Use of adjuvant trastuzumab in women with human epidermal growth factor receptor 2 (HER2)-positive breast cancer by race/ethnicity and education within the National Comprehensive Cancer Network




Trastuzumab for human epidermal growth factor receptor 2 (HER2)-positive breast cancer is highly efficacious yet costly and time-intensive, and few data are available about its use. The authors of this report examined receipt and completion of adjuvant trastuzumab by race/ethnicity and education for women with HER2-positive disease.


The National Comprehensive Cancer Network Breast Cancer Outcomes Database was used to identify 1109 women who were diagnosed with stage I through III, HER2-positive breast cancer during September 2005 through December 2008 and were followed for ≥1 year. The authors used multivariable logistic regression to assess the association of race/ethnicity and education with the receipt of trastuzumab and, among those women who initiated trastuzumab, with the completion of > 270 days of therapy.


The cohort was 75% white, 8% black, and 9% Hispanic; and 20% of women had attained a high school degree or less. Most women (83%) received trastuzumab, and no significant differences were observed according to race/ethnicity or socioeconomic status. Among the women who initiated trastuzumab, 73% of black women versus 87% of white women (P = .007) and 70% of women with less than a high school education versus 90% of women with a college degree completed > 270 days of therapy (P = .006). In adjusted analyses, black women (vs white women) and women without a high school degree (vs those with a college degree) had lower odds of completing therapy (black women: odds ratio, 0.45; 95% confidence interval, 0.27-074; white women: odds ratio, 0.27, 95% confidence interval, 0.14-0.51).


Differences in completing trastuzumab therapy were observed according to race and educational attainment among women who received treatment at National Comprehensive Cancer Network centers. Efforts to assure the appropriate use of trastuzumab and to understand treatment barriers are needed and may lead to improved outcomes.

The authors report differences in the rate at which patients complete treatment with trastuzumab according to race and education among women who receive treatment at National Comprehensive Cancer Network centers. Efforts to assure the appropriate use of trastuzumab and to understand treatment barriers are needed and may lead to improved outcomes. Cancer 2013. © 2012 American Cancer Society.


Human epidermal growth factor receptor 2 (HER2) is an important mediator of cell proliferation and differentiation.1,2 Although women with HER2-positive breast cancers traditionally have had poor prognoses,3 the natural history of this disease subtype changed dramatically with the development and widespread use of trastuzumab. Since June 2005, adjuvant treatment guidelines have recommended that standard chemotherapy be supplemented with trastuzumab for patients with HER2-positive cancers4 after large-scale randomized trials demonstrated an approximately 0.50 risk of disease recurrence compared with chemotherapy alone, with few adverse events.5-7 Trastuzumab administration involves a 30-minute to 90-minute infusion administered weekly or every 3 weeks for 1 year with an estimated total cost of > $100,000 when trastuzumab is incorporated into anthracycline-based chemotherapy regimens.8

Previous research has demonstrated disparities in the receipt of adjuvant therapies, including radiation and endocrine therapy, for black woman compared with white women.9-12 In addition, black women experience worse breast cancer outcomes than white women, even when matched for known prognostic factors.13 Although the reasons for racial disparities in outcomes are multifactorial, differences in receipt of efficacious treatments likely contribute. Differences in the receipt of breast cancer care and outcomes by socioeconomic status (SES) also have been observed,14-18 although studying associations of treatment and outcomes with SES is often reliant on area-level measures of education and income.

Although trastuzumab has substantially improved outcomes for women with HER2-positive breast cancer, a risk for disparity has arisen because of the high cost and time-intensive commitment required for therapy. The identification of potential treatment differences among women with HER2-positive breast cancer is crucial, yet few data are available about the receipt of trastuzumab. We examined disparities in the receipt and completion of adjuvant trastuzumab in a large cohort of women who were treated in National Comprehensive Cancer Network (NCCN) centers across the United States.


