Disparities in timely treatment among young women with breast cancer

Abstract

Background:

Although evidence suggests worse breast cancer-specific survival associated with treatment delay beyond 90 days, little is known regarding the sociodemographic predictors of delays in cancer-directed surgery among young women with breast cancer. This is particularly notable, given that 5–10% of new diagnoses occur in younger women aged <40 years, commonly with more aggressive features than in older women.

Methods:

We used the National Cancer Database (2004–2017) to assess sociodemographic disparities in delay of upfront surgery beyond 90 days among young women with non-metastatic breast cancer, using multivariable logistic regression and predictive marginal modeling.

Results:

Black women experienced treatment delays more frequently than white women (aOR: 1.93 [95% CI: 1.76–2.11], p < 0.001). Adjusted rates of treatment delay were 4.91% [95% CI: 4.51%–5.30%] and 2.60% [95% CI: 2.47%–2.74%] for Black and white women, respectively, and 2.97% [95% CI: 2.83%–3.12%], 2.36% [95% CI: 2.03%–2.68%], and 1.18% [95% CI: 0.54%–1.81%] for women from metro, urban, and rural areas, respectively.

Conclusion:

These results suggest that improving access to timely treatment may be leveraged as a means through which to lessen the breast cancer disparities experienced by Black women.

1. Introduction

Breast cancer is the most common non-cutaneous cancer among women in the United States, comprising 14.8% of all new cancer cases and 7.2% of all cancer deaths annually.¹ Although breast cancer is considered a disease primarily impacting older women, 5–10% of new diagnoses occur in younger women aged <40 years.¹ Among young women, breast cancer often presents with more aggressive characteristics and worse prognosis, rendering timely initiation of treatment important.¹ Despite data that suggest decrements in overall survival and breast cancer-specific survival associated with treatment delay beyond 90 days,² young women may experience delays in the receipt of surgery for various reasons. These reasons may include structural barriers and inequities in access to cancer care,³ patient preferences and concerns about adverse effects of treatment,⁴ need to coordinate oncologic and reconstructive surgery,⁵ fertility preservation efforts,⁶ awaiting genetic testing results that may impact surgical decision-making,⁷ and competing demands more commonly faced by young people with cancer such as care for young children and other family members and fewer financial assets.^3,8,9 While treatment delays are well-documented among older women diagnosed with breast cancer from low-socioeconomic status (SES) strata or minority racial/ethnic backgrounds,^10–12 little is known about the broader sociodemographic predictors associated with treatment delay among young women with breast cancer at a national level. Therefore, we used the National Cancer Database (NCDB) to assess sociodemographic disparities in delay of upfront surgery among young women with non-metastatic breast cancer.

2. Methods

Using the patient-deidentified NCDB, a hospital-based cancer registry including ~70% of new cancer diagnosis in the US, we selected women aged <40 years diagnosed from 2004 to 2017 with ductal carcinoma in situ (DCIS, referred to as Stage 0) and non-metastatic invasive breast cancer (Stage I-III). We excluded patients with primary stage IV disease or missing stage data due to anticipated differences in treatment trajectory compared to those with potentially curative disease. Only patients who received surgery as their first course of treatment were included to focus our study on delays to initial surgical resection, and we acknowledge that different factors may be at play among women who receive neoadjuvant treatment.¹³ For example, human epidermal growth factor receptor 2 (HER2)–positive breast cancer and triple-negative breast cancer (TNBC) patients often receive chemotherapy prior to surgery and/or radiation therapy, and therefore may be excluded disproportionately from the sample.^14,15 The primary dependent variable of interest was delay of treatment, defined as receipt of the first course of surgery 90 days or more following the date of initial diagnosis, as done in prior studies.² The 90-day threshold was chosen because of evidence suggesting that expeditious treatment within 90 days may improve overall survival and breast cancer-specific survival.²

