Abstract
Background:
Recent reports suggest that racial differences in breast cancer incidence rates have decreased. We examined whether these findings apply to breast cancer mortality while considering age, period, and cohort influences on both absolute and relative measures of breast cancer mortality.
Methods:
Using publicly available datasets (CDC Wonder, Human Mortality Database), we developed an age-period-cohort model of breast cancer mortality and breast cancer deaths as a proportion of all deaths during 1968–2019 among all women and by five race/ethnicity groups with sufficient numbers for estimation: Hispanic (all races), American Indian/Alaska Native and Asian/Pacific Islanders (regardless of ethnicity), non-Hispanic Black, and non-Hispanic White.
Results:
Initially increasing after 1968, age-adjusted breast cancer mortality rates have decreased among all racial/ethnic groups since 1988. The age-adjusted percent of all deaths due to breast cancer also has been declining for non-Hispanic White women since about 1990 while increasing or holding steady for other race/ethnic groups. In 2019, the age-adjusted percent of deaths due to breast cancer for women was highest for Asian/Pacific Islanders (5.6%) followed by non-Hispanic Black (4.5%), Hispanic (4.4%), non-Hispanic White (4.1%), and American Indian/Alaska Native women (2.6%).
Conclusions:
Breast cancer mortality disparities are now greater on both relative and absolute scales for non-Hispanic Black women, and using the relative scale for Asian/Pacific Islander and Hispanic women, compared with non-Hispanic White women for the first time in 50 years.
Keywords: breast cancer, mortality, disparities, age-period-cohort modeling
Introduction
Breast cancer is the second leading cause of cancer death in U.S. women.1 Breast cancer mortality has decreased by about 40% since 1975 in all racial and ethnic groups combined.2–4 However, when trends for White women are compared with trends for women of other races and Hispanic ethnicity, long-standing racial disparities in breast cancer mortality are evident through descriptive analyses of surveillance data with greater mortality declines observed for White women.4–6 Based on data through 2015, DeSantis2 suggested that racial-ethnic disparities in mortality rates have stabilized with death rates 39% higher among non-Hispanic Black women compared with non-Hispanic White women. In another analysis of breast cancer incidence rates, Davis Lynn7 concluded that the Black-White breast cancer mortality disparity is unlikely to worsen since incidence was projected to increase for non-Hispanic White women but decrease for non-Hispanic Black women; their analysis was motivated by the finding that while recent patterns suggested that breast cancer incidence rates for Black women had converged with incidence rates for White women, their modeling projections showed that these incidence rates would diverge after 2015.
Many complex health system factors influence breast cancer mortality, and these factors impact women of different races and ethnicities in dissimilar ways. Recent use of mammography screening for the early detection of breast cancer varies by race and ethnicity, ranging from 65% of non-Hispanic Asian women to 79% of Hispanic women.8 Variations in access to breast cancer care by race and ethnicity have been documented in numerous studies farther along the cancer care continuum including follow-up after an abnormal exam,9–11 timeliness of initiation of treatment,12,13 and concordance of therapy with recommended guidelines.14 Availability, use, and timing of breast health services differently by racial subgroups reflect structural racism rather than putative underlying biologic differences,15 and in this study we observe the definition of race as a social-political construct rather than a reflection of biology or genetics.16,17
Studies of breast cancer mortality disparities tend to compare breast cancer mortality rates across racial and ethnic groups on an absolute scale.18 As an alternative to measuring the mortality burden of breast cancer on the absolute scale, breast cancer mortality can also be measured as a proportion of all deaths. Measured as a fraction of all deaths, the breast cancer proportion reflects the impact of mammography screening and breast cancer treatment as well as progress in the prevention and treatment of competing causes of mortality within each group of interest. When the burden of breast cancer is measured on this relative scale—as a proportion of all deaths—trends reveal whether breast cancer is increasing or decreasing relative to all other causes of mortality which, also, vary by race and ethnicity. In some settings, health outcomes considered within the context of the total mortality burden within each race and ethnic group may be a more salient center of focus than the mortality rates of the total population, the majority population, or the “best off” group.19
In 2009, Harper4 used Surveillance, Epidemiology, and End Results (SEER) program data through 2004 to show that absolute racial and ethnic group disparities in breast cancer declined across multiple outcomes—including 5-year breast cancer-specific probability of death and mortality—but relative disparities increased by 17% for 5-year probability of death and by 24% for mortality. Analysis by Harper and colleagues used age-adjusted population averages for comparisons among women 50 and older. We updated and extended this analysis using age-period-cohort modeling and publicly available mortality data through 2019 to examine progress against racial and ethnic disparities using both an absolute measure (breast cancer mortality rate) and a relative measure, defined as the percent of all deaths due to breast cancer. By considering a relative measure of mortality burden, the breast cancer burden for each race and ethnic group is compared with the overall mortality risks for their own group, providing an alternative to placing non-Hispanic White women as the reference standard.
Methods
Previously we estimated the overall proportion of deaths attributable to breast cancer by age and year of death for U.S. women based on data available through 2014;20,21 this proportion was highest (4.1% to 12.9%) for women in their 40s and 50s born during 1900–2000. Here we report an updated model of the percent of all deaths attributable to breast cancer for non-Hispanic Black/African American (hereafter called Black), American Indian/Alaska Native, Asian/Pacific Islander, Hispanic (all races), and non-Hispanic White women. Only publicly available anonymous data were used; this study was exempt from review as determined by the University of Wisconsin Health Sciences Institutional Review Board.