Data Source

Since 1997, the NCCN Breast Cancer Outcomes Database Project has evaluated patterns and outcomes of cancer care.19 It prospectively captures information on patient and tumor characteristics, treatments, and outcomes abstracted from patients, medical records, and institutional systems for women with breast cancer who receive treatment at member institutions.20-22 Patients are included in the database if they receive all or some of their treatment at a reporting center; those with 1-time consultations are not included. Eight centers contributed data to the current analysis: City of Hope National Medical Center, The University of Texas M. D. Anderson Cancer Center, Fox Chase Cancer Center, Dana-Farber Cancer Institute, Roswell Park Cancer Institute, H. Lee Moffitt Cancer Center, the University of Michigan Cancer Center, and Ohio State University. NCCN data include information abstracted from medical records on age, HER2 status, other tumor characteristics, comorbidity, and treatments received as well as patient-reported information on race/ethnicity, educational attainment, insurance, and employment. Data collected by the NCCN are subject to rigorous quality assurance and routine on-site audits of source documents against submitted data.

Study Cohorts

We identified 2 study cohorts. We examined receipt (ie, initiation) of trastuzumab in Cohort 1 and completion of trastuzumab therapy in Cohort 2. Cohort 1 included adult women with a first diagnosis of stage I, II, or III breast cancer that overexpressed HER2, defined as a “positive” result by fluorescence in situ hybridization or a score of 3 +, “high positive,” or “positive not otherwise specified” by immunohistochemistry. We restricted this cohort to 1145 women who were diagnosed with breast cancer between September 2005 and December 2008 who had follow-up data through December 2009. After excluding 36 women who had < 365 days of follow-up at the reporting institution, the first cohort included 1109 women.

For the second cohort, we focused on the 925 women from Cohort 1 who initiated any trastuzumab. We excluded 17 patients who had a break in their trastuzumab regimen > 30 days (because of uncertainty in interpreting the reasons for treatment break), 2 women who had 0 days of trastuzumab documented, and 18 women who stopped trastuzumab for progression, transfer of care, or death. Cohort 2 included 888 women.

Variables of Interest

Our dependent variables of interest were 1) receipt of adjuvant/neoadjuvant trastuzumab for the first cohort and 2) completion of adjuvant trastuzumab for the second cohort. Receipt of trastuzumab was defined as having a reported start date for trastuzumab at any time after diagnosis but before any recurrence. Completion of trastuzumab was defined as documented receipt of > 270 days of treatment, which was calculated by summing the duration of all trastuzumab treatments received in the neoadjuvant and/or adjuvant setting (considering all women who received ≥ 75% of the recommended 1 year of trastuzumab therapy as having completed therapy). For the 62 women who did not have an end date documented for trastuzumab, we used the date of last follow-up with medical oncology at the NCCN institution as a proxy for the end date. By using this definition, 59 of 62 women were categorized as having completed trastuzumab, and 3 of 62 women were defined as not having completed trastuzumab.

Our independent variables of interest were race/ethnicity and education. Race/ethnicity was categorized as non-Hispanic white, non-Hispanic black, Hispanic, and Asian/Pacific Islander/other. We defined education as college/postgraduate degree, some college/junior college, high school graduate, less than high school graduate, and unknown. Additional independent variables at diagnosis (categorized as indicated in Table 1) included age, employment, insurance, comorbidity (based on Charlson and Katz scoring systems23,24 from medical records and patient surveys), institution (blinded, labeled A through H), diagnosis year, disease stage, tumor grade, and hormone receptor status.

Table 1. Patient Characteristics by Race/Ethnicity in Cohort 1
 No. of Patients (%) 
CharacteristicOverall, n = 1109White, n = 837Black, n = 93Hispanic, n = 103Asian/Pacific Islander/Other, n = 76Pa
  • Abbreviations: NE, not estimable.

  • a

    P values were calculated with the chi-square test for characteristics that were associated with race/ethnicity.

  • b

    This value was NE because of small cell counts.

