Secondary Logo

Journal Logo

Women With Breast Cancer in the Veterans Health Administration

Demographics, Breast Cancer Characteristics, and Trends

Colonna, Sarah MD*,†,‡; Halwani, Ahmad MD*,†,‡; Ying, Jian PhD*,†,‡; Buys, Saundra MD*,†; Sweeney, Carol PhD*,†

doi: 10.1097/MLR.0000000000000299
Complex Chronic Conditions

An increasing number of women are being cared for within the Veterans Health Administration (VA). However, the demographics and trends of women with breast cancer at the VA has not been documented. We describe the demographics and breast cancer characteristics of the 4445 women enrolled in the VA and reported to the Department of Veterans Affairs Central Cancer Registry diagnosed with breast cancer from 1995 to 2012. The cases of breast cancer per year increased over time to 365 in 2012. Black women represented only 16% of women diagnosed with breast cancer in the VA in 1995–1999 but increased to 25% by 2010–2012 (P<0.001). The median age at diagnosis in 1995–1999 was 58.4 and decreased to 56.8 by 2010–2012 (P<0.02). The fraction of breast cancers that were node negative was 45% in 1995–1999 and increased to 64% in 2010–2012; correspondingly, women presented at an earlier stage in more recent years (P<0.001). Urban women with breast cancer cared for within the VA are more likely to be younger (P=0.04) and nonwhite (P<0.001) compared with rural women, but the breast tumor characteristics appear similar. Oncology physicians at the VA must be prepared to care for breast cancer among women as the number of cases is growing. With only 365 women diagnosed with breast cancer at the VA as per year 2012 and nearly 150 treating VA facilities, the number of breast cancer patients seen by a particular physician could be quite low, and this fact suggests a need for an evaluation of the quality and outcomes of breast cancer care at the VA.

*Huntsman Cancer Institute

Department of Internal Medicine, University of Utah

George E. Whalen Veterans Health Administration, Salt Lake City, UT

A.H.: Salary support from industry sponsored trials (KHK, BMS, SGN). S.B.: Grant support through NCI. C.S.: NIH URR025764/UL1TR001067, R01CA129059, R01DK099098, HHSN267200700015C, HHSN267200700015C, KM1-CA1567, HHSN261201000026, PhRMA Foundation Mentored Scholars and Professional Education Program Award. The authors declare no conflict of interest.

The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.

Reprints: Sarah Colonna, MD, Department of Hematology/Oncology, University of Utah, 500 Foothill Dr., Salt Lake City, UT 84148. E-mail:

Women comprise a rapidly growing proportion of patients at the Veterans Health Administration (VA). There are now 2.2 million women veterans residing in the United States, representing nearly 10% of all living veterans, and in 2011, women constituted nearly 15% of the active duty-armed forces.1–3 The number of women veterans enrolled at the VA nearly doubled from 159,630 to 292,921 from 2000 to 2009, thus only 32% of women veterans were enrolled at the VA and further, only about 19% of women veterans utilized the VA for healthcare.4 A recent publication documented a 47% increase in women seen as new patients within the VA outpatient setting between 2003 and 2009, and 65% of those women were below the age of 45 years.5

Previous comparisons of men and women veterans demonstrate that women veterans receiving VA care are more likely to be an ethnic minority, have low income and poor health status.6,7 A survey of women veterans revealed that women who did not use the VA for medical care reported a greater perception that providers were not skilled in women’s healthcare issues.8 In the past decade, the VA has placed a high priority on women’s health and has developed Women’s Health Centers to care for women in a more consistent sex-oriented manner.9,10

Access to care for rural women veterans is also a concern as rural women with breast cancer in other healthcare systems have been shown to present at a later stage and have decreased access to subspecialized breast cancer care.11,12 Rural women veterans receive less sex-specific care within the VA but it is unknown whether rural women receive more non-VA medical care than urban women.13–15 Within areas of women’s health such as obstetrics and gynecology at the VA, there have been clearly documented increases in the volume of women and the clinical care being delivered has increased.16 However, the care of women with breast cancer requires specialized multidisciplinary coordination, and the trends for treatment of breast cancer within the VA system have not been formally documented.

