a radiation therapy center, inability to alter work schedule to accommodate treatment regimens, and the psychologic and emotional sequelae of breast carcinoma treatments, must be addressed in a culturally appropriate manner.
Differences in breast carcinoma treatment patterns for Chinese, Japanese, and Filipino women compared with white women underscore the importance of looking at treatment patterns separately for API subgroups. The reasons for these racial/ethnic treatment differences remain unclear, although in this study they were not due to differences in tumor and socioeconomic characteristics. Additional research is needed to better understand the cultural differences that may influence breast carcinoma treatment choices so that all women will have the opportunity to receive treatment that is therapeutically effective.
The authors thank Cynthia O'Malley, Ph.D., and Scarlett Lin, M.P.H., for their thoughtful reviews.
- 13Asian and Pacific Islander Center for Census Information and Services. Our ten years of growth: a demographic analysis on Asian and Pacific Islander Americans. San Francisco: APICCIS, 1992.
- 14MillerBA, KolonelLN, BernsteinL, et al., eds. Racial/ethnic patterns of cancer in the United States, 1988–1992. NIH Pub. No. 96–4104. Bethesda: National Cancer Institute, 1996.
- 15Cancer incidence in Chinese, Japanese and Filipinos in the US and Asia, 1988–1992. Union City, CA: Northern California Cancer Center, 1999., , , et al.
- 16RiesLAG, EisnerMP, KosaryCL, et al., eds. SEER cancer statistics review, 1973–1998. Bethesda: National Cancer Institute, 2001.
- 17FlemingID, CooperJS, HensonDE, et al., eds. American Joint Committee on Cancer staging manual, 5th edition. Philadelphia: J.B. Lippincott, 1997.
- 18PercyC, Van HoltenV, MuirC, eds. International classification of diseases for oncology, 2nd ed. Geneva: World Health Organization, 1990.
- 19Bureau of the Census. Census of population and housing,1990: summary tape file 3 on CD-ROM (machine-readable data files). Washington, DC: Bureau of the Census, 1992.
- 20National Cancer Institute. Physician's data query information for health care professionals, breast cancer. Bethesda: National Cancer Institute, 1995.
- 21SAS for UNIX. Version 6.12. Cary, NC: Statistical Analysis Software, 1996.
- 22Stata 5.0 for Windows 95/NT. College Station, TX: Stata Corporation, 1998.
- 40Overview of Asian and Pacific Islander Americans. In: ZaneNWS, TakeuchiDT, YoungKNJ. Confronting critical health issues in Asian and Pacific Islander Americans. Thousand Oaks, CA: Sage, 1994: 3–21., .
- 43Difficulty of obtaining wide tumour to deep fascial margins in Asian women with breast cancers: implications on treatment and prognosis. Ann Acad Med Singapore. 1996; 23: 241–444., , , , .
- 47Early Breast Cancer Trialists' Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy: 133 randomised trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet. 1992; 339: 1–15.
- 48Early Breast Cancer Trialists' Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy: 133 randomised trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet. 1992; 339: 71–85.