  • c

    Hormone receptor status was defined as positive if tumors were positive for either estrogen receptor or progesterone receptor and was defined as negative if tumors were negative for either estrogen receptor or progesterone receptor and were not positive for either receptor.

Age at diagnosis, y     .17
 <50496 (45)363 (43)38 (41)52 (50)43 (57) 
 50-59325 (29)241 (29)32 (34)29 (28)23 (30) 
 60-69183 (17)147 (18)13 (14)16 (16)7 (9) 
 ≥70105 (9)86 (10)10 (11)6 (6)3 (4) 
Education     <.0001
 College/graduate degree404 (36)314 (38)26 (28)28 (27)36 (47) 
 Some college218 (20)165 (20)28 (30)18 (17)7 (9) 
 High school degree161 (15)118 (14)12 (13)22 (21)9 (12) 
 <High school54 (5)24 (3)6 (6)21 (20)3 (4) 
 Other/unknown272 (25)216 (26)21 (23)14 (14)21 (28) 
Employment     <.0001
 Employed/student577 (52)444 (53)46 (49)51 (50)36 (47) 
 Homemaker178 (16)130 (16)6 (6)30 (29)12 (16) 
 Unemployed69 (6)36 (4)15 (16)12 (12)6 (8) 
 Retired153 (14)125 (15)18 (19)6 (6)4 (5) 
 Other/unknown132 (12)102 (12)8 (9)4 (4)18 (24) 
Insurance     NEb
 Managed784 (71)615 (73)57 (61)60 (58)52 (68) 
 Indemnity34 (3)30 (4)1 (1)0 (0)3 (4) 
 Medicaid/indigent/self-pay107 (10)45 (5)16 (17)31 (30)15 (20) 
 Medicare167 (15)134 (16)17 (18)11 (11)5 (7) 
 Other/unknown17 (2)13 (2)2 (2)1 (1)1 (1) 
No. of comorbidities     NEb
 0913 (82)695 (83)76 (82)77 (75)65 (86) 
 1148 (13)105 (13)14 (15)21 (20)8 (11) 
 ≥248 (4)37 (4)3 (3)5 (5)3 (4) 
Year of diagnosis     .21
 2005115 (10)93 (11)8 (9)8 (8)6 (8) 
 2006399 (36)296 (35)40 (43)41 (40)22 (29) 
 2007383 (35)285 (34)33 (35)39 (38)26 (34) 
 2008212 (19)163 (19)12 (13)15 (15)22 (29) 
Disease stage     .24
 I422 (38)326 (39)30 (32)34 (33)32 (42) 
 II418 (38)322 (38)37 (40)36 (35)23 (30) 
 III269 (24)189 (23)26 (28)33 (32)21 (28) 
Tumor grade     NEb
 High749 (68)549 (66)68 (73)80 (78)52 (68) 
 Low/intermediate325 (29)258 (31)22 (24)21 (20)24 (32) 
 Unknown35 (3)30 (4)3 (3)2 (2)0 (0) 
Hormone receptor statusc     .03
 Positive684 (62)304 (36)41 (44)40 (39)40 (53) 
 Negative425 (38)533 (64)52 (56)63 (61)36 (47) 

Statistical Analysis

We compared rates of trastuzumab receipt and completion by race/ethnicity, education, and other patient characteristics using individual chi-square tests. We used 2 logistic regression models to assess the probability of 1) receipt of trastuzumab and 2) completion of trastuzumab by race/ethnicity and education, adjusting for the independent variables listed above. We used generalized estimating equations to account for clustering at the institutional level.

We performed multiple sensitivity analyses for Cohort 2. First, we repeated the analyses restricted to patients with stage II and III disease, because most of the adjuvant clinical trials with trastuzumab included higher risk patients, and some patients with stage I disease may have been treated differently. Second, we repeated the analyses after redefining the completion of trastuzumab as receipt of ≥ 350 days of treatment among patients who had ≥ 18 months of follow-up. Third, because 59 of 62 women without documented treatment end dates were defined as having completed trastuzumab, we repeated the analyses assuming that all 62 of these women did not complete therapy. Finally, because the development of cardiotoxicity can have an impact on treatment, we repeated the models with inclusion of a binary variable for anthracycline receipt.