In this paper, we describe the demographics and trends of women diagnosed with invasive breast cancer who are enrolled at the VA over the past 2 decades and discuss the implications for VA healthcare and research needs for this population.

Back to Top | Article Outline


The Department of Veterans Affairs Central Cancer Registry (VACCR) has collected information about all cancer diagnoses from the 132 hospital-based cancer registries within the VA since 1995.17 After Institutional Review Board and Veterans Affairs Research and Development approval, data about all women enrolled at the VA who were diagnosed with breast cancer within the VACCR from 1995 until 2012 was obtained. Administrative VA data were also accessed to obtain the veteran/nonveteran status of women.

The dataset from VACCR contained patient date of birth, ethnicity, zip code of patient residence at the time of diagnosis, and VA designation of rural or urban status. There were 29 separate categories for race and ethnicity and 8 further categories defining Hispanic origin. Rural/urban data were divided into 9 population categories ranging from the most rural, defined as a county with fewer than 2500 residents and not adjacent to a metropolitan area, to the most urban category, defined as a county within a metropolitan area with more than 1 million people. The registry also included details about the diagnosis of breast cancer including date of diagnosis, histology subtype, tumor size in millimeters, lymph node status (positive or negative), American Joint Committee on Cancer stage at the time of diagnosis, and estrogen receptor (ER) and progesterone status in positive or negative format. Human epidermal growth factor 2 status was not obtained.

The VACCR also gathered treatment information including whether adjuvant chemotherapy was administered, the specific chemotherapy regimen, date of chemotherapy administration, adjuvant hormone therapy (yes or no), and the date hormone therapy was initiated. All the VACCR data are obtained exclusively from within the VA electronic medical record system, with new cases identified by ICD diagnoses, pathology, radiology, and clinical records. VA cancer registrars capture care administered by outside facilities only if it is documented by VA clinicians within the VA electronic medical record system.

Women with histologic codes for noninvasive cancer or histologies that were not carcinoma were excluded. Tumor grades 3 and 4 were collapsed into 1 category denoting high grade because grade 4 is not a widely used category and there were very few grade 4 tumors. ER status that was coded as borderline was included in the ER-positive category, because an ER of 1% or more is now considered to be positive, and there is no longer a borderline category as there was historically. The 37 categories of race and ethnicity were collapsed into the broader categories of white, black, Hispanic/Latino, and other to protect patient identity as there were small numbers in several ethnic categories. Women were defined as urban if they lived in a county within a metropolitan area and all other women were defined as rural.

We conducted descriptive analyses, compiling contingency tables and calculating percent distributions. The temporal trends of subject and tumor characteristics were tested by different appropriate regression models continuously over each individual year of diagnosis according to class of the dependent variables. Specifically, patient age was tested by linear regression; tumor size was tested by quantile regression. Dichotomous variables including rural or urban, lymph node status, ER status, and adjuvant hormone and chemotherapy treatment rates were tested by logistic regression. Ethnicity and histology were tested by multinomial logistic regression and ordinal variables including grade and stage were tested by ordinal logistic regression. Rural and urban populations were compared using logistic regression. Observations that were “unknown” or missing for the specific variable tested were excluded from the analysis for that particular statistical test. The data were analyzed with STATA statistical software.18

Back to Top | Article Outline


There were initially 6111 women identified by VACCR diagnosed with breast pathology from 1995 to 2012. Nonepithelial malignancies (including sarcoma and lymphoma, n=79) and in situ cancers listed either as noninvasive histology codes (n=1028) or by extent of disease codes (n=151) were excluded. There were 408 records excluded because they were duplicated entries for individual women. We retained only the tumor information for each woman with the earliest date of diagnosis for analysis, leaving 4445 women for our analysis. The majority of women within the VACCR cohort were veterans, at 78%, with only 5% documented as nonveterans, with the remaining 17% of women with missing data for veteran status. Demographic information was nearly complete; however, there was significant data missing for tumor-related information, such as tumor size and grade, lymph node status, and ER status (Tables 1, 2).