The primary independent variables of interest were year of diagnosis, race (white vs. Black, the two largest racial groups in the NCDB), age, insurance status, zip code-wide educational attainment (zip code-wide percentage without high school diploma), zip code-wide median income, Charlson-Deyo combined comorbidity index (CDCC), clinical stage, distance from treatment facility, rural/urban county, laterality, tumor size and grade, receptor status (ER, PR, HER2), and treatment modality in addition to surgery (to account for adjuvant treatments with radiation therapy, chemotherapy, and/or hormone therapy). Multivariable logistic regression defined adjusted odds ratios (aOR) with 95% confidence intervals (95% CI) of delaying treatment across all patients and separately for stages 0/1 and 2/3; additional stage-stratified models defined aORs for women with private insurance/managed care. Predictive marginal modeling, which uses the multivariable regression models to predict probability of delay for women stratified by race and age and holding all other covariates at the mean, provided estimates of the percentages of women experiencing treatment delay by patient group. Supplementary models provided estimates of the percentages of women experiencing treatment delay by patient group for stages 0, 1, 2, and 3 separately. Analyses were performed with Stata/SE 16.1 (Stata-Corp, College Station, TX). This study was exempt from IRB review given the use of de-identified data.

3. Results

Among the 74,304 women who met inclusion criteria, the median age was 37 (Interquartile Range [IQR] 34–39) years. In the primary model across all patients, Black women experienced treatment delays more frequently than white women (aOR: 1.93 [95% CI: 1.76–2.11], p < 0.001; unadjusted rates: 8.11% vs. 3.84%). Additionally, lower educational attainment (zip code-wide percentage without high school diploma, aOR for <6.3% vs. ≥17.6%: 0.57 [95% CI: 0.50–0.66], p < 0.001; unadjusted rates: 3.44% vs. 6.31%), increasing levels of income (aOR for ≥$63,333 vs. <$40,227: 1.29 [95% CI: 1.13–1.48], p < 0.001; unadjusted rates: 4.00% vs. 5.53%), and decreasing rurality (aOR for rural vs. metro: 0.39 [95% CI: 0.22–0.67], p = 0.001; unadjusted rates: 2.78% vs. 4.48%) were significantly associated with higher rates of treatment delay in the primary model (Table 1). These racial, urban/ rural, educational attainment, and income disparities largely persisted in sensitivity models restricted to women with private insurance/ managed care and when stratified by Stage 0/1 vs. Stage 2/3.

Table 1.

Association between sociodemographic characteristics and treatment delay for young women with breast cancer, 2004–2017

	OR (95% CI)
Characteristic	All Patients	Stage 0/1	Stage 2/3	Stage 0/1, Insured Patients Only	Stage 2/3, Insured Patients Only
No.	74,304	37,745	36,559	31,667	28,124
Race
White	1 [Reference]	1 [Reference]	1 [Reference]	1 [Reference]	1 [Reference]
Black	1.93 (1.76–2.11) ***	1.94 (1.71–2.20) ***	1.93 (1.69–2.22) ***	2.05 (1.76–2.37) ***	2.02 (1.69–2.41) ***
Zip Code-Wide Percent Without High School Education
17.6% or more	1 [Reference]	1 [Reference]	1 [Reference]	1 [Reference]	1 [Reference]
10.9% to 17.5%	0.83 (0.75–0.93) ***	0.80 (0.69–0.93) **	0.89 (0.76–1.05)	0.85 (0.70–1.04)	1.00 (0.80–1.23)
6.3% to 10.8%	0.68 (0.60–0.77) ***	0.70 (0.59–0.82) ***	0.68 (0.56–0.82) ***	0.83 (0.68–1.02)	0.79 (0.62–1.01)
Less than 6.3%	0.57 (0.50–0.66) ***	0.56 (0.46–0.68) ***	0.61 (0.49–0.75) ***	0.62 (0.49–0.78) ***	0.72 (0.55–0.94) *
Zip Code-Wide Median Household Income
Less than $40,227	1 [Reference]	1 [Reference]	1 [Reference]	1 [Reference]	1 [Reference]
$40,227 – $50,353	1.07 (0.95–1.21)	1.20 (1.01–1.42) *	0.94 (0.79–1.12)	1.14 (0.91–1.42)	0.88 (0.69–1.13)
$50,354 – $63,332	1.04 (0.92–1.18)	1.23 (1.03–1.47) *	0.82 (0.68–0.99) *	1.10 (0.87–1.38)	0.75 (0.58–0.96) *
$63,333 +	1.29 (1.13–1.48) ***	1.43 (1.18–1.73) ***	1.13 (0.92–1.38)	1.35 (1.06–1.71) *	1.07 (0.82–1.39)
Rural/Urban Household Designation
Metro	1 [Reference]	1 [Reference]	1 [Reference]	1 [Reference]	1 [Reference]
Urban	0.79 (0.68–0.91) **	0.76 (0.63–0.93) **	0.85 (0.68–1.07)	0.84 (0.66–1.06)	1.05 (0.79–1.39)
Rural	0.39 (0.22–0.67) ***	0.47 (0.23–0.92) *	0.32 (0.13–0.80) *	0.66 (0.30–1.42)	0.41 (0.13–1.36)
Insurance Status
Not Insured	1 [Reference]	1 [Reference]	1 [Reference]	N/A	N/A
Private Insurance/Managed Care	0.53 (0.45–0.62) ***	0.45 (0.36–0.56) ***	0.60 (0.48–0.76) ***	N/A	N/A
Medicaid	1.00 (0.84–1.18)	0.93 (0.73–1.20)	1.03 (0.81–1.32)	N/A	N/A
Medicare	0.99 (0.77–1.27)	0.84 (0.59–1.19)	1.16 (0.82–1.65)	N/A	N/A
Other Government	0.67 (0.48–0.93) *	0.63 (0.41–0.97) *	0.66 (0.39–1.10)	N/A	N/A
Unknown	0.68 (0.50–0.92) *	0.57 (0.37–0.87) **	0.77 (0.50–1.20)	N/A	N/A