All reported P values were 2-sided, and statistical analyses were conducted using SAS version 9.2 (SAS Institute, Cary, NC). Because analyses used previously collected data without identifying patient information, the study was considered exempt from review by the Institutional Review Board at Dana-Farber Cancer Institute.


Patient and tumor characteristics for Cohort 1 are presented in Table 1. Most women were white (75%), 8% were black, 9% were Hispanic, and 7% were Asian/Pacific Islander/other race/ethnicity. Approximately 36% of women completed college/graduate school, whereas 20% of women attained a high school degree or less. There were no significant differences in age, comorbidity, diagnosis year, stage, or tumor grade by race/ethnicity, although white women had the highest percentage of hormone receptor-negative cancers. Overall, white women and those of Asian/Pacific Islander/other race had higher educational attainment than black and Hispanic women. White women were most likely to be insured by managed care; whereas black, Hispanic, and Asian/other women had higher rates of Medicaid/indigent/self-pay coverage.

Among the 1109 women in Cohort 1, 83% initiated trastuzumab. Unadjusted rates of trastuzumab receipt did not differ by race/ethnicity (range, 82%-90%; P = .151) (Fig. 1, top) or education (range, 81%-85%; P = .91) (Fig. 1, bottom). In adjusted analyses, we did not observe racial/ethnic or education differences in treatment initiation (Table 2). With regard to other variables, women aged ≥ 70 years (vs < 50 years), those with the highest (vs the lowest) comorbidity, and those with lower/intermediate (vs higher) tumor grade had lower odds of treatment receipt; whereas women who were diagnosed during 2006 to 2008 (vs 2005) and those who had stage II and III disease (vs stage I disease) had higher odds of treatment receipt. Compared with the reference institution, all but 1 institution had lower odds of trastuzumab initiation (Table 2).

Figure 1.

These charts illustrated the unadjusted percentages of trastuzumab receipt and completion by (Top) race/ethnicity and (Bottom) education. Overall P values were calculated with the chi-square test.

Table 2. Unadjusted Proportions and Adjusted Odds for Receipt and Completion of Trastuzumab
VariableNo. of Patients Receiving Trastuzumab (%), N = 1109Unadjusted PaAdjusted OR for Receipt of Trastuzumab [95% CI]bNo. of Patients Completing Trastuzumab (%), N = 888Unadjusted PaAdjusted OR for Completion of Trastuzumab Among Women Initiating Treatment [95% CI]b
  • Abbreviations: CI, confidence interval; HS, high school; OR, odds ratio.

  • a

    Differences in unadjusted percentages of the receipt and completion of treatment were examined using chi-square tests.

  • b

    Generalized estimating equations were used to account for clustering at the institution level, and analyses were adjusted for all variables in the table.

  • c

    Hormone receptor status was defined as positive if tumors were positive for either estrogen receptor or progesterone receptor and was defined as negative if tumors were negative for either estrogen receptor or progesterone receptor and were not positive for either receptor.

  • d

    Institutions are deidentified for this analysis, and sample sizes are not provided.