The number of women diagnosed with breast cancer increased from 117 diagnoses in 1995–365 in 2012 (Fig. 1). The distribution of race and ethnicity of women with breast cancer at the VA changed over time (P<0.001). Most notably, there were significantly more black women cared for at the VA for breast cancer in recent years, a growth from 16% in 1995–1999 to 25% in 2010–2012 (Table 1, Fig. 2).





The age of breast cancer diagnosis has decreased over time (P=0.02). During the late 1990s, 24% of breast cancer diagnoses occurred among women aged 70–79 years, contrasted with only 7% in the 2010s. The increasing number of diagnoses over time seems to be occurring mostly among women in their 50s and 60s (Table 1, Fig. 2). There has been a significant shift from rural to urban in the population of women with breast cancer at the VA. More cases were drawn from rural regions of the United States in the 1990s (21%) than in the 2010s (18%, P=0.04; Table 1).

The reported tumor characteristics of breast cancer among women diagnosed at the VA have changed over time. In recent years, the VACCR classified breast cancer as invasive ductal carcinoma more often, whereas the proportion described as other or unspecified types of carcinomas declined (P<0.001; Table 2). The change in distribution of size of the primary breast tumor over time did not reach significance (P=0.10), however, tumors were less likely to involve lymph nodes in recent years; after excluding tumors with missing node status, 65% were node negative in 2010–2012 increasing from 58% in 1995–1999 (P<0.001). Lymph node status within the VACCR was coded as unknown for 22% of cancers diagnosed in the 1990s but unknown for only 2% of cancers diagnosed in 2010–2012.

Corresponding to the decreasing lymph node involvement over time, breast cancers in recent years are more likely to be diagnosed at a lower American Joint Committee on Cancer stage than those diagnosed in the 1990s and early 2000s. In particular, the number of stage 1 diagnoses has increased from 35% in 1995–1999 to 45% in 2010–2012 (P<0.001). The grading of breast tumors by pathologists has not changed significantly over time, although the VACCR data for this category is more completely documented in recent years. The number of ER-positive breast tumors has increased over time, from 47% in 1995–1999 to 65% in 2010–2012 (P=0.05; Table 2).

Rates of adjuvant hormone therapy administration did not change significantly over time (P=0.31); however, reported chemotherapy administration in the adjuvant setting increased substantially from 10% in 1995–1999 to 46% in 2010–2012 (Table 2; P<0.001).

Women living in urban areas of the United States diagnosed with breast cancer were more likely to have been diagnosed at a younger age (57 compared with 60 for rural women, P=0.04), and urban women were significantly more likely to be an ethnic minority (31% urban nonwhite vs. 13% rural nonwhite, P<0.001). Breast cancer characteristics of women (lymph node status, stage of breast cancer, and ER status) did not differ significantly between rural and urban women (P=0.98, 0.12, 0.9, respectively) (Table 3).



Back to Top | Article Outline


The number of women diagnosed with breast cancer and receiving oncologic care at the VA has increased steadily over time, which is likely a result of multiple factors. Increasing numbers of women serving in the armed forces in recent years has contributed to the change in the face of breast cancer at the VA. Although representing only an estimated 0.1% of the 226,870 breast cancer diagnoses in the United States in 2012, the number of women diagnosed with breast cancer at the VA has more than tripled from 1995 to 2012 and from 2000 to 2009 increased by about 60%, which is similar to the 83% increase in all women enrolled at the VA from 2000 to 2009.2,19 The caseload at the VA would also be affected by women’s decisions about where to receive their medical care as not all women elect to receive care at the VA shown by a 2003 national survey of women veterans, when only about one third of all women veterans were enrolled at the VA and receiving medical care there.20

The number of black women diagnosed with breast cancer within the VA has markedly increased in recent years from this analysis of VACCR data, from 16% to 25% from 1995 to 2012, and this is likely a result of increasing numbers of black women serving in the US Armed Forces in recent years. The percentage of women veterans who were black in 2000 was 17%, and 1 decade later, in 2010, that figure nearly doubled to 31% of women in active-duty military being black. The commensurate general population of black women in the United States aged 18–44 years in 2010 was only 15%.21,22