^***:p ≤ 0.001

^**:p ≤ 0.01

^*:p < 0.05.

Marginal modeling-derived estimates of treatment delay by patient group are presented in Fig. 1. Across the entire cohort, adjusted rates of treatment delay were 4.91% [95% CI: 4.51%–5.30%] and 2.60% [95% CI: 2.47%–2.74%] for Black and white women, respectively, and 2.97% [95% CI: 2.83%–3.12%], 2.36% [95% CI: 2.03%–2.68%], and 1.18% [95% CI: 0.54%–1.81%] for women from metro, urban, and rural areas, respectively. With regards to educational attainment, adjusted rates of treatment delay were 3.93% [95% CI: 3.56%–4.30%] in areas of lowest education and 2.29% [95% CI: 2.09%–2.49%] in areas of greatest education. Finally, income was also associated with disparities in rates of treatment delay: 2.52% [95% CI: 2.25%–2.80%] for <$40,227 and 3.24% [95% CI: 3.02%–3.47%] for ≥$63,333 (Fig. 1A). Joint chi-squared tests of constituent contrasts across all marginal modeling results resulted in p < 0.001, indicating significant differences between rates of treatment delay by race, rural/urban county, educational attainment, and income. Subgroup analyses by stage (Stage 0/1, Fig. 1B; Stage 2/3, Fig. 1C) and upon restriction to insured patients only (Stage 0/1, insured, Fig. 1D; Stage 2/3, insured, Fig. 1E) illustrate the persistence of racial disparities in treatment delay among young women with breast cancer. Marginal modeling results for individual stages are presented in the Supplementary Figure.

Fig. 1.

Predictive marginal modeling estimates of treatment delay for young women with breast cancer across sociodemographic characteristics, 2004–2017.

4. Discussion

Based on data from a national database of young women with curable breast cancer, Black women were twice as likely to experience a treatment delay compared to their white peers. Furthermore, treatment delay was significantly greater among women living in metro/urban areas, as well as in women from areas with lower education and higher median income. Racial, educational, and income disparities persisted among patients with private insurance/managed care.