Race/ethnicity .151  .007 
 White687 (82) 1.00576 (87) 1.00
 Black80 (86) 1.02 [0.65-1.59]53 (73) 0.45 [0.27-0.74]
 Hispanic93 (90) 1.40 [0.86-2.28]76 (85) 1.40 [0.79-2.48]
 Asian/Pacific Islander/other65 (86) 1.06 [0.44-2.54]57 (90) 1.78 [0.99-3.21]
Age at diagnosis, y < .0001  .022 
 <50445 (90) 1.00372 (88) 1.00
 50-59278 (86) 0.75 [0.39-1.44]226 (85) 0.79 [0.57-1.10]
 60-69145 (79) 0.60 [0.36-1.01]124 (88) 1.12 [0.42-3.01]
 ≥7057 (54) 0.18 [0.08-0.39]40 (73) 0.58 [0.13-2.52]
Education .911  .006 
 College/graduate degree341 (84) 1.00294 (90) 1.00
 Some college182 (83) 0.94 [0.67-1.32]151 (87) 0.71 [0.37-1.34]
 HS degree131 (81) 0.84 [0.58-1.23]102 (82) 0.50 [0.33-0.77]
 >HS degree46 (85) 1.58 [0.98-2.52]31 (70) 0.27 [0.14-0.51]
 Other/unknown225 (83) 1.54 [0.62-3.85]184 (85) 0.72 [0.38-1.36]
Employment .002  .492 
 Employed/student491 (85) 1.00415 (88) 1.00
 Homemaker148 (83) 0.89 [0.69-1.15]114 (83) 0.79 [0.53-1.17]
 Unemployed61 (88) 1.11 [0.56-2.21]49 (86) 1.30 [0.36-4.78]
 Retired110 (72) 1.37 [0.96-1.94]87 (81) 0.89 [0.32-2.48]
 Other/unknown115 (87) 1.82 [0.50-6.69]97 (87) 1.27 [0.48-3.32]
Insurance < .0001  .502 
 Managed682 (87) 1.00571 (87) 1.00
 Indemnity28 (82) 0.78 [0.32-1.88]20 (80) 0.85 [0.44-1.65]
 Medicaid/indigent81 (90) 0.74 [0.41-1.36]64 (83) 0.99 [0.62-1.59]
 Self-pay13 (76) 0.42 [0.17-1.05]11 (85) 0.75 [0.29-1.99]
 Medicare107 (64) 0.59 [0.32-1.06]84 (81) 0.89 [0.34-2.35]
 Other/unknown14 (82) 0.78 [0.18-3.40]12 (86) 0.53 [0.24-1.15]
No. of comorbidities .0008  .112 
 0779 (85) 1.00648 (86) 1.00
 1112 (76) 0.77 [0.51-1.16]89 (86) 1.27 [0.69-2.34]
 ≥234 (71) 0.50 [0.26-0.97]25 (74) 0.60 [0.28-1.28]
Year of diagnosis .012    
 200584 (73) 1.0068 (85).8031.00
 2006343 (86) 2.28 [1.20-4.32]282 (85) 1.08 [0.58-1.99]
 2007322 (84) 2.12 [1.32-3.41]257 (86) 1.21 [0.63-2.35]
 2008176 (83) 2.40 [1.08-5.32]155 (88) 1.20 [0.61-2.34]
Disease stage <.0001  .040 
 I287 (68) 1.00247 (89) 1.00
 II384 (92) 4.98 [3.95-6.29]319 (86) 0.76 [0.46-1.27]
 III254 (94) 6.77 [4.77-9.60]196 (82) 0.64 [0.34-1.21]
Tumor grade <.0001  .586 
 High659 (88) 1.00545 (86) 1.00
 Low/intermediate242 (75) 0.49 [0.29-0.82]197 (84) 0.71 [0.35-1.43]
 Unknown24 (69) 0.38 [0.21-0.68]20 (91) 1.71 [0.70-4.19]
Hormone receptor statusc .0007  .052 
 Positive550 (80) 0.80 [0.56-1.14]463 (88) 1.41 [0.75-2.66]
 Negative375 (88) 1.00299 (83) 1.00
Institutiond .0008  .008 
 A— (90) 1.00— (82) 1.00
 B— (85) 0.82 [0.70-0.98]— (83) 1.09 [0.80-1.48]
 C— (79) 0.58 [0.49-0.69]— (90) 1.45 [1.24-1.70]
 D— (78) 0.57 [0.35-0.93]— (89) 1.79 [1.36-2.37]
 E— (72) 0.46 [0.37-0.56]— (87) 1.74 [1.38-2.20]
 F— (84) 0.62 [0.40-0.95]— (77) 0.63 [0.46-0.88]
 G— (85) 0.91 [0.73-1.14]— (92) 2.78 [2.29-3.39]
 H— (85) 0.70 [0.59-0.84]— (94) 2.95 [2.42-3.60]