In addition, the age of diagnosis of breast cancer within the VA population has decreased over time. The largest portion of women veterans alive, at 27.7%, served during Operation Enduring Freedom/Operation Iraqi Freedom underscoring the young age of many women veterans.4 In comparison, Surveillance Epidemiology and End Results data reports that for the years 2010–2011 the proportion of incident breast cancers diagnosed in women in their 70s and 80s was 18% and 12%, respectively, data described here from the VACCR for 2010–2012, reported, 7% for women in their 70s and only 4% for women in their 80s.23 The real and proportional increase among women veterans aged 45–64 (30% to 44%) and the proportional decrease in veterans over age 65 (19% to 14%) documented from 2000 to 2009 likely accounts for much of this difference.2 In addition, the VA has documented very high mammography rates, at 87% of eligible women in 2011, which could be another reason for the younger age at diagnosis within the VA.24

We observed a stage shift of breast cancer at the VA over time, with fewer women diagnosed with lymph node involvement, a corresponding lower stage at diagnosis, and a larger proportion of tumors that were ER positive. This change is similar to what has been observed for breast cancer among the general population, with increased obesity in postmenopausal women being associated with more indolent ER-positive tumors, and increasing mammography rates diagnosing breast cancers at an earlier stage.25

The percent of women reported as receiving adjuvant chemotherapy for breast cancer at the VA has increased greatly over time. In general, younger women are more likely to receive adjuvant chemotherapy than older women and, as documented above, women treated within the VA system are currently younger than previously, thus there could be a real increase in the use of chemotherapy in VA breast cancer patients. Capture of chemotherapy and hormonal treatment in cancer registry records is recognized to be incomplete.26–28 For example, the cancer registry chemotherapy variable is estimated to be 68% sensitive relative to Medicare claims.29 Therefore, it should be considered that some or all of the trend with time could be a result of a change in reporting.

Women with breast cancer at the VA became more urban by zip code-classification over time, which mirrors the overall changes of the US population.30 Further, black women are much more likely to live in urban areas, and the increase in black women with breast cancer at the VA occurred primarily within urban areas. Our comparison of rural and urban status among women diagnosed with breast cancer at the VA revealed that urban women are more likely to be younger and of an ethnic minority, consistent with known differences between rural and urban women in the general population, however, the presenting characteristics of breast cancer were not different between these 2 populations.31 The outcomes of breast cancer among rural women should be further evaluated, considering both the region where the patient lives and the distance from the closest hospital providing subspecialty care.

This description of breast cancer at the VA is limited by several factors. First, it relies nearly exclusively on 1 source of data, the VACCR, which does not include all information of interest to describe cancer patients, treatment, and outcomes. However, cancer registry data, in general, are considered highly complete with respect to case ascertainment and valid for cancer diagnosis and patient demographics.32,33

Second, women receiving care at the VA may receive portions of their care at various non-VA facilities, adding a layer of difficulty when interpreting VA datasets. There are 3 main arms of breast cancer treatment—radiation, surgery, and medical oncology. Nearly all radiation oncology for cancer treatment of VA patients is provided at outside facilities by non-VA care, however, the majority of surgical and medical treatment for breast cancer is provided by VA physicians. Luther et al34 estimated that for women within the VACCR from 2000 to 2006, only about 10% received breast cancer surgery at an outside facility. And although exact numbers of women treated by medical oncology for breast cancer outside the VA are unknown, most women are likely treated adjuvantly with chemotherapy or hormone therapy by VA medical oncologist as there are 134 VA facilities with cancer treating capabilities, which includes by necessity a medical oncologist.17 As there are so few women diagnosed with breast cancer at the VA per year (365 in 2012) and there are nearly 150 treating VA facilities, the number of breast cancer patients seen by a particular physician could be quite low, and this fact calls for an evaluation of the quality of care and outcomes of the women being treated for breast cancer at the VA.