Racial disparities experienced by Black women with breast cancer extend throughout the cancer continuum. The incidence of breast cancer is higher among young Black women than young white women (a trend that appears to reverse after menopause).^16,17 Black women are at greater risk of triple-negative disease,¹⁸ more likely to present with more advanced disease,¹⁹ experience more limited access to high-quality care,²⁰ and experience worse health-related quality of life in the survivorship setting.³ Furthermore, mortality of breast cancer is higher among young Black women compared to young white women.²¹ The present study adds to the literature by highlighting another facet of healthcare access – that is, access to timely treatment – wherein Black women experience disparate care. Racial disparities in treatment delays among young women are concerning given myriad studies that suggest detrimental effects of treatment delays on breast cancer outcomes, which include worse overall survival and breast cancer-specific survival.^22,22 Fortunately, the corollary to these findings is that improving access to timely treatment may be leveraged as a means through which to mitigate the disparities in survival outcomes experienced by Black women.

There are likely complex reasons underlying these disparities, including systemic racism towards marginalized groups; younger women of color are particularly vulnerable given their position at the intersection of race/ethnicity, gender, and age. Additionally, implicit bias in physician recommendations, in combination with cultural beliefs and the effects of centuries of injustice carried out by the medical system towards Black people, may lead to treatment delays for Black individuals with cancer.²³ Our findings build upon the growing body of literature assessing disparities in the often overlooked adolescent and young adult population, where racial disparities have been demonstrated in both BRCA genetic testing and breast cancer mortality.^21,24 It is important to note that young adult cancer survivors may face significant competing family and work demands, as well as less time spent accumulating financial assets, resulting in fewer sick days or allowable time off work and challenges with child care and transportation, impacting patients’ ability to seek timely treatment.^8,25 These intersectional disparities of race, financial status, age, and sex are important to explore and mitigate.

While our results related to race and educational attainment are well-explained by access to care barriers and the social determinants of health, we found that women in rural regions experienced treatment delays less often compared to women in urban and metro areas. Previous research has described such a “rural reversal,” in which the likelihood of late-stage diagnosis was highest among residents of densely populated Chicago, decreasing along a continuum from suburbs and smaller metropolitan areas to rural settings.²⁶ These disparities may represent factors uncaptured in our analysis such as waiting time to schedule surgery, likelihood of following physician recommendation, or differences in the distribution of patients by race and age, which may vary across the urban-rural continuum.²⁶ Similarly, a slight increase in treatment delay for women in wealthier neighborhoods may be indicative other uncaptured differences including greater access to genetic testing, plastic surgery, and oncofertility, as well as a higher likelihood to pursue alternative medicine.²⁷ Similarly, women from areas with higher median income experienced higher rates of treatment delay, which appears paradoxical to our findings regarding educational attainment. However, previous literature indicates that both high- and low-income women with breast cancer experience similarly high rates of delay for diagnostic testing, diagnosis, and chemotherapy when enrolled in high-deductible health plans due to high out-of-pocket spending obligations.^28,29 Additionally, working women diagnosed with breast cancer have reported concerns related to obtaining paid sick leave and disability benefits, as well as inflexible work schedules, in their ability to receive timely treatment.³⁰ Therefore, the different rates of treatment delay associated with income compared with education may be reflective of complex factors uncaptured in our analysis and require additional investigation.

Limitations of this study include the retrospective design inherent to large database studies. Although the large sample size allows comparison of disparities while adjusting for many possible confounders, it does not capture all potential variables like factors that contribute to patient-related delays to care. In this way, our study is limited by the inability to ascertain the reason for the observed disparities. The reasons associated with treatment delays as well as other quality metrics such as treatment refusal and treatment non-completion in breast and other cancers^31–34 merit further exploration with quantitative and qualitative analyses. It would also be critical to assess the disparities experienced by women who receive neoadjuvant chemotherapy as their first line of treatment, particularly for HER2-positive breast cancer or triple-negative breast cancer, which disproportionately affect young Black women; factors are likely different given the challenges of setting up chemotherapy and constitute important areas of future research.^35,36 Lastly, our findings are limited by racial categorizations, which cannot fully capture the diversity of ways in which people identify, and unique human experiences as they relate to a cancer diagnosis.

Overall, this study highlights stark disparities in timely treatment for young women with potentially curable cancers. We hope our findings serve as a call to action to assess reasons behind treatment delays that fall along racial lines. Importantly, efforts are needed to design, implement, and assess ways to mitigate these disparities in access to timely surgery for young women with breast cancer.