Among the 888 women in Cohort 2, 86% completed ≥ 270 days of treatment with a median treatment duration of 350 days (range, 2-1261 days). The rates of treatment completion were lowest for black women compared with other women (Fig. 1, top) and for women with less than a high school education (Fig. 1, bottom). In adjusted analyses (Table 2), black women had significantly lower odds of completing trastuzumab compared with white women (odds ratio [OR], 0.45; 95% confidence interval [CI], 0.27-0.74). In addition, we observed differences in therapy completion by education; women who had no high school diploma and those who had a high school diploma but no college education had lower odds of completion compared with those who had a college degree (OR, 0.27 [95% CI, 0.14-0.51] and 0.50 [95% CI, 0.33-0.77], respectively). We observed no differences in the odds of completing trastuzumab therapy among women in Cohort 2 by age, comorbidity, diagnosis year, stage, or tumor grade; however, we did observe institutional variation: Five centers had higher odds, 1 center had lower odds, and 1 center had similar odds of completing therapy compared with the reference institution (Table 2).

In a sensitivity analysis for the second cohort, after it was restricted to patients with stage II or III disease (n = 612), the results for race/ethnicity and education were similar to the initial model (black women vs white women: OR, 0.45; 95% CI, 0.23-090; less than high school education vs college degree: OR, 0.27; 95% CI, 0.14-0.51). In an analysis restricted to patients who had 18 months of follow-up (n = 731) with trastuzumab completion defined as treatment for > 350 days, black women had persistently lower odds of completion compared with white women (OR, 0.47; 95% CI, 0.31-0.72), although the findings for education were no longer statistically significant (OR for less than high school education, 1.12 [95% CI, 0.62-2.02]; OR for high school diploma, 0.86 [95% CI, 0.64-1.16] both vs college degree). In the model that categorized 62 women without a completion date as having not completed trastuzumab, the results also were similar to the original analysis (data not shown). Finally, in the model with the anthracycline variable, the findings for race and education were unchanged (data not shown). Employment was not significant in any of the models performed.


We examined the receipt of trastuzumab, an efficacious yet costly and time-intensive treatment, by race/ethnicity and education in a large cohort of women with HER2-positive breast cancer who received treatment at NCCN centers across the United States. We observed no disparities in initiation of trastuzumab by race/ethnicity and education but found differences in treatment completion, with black women and those with lower educational attainment having significantly lower odds of therapy completion compared with white and college-educated women. Our findings were robust to multiple sensitivity analyses. To our knowledge, this analysis represents the first examination of differences in the receipt of trastuzumab.

Given the substantial improvement in outcomes from receipt of trastuzumab for HER2-positive breast cancers, the overall high rates of receipt and completion of therapy observed in the current study were reassuring, suggesting substantial agreement with treatment recommendations among practicing oncologists.4 Moreover, despite the burden of ongoing infusional treatments, 86% of women who initiated trastuzumab completed ≥ 270 days of therapy. This is higher than estimates of completing standard infusional chemotherapy among patients who initiated adjuvant chemotherapy for colon cancer, in which studies have demonstrated variable 6-month completion rates of 58% to 78%, with lower rates of completion among older patients and those with greater comorbidity.25-27 Nevertheless, despite the relatively high rates of trastuzumab completion in our analysis, the disparities we observed raise concerns. Although it is possible that patients still may benefit from shorter durations of trastuzumab,28 this has not been well studied, and guidelines recommend 1 year of therapy for all patients.4