Breast cancer treatments and outcomes need to be further explored with attention to patient comorbidities and breast cancer volume of the treating facility. To properly compare those outcomes, it will be necessary to utilize multiple VA datasets to ensure data completeness and accuracy and to appropriately control for the many confounding factors within the population of women receiving care at the VA. Methods have been developed to obtain and validate information through VA electronic medical records, including use of text fields and pharmacy data.35,36 Further research on cancer treatment within the VA should include validating methods to obtain information on adjuvant treatment, comorbidities, and functional status to better evaluate treatment trends and receipt of guideline-concordant therapies.

Given the rapidly increasing number of women at the VA, it is key that both the clinical care and the research goals include specialization in breast cancer care and validation of methods to evaluate breast cancer outcomes. This description of women with breast cancer at the VA from 1995 to 2012 is an initial step necessary to define parameters by which breast cancer at the VA can be judged to improve both the quality of medical care and outcomes for women with breast cancer.

Back to Top | Article Outline


The authors acknowledge biostatistics core resources support from NIH 1ULTR001067 and Huntsman Cancer Institute P30CA042014. The authors acknowledge the VACCR for allowing access to the dataset for breast cancer at the VHA hospitals and for aiding them in procurement and definition of the data. This material is the result of work supported with resources from the VA Informatics and Computing Infrastructure (VINCI) and the use of facilities at the George E. Whalen VA Medical Center.

Back to Top | Article Outline


1. Department of Veterans Statistics at a Glance: Prepared by the National Center for Veterans Analysis and Statistics, 2014. Available at: Accessed December 19, 2014.
2. Frayne SM, Phibbs CS, Friedman SA, et al.. Sourcebook: Women Veterans in the Veterans Health Administration Volume 1 Sociodemographic Characteristics and Use of VHA Care. 2010.Washington, DC: Women’s Health Evaluation Initiative, Women Veterans Health Strategic Health Care Group, Veterans Health Administration, Department of Veterans Affairs.
3. Burrelli DF. Women in combat: issues for Congress. CRS Report for Congress. 2013.
4. National Center for Veterans Analysis and StatisticsAmerica’s Women Veterans: Military Service History and VA Benefit Utilization Statistics. 2011.Washington, DC: National Center for Veterans Analysis and Statistics, Department of Veterans Affairs.
5. Friedman SA, Phibbs CS, Schmitt SK, et al.. New women veterans in the VHA: a longitudinal profile. Womens Health Issues. 2011;21supplS103–S111.
6. Frayne SM, Yu W, Yano EM, et al.. Gender and use of care: planning for tomorrow’s Veterans Health Administration. J Womens Health (Larchmt). 2007;16:1188–1199.
7. Skinner K, Sullivan LM, Tripp TJ, et al.. Comparing the health status of male and female veterans who use VA health care: results from the VA Women’s Health Project. Women Health. 1999;29:17–33.
8. Washington DL, Yano EM, Simon B, et al.. To use or not to use. What influences why women veterans choose VA health care. J Gen Intern Med. 2006;21suppl 3S11–S18.
9. Bean-Mayberry B, Yano EM, Bayliss N, et al.. Federally funded comprehensive women’s health centers: leading innovation in women’s healthcare delivery. J Womens Health (Larchmt). 2007;16:1281–1290.
10. Yano EM, Goldzweig C, Canelo I, et al.. Diffusion of innovation in women’s health care delivery: the Department of Veterans Affairs’ adoption of women’s health clinics. Womens Health Issues. 2006;16:226–235.
11. Onitilo AA, Liang H, Stankowski RV, et al.. Geographical and seasonal barriers to mammography services and breast cancer stage at diagnosis. Rural Remote Health. 2014;14:2738.
12. Nguyen-Pham S, Leung J, McLaughlin D. Disparities in breast cancer stage at diagnosis in urban and rural adult women: a systematic review and meta-analysis. Ann Epidemiol. 2014;24:228–235.
13. Weeks WB, Wallace AE, West AN, et al.. Research on rural veterans: an analysis of the literature. J Rural Health. 2008;24:337–344.
14. Brooks E, Dailey N, Bair B, et al.. Rural Women Veterans demographic report: defining VA users’ health and health care access in rural areas. J Rural Health. 2014;30:146–152.
15. Nayar P, Yu F, APentang B. Improving care for rural veterans: are high dal users different? J Rural Health. 2014;30:139–145.
16. Mattocks KM, Frayne S, Phibbs CS, et al.. Five-year trends in women veterans’ use of VA maternity benefits, 2008-2012. Womens Health Issues. 2014;24:e37–e42.
17. Zullig LL, Jackson GL, Dorn RA, et al.. Cancer incidence among patients of the US Veterans Affairs Health Care System. Mil Med. 2012;177:693–701.
18. StataCorp. Stata Statistical Software: Release 12. 2011.College Station, TX: StataCorp LP.
19. American Cancer Society. Cancer Facts & Figures. 2012American Cancer SocietyAvailable at: Accessed October 8, 2014.
20. Ouimette P, Wolfe J, Daley J, et al.. Use of VA health care services by women veterans: findings from a national sample. Women Health. 2003;38:77–91.
21. Patten Eileen, Parker Kim. Women in the US Military: growing share, distinctive profilePew Social and Demographic trends; 2011 Available at: Accessed October 8, 2014.
22. Office of Policy and Planning. Women veterans: past, present and future. 2007. Available at: Accessed October 8, 2014.
23. National Cancer Institute DCCPS Surveillance Research Program Cancer Statistics Branch. Surveillance, Epidemiology, and End Results (SEER) Program. SEER*Stat Databases: Incidence—SEER 13 Regs Public-Use. 2006. Available at: Accessed October 8, 2014.
24. US Department of Veterans Affairs. VA leads nation in breast cancer screening rates. US Census Bureau, American Community Survey PUMS. 2011. Available at: Accessed October 1, 2014.
25. Bleyer A, Welch HG. Effect of three decades of screening mammography on breast-cancer incidence. N Engl J Med. 2012;367:1998–2005.
26. Bickell NA, McAlearney AS, Wellner J, et al.. Understanding the challenges of adjuvant treatment measurement and reporting in breast cancer: cancer treatment measuring and reporting. Med Care. 2013;51:e35–e40.
27. Malin JL, Kahn KL, Adams J, et al.. Validity of cancer registry data for measuring the quality of breast cancer care. J Natl Cancer Inst. 2002;94:835–844.
28. Bickell NA, Chassin MR. Determining the quality of breast cancer care: do tumor registries measure up? Ann Intern Med. 2000;132:705–710.
29. Noone AM, Lund JL, Mariotto A, et al.. Comparison of SEER treatment data with Medicare claims. Med Care. 2014[Epub ahead of print].
30. United States Census. Bureau. Growth in Urban Population Outpaces Rest of Nation, Census Bureau Reports; 2010. Available at: Accessed October 1, 2014.
31. US Department of Health and Human Services, Health Resources and Services Administration, Maternal and Child Health Bureau. Women’s Health USA 2012. 2013.Rockville, MD: US Department of Health and Human Services.
32. Gomez SL, Glaser SL. Misclassification of race/ethnicity in a population-based cancer registry (United States). Cancer Causes Control. 2006;17:771–781.
33. West CN, Geiger AM, Greene SM, et al.. Race and ethnicity: comparing medical records to self-reports. J Natl Cancer Inst Monogr. 2005;35:72–74.
34. Luther SL, Neumayer L, Henderson WG, et al.. The use of breast-conserving surgery for women treated for breast cancer in the Department of Veterans Affairs. Am J Surg. 2013;206:72–79.
35. Gundlapalli AV, Redd A, Carter M, et al.. Validating a strategy for psychosocial phenotyping using a large corpus of clinical text. J Am Med Inform Assoc. 2013;20e2e355–e364.
36. DeLisle S, Kim B, Deepak J, et al.. Using the electronic medical record to identify community-acquired pneumonia: toward a replicable automated strategy. PLoS One. 2013;8:e70944.

breast cancer; women veterans; VA

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.