Possible explanations for the differences in rates of treatment completion include toxicity, variable patient preferences, institutional variations in toxicity assessments, and provider and patient adherence, which may be amplified by the need for frequent infusions over a prolonged period. Previous studies have suggested that black women may have an increased risk of cardiotoxicity with anthracycline chemotherapy,29-32 although detailed toxicity assessments for minority patients receiving trastuzumab have not been reported. Lower adherence also has been reported for nonwhite women and women of lower SES who initiate adjuvant hormone therapy,33-35 in which 5 years of adjuvant therapy is recommended.4 Barriers to long-term adherence, including patient/family preferences to discontinue therapy early, may be particularly relevant for women of lower SES, who may have more difficulty freeing themselves from work and childcare commitments. Although we observed lower odds of treatment completion for those who did not graduate high school, we did not observe differences in completion of treatment by employment. However, this may have resulted from our inability to distinguish types of jobs among employed women. We also did not observe differences in receipt or completion of treatment by insurance, suggesting broad insurance coverage for trastuzumab. It is noteworthy that our cohort reflected a group that sought care at a NCCN center and included very few patients without insurance.

Although specific barriers to completion of care could not be elucidated in our study, our results suggest that the development of interventions may be necessary to assure completion of therapy for women who do not experience serious toxicity. Because of the need to continuously calculate the number of trastuzumab infusions a patient has received, enhanced electronic medical records with provider reminders for a patient's expected end date of therapy may help to assure that all doses are delivered. Strategies also may focus on patient adherence, with patient reminders and/or provider alerts if a patient misses an infusion appointment. Enhanced patient education about the benefits of trastuzumab and the expected duration of therapy may be particularly useful for those with lower educational attainment.

We observed lower rates of treatment receipt for women with lower risk cancers, for older women, and for women with greater comorbidity. These findings were not surprising given the uncertainty about the optimal systemic treatment strategy for lower risk disease and concerns about toxicity, competing risks, or perceived lower benefits of treatment in older and sicker populations. Although trastuzumab is well tolerated in most women, it is initially administered with chemotherapy. Consequently, decisions to treat older and sicker women are likely based on concerns about tolerance of chemotherapy rather than trastuzumab. Other studies have demonstrated lower rates of adjuvant chemotherapy for older women with breast cancer,36,37 but few data are available regarding the rates of trastuzumab receipt for women of various ages.38,39 Among women who initiated trastuzumab in our study, we observed no differences in the odds of completing therapy by age or comorbidity; thus, providers may be selecting patients to receive treatment who will optimally tolerate it.

The strengths of this analysis include our ability to examine detailed information on the receipt of infusional treatments for a large cohort of women with information on HER2 status and individualized information on employment and education: variables that are not currently available in other cancer registries. However, we acknowledge several limitations. First, the care patterns in NCCN centers may not be generalizable to other settings, and the numbers of minority women were relatively small. Nevertheless, assessment of care at academic centers is of interest. Second, although NCCN data on trastuzumab duration may be imperfect, our findings were robust to sensitivity analyses and rigorous data-quality checks. Third, detailed information on why trastuzumab was stopped was not available, nor did analyses account for other factors that may have influenced treatment, such as social supports, marital status, or distance to the medical center. Finally, because of the small numbers of women in some subgroups (ie, employment, insurance), we may have had insufficient power to detect small differences in the outcomes of interest.

In conclusion, we observed differences in completion, but not initial receipt, of adjuvant trastuzumab by race and education. Further work is needed to better understand differences in adherence, reasons for premature cessation, and differences in tolerance to therapy. With the expected substantial improvement in outcomes for women who receive 1 year of adjuvant trastuzumab, improved and appropriate administration of this agent may have a significant impact on disease recurrence and survival for women with HER2-positive breast cancer. Interventions aimed at assuring the receipt of guideline care for all women with HER2-positive breast cancers should be a priority.


This work was funded in part by the Susan G. Komen for the Cure Foundation and by grant P50 CA89393 from the National Cancer Institute to Dana-Farber Cancer Institute.


The authors made no disclosures.