As described in greater detail in eAppendix 1, counts of deaths due to malignant neoplasms (ICD-10 C00-C97), malignant breast cancer (ICD-10 C50), disease of the circulatory system (ICD-10 I00-I99), diseases of the respiratory system (ICD-10 J00-J98), and all-causes for single years of age 0–99 and populations for single years of age 0–84 were obtained for all women, Hispanic women (all races), White women, non-Hispanic White women, Black women, non-Hispanic Black women, American Indian/Alaska Native women, and Asian/Pacific Islander women from the Detailed Mortality file on CDC WONDER for 1999–2019.22 CDC WONDER is an online national public health data resource providing an ad-hoc query system and downloadable datasets regarding births, population counts, morbidity and mortality, and other health-related topics.23 Due to small numbers, American Indian/Alaska Native and Asian/Pacific Islander women included both Hispanic and non-Hispanic women. Deaths in women from all malignant neoplasms (ICD-8 140–209, ICD-9 140–208), malignant breast cancer (ICD-8 and -9 174), circulatory system diseases (ICD-8 390–458, ICD-9 390–459), respiratory diseases (ICD-8 and -9 460–519), and all-causes and populations for age groups (<1, 1–4, 5–9, 10–14, 15–19, 20–24, 25–34, 35–44, 45–54, 55–64, 65–74, 75–84) were obtained for all races combined, White women, and Black women from the Compressed Mortality files on CDC WONDER for calendar years 1968–1978 (ICD-8) and 1979–1998 (ICD-9).22 Female all-cause mortality life tables for single years of age 0–119 were obtained for all races combined from the Berkeley Mortality Database for the 1900–2000 birth cohorts.24
We used a series of age-period-cohort (APC) models25 to estimate the proportion of deaths due to breast cancer among all women using binomial logistic regression models,26 for all races combined and for individual race/ethnic groups. Age, period (year of death), and cohort (year of birth) were entered into each model as additive natural cubic splines.27 Race-specific models borrowed strength from the model for all races combined by using estimates from the all races-combined model as an offset term; the model for non-Hispanic Blacks (Whites) borrowed strength from the model for Blacks (Whites) by using estimates from the Black (White) model as an offset term. Similar to the identification strategy of Carstensen,25 age and cohort effects for all races combined were penalized towards linear terms, while the period effect for all races combined and all effects for race- and ethnicity-specific models were penalized towards no effect. Smoothing parameters for the splines were selected using generalized cross validation with a BIC-like penalty to prevent over-fitting.28
Standard errors for the estimated proportion of deaths due to breast cancer on the logit scale were obtained using the delta method. 95% confidence intervals were obtained on the logit scale using a Wald-type procedure and back-transformed to the proportion scale.
Similar APC models were used to estimate race- and ethnicity-specific hazard ratios for all-cause mortality. Values were estimated for ages 0–119 and birth cohorts 1900–2000. Breast cancer-specific mortality was calculated by multiplying the proportion of deaths due to breast cancer by all-cause mortality. Standard errors for the estimated mortality rates on the logit scale were obtained using the delta method. 95% confidence intervals were obtained on the log scale using a Wald-type procedure and back-transformed to the rate scale.
Age-adjusted rates and proportions used the female population for all races combined in 2019 for ages 18–84. Analyses were conducted using the mgcv26,27 and ggplot229 packages in R v4.0.2.30
Results
Age-adjusted death rates have been decreasing over the entire time frame of the analysis (1968–2019) from all causes combined (eFigure 1) as well as circulatory diseases (eFigure 2). Cancer mortality rates for all types combined other than breast cancer have decreased for the five race/ethnicity groups over the past three decades (eFigure 3). Age-adjusted respiratory disease mortality rates through 2019 have varied but remained fairly stable year-to-year (eFigure 4). Within age, race, and ethnicity groups, all-cause mortality rates in 2019 are lower than rates in 1979 except for 30 and 40 year-old non-Hispanic White women (increasing from 54 to 81 per 100,000 and 148 to 161 per 100,000, respectively) and 30 year-old American Indian/Alaska Native women (79 to 100 per 100,000) (eTable 1).
Age-adjusted breast cancer mortality rates decreased for all five race/ethnicity groups since about 1988 (Figure 1). Within each race and ethnicity group, breast cancer mortality rates increased by age across all years of birth (birth cohorts) (Figure 2) and years of death (Figure 3). Breast cancer mortality rates decreased over time for non-Hispanic White women of all ages and for women of other races and Hispanic ethnicity at younger ages (Table 1). However, for 80 year-old Asian/Pacific Islander, Hispanic, and non-Hispanic Black women, breast cancer mortality rates were higher in 2019 than in 1979. In 2019, breast cancer mortality rates for 50 year-old women were highest for non-Hispanic Black women at 39 per 100,000 (95% CI 38, 40) followed by non-Hispanic White women at 22 per 100,000 (95% CI 22, 23), equal to 18 per 100,000 (95% CI 17, 18) for both Asian/Pacific Islander women and Hispanic women, and 13 per 100,000 (95% CI 12, 14) for American Indian/Alaska Native women.
Figure 1.
Age-adjusted breast cancer mortality rates per 100,000 women according to year of death and race/ethnicity, 1968–2019, United States.
Figure 2.
Breast cancer mortality rates per 100,000 women according to year of birth, age and race/ethnicity, 1900 to 1985, United States for (a) 1900, (b) 1905, (c) 1910, (d) 1915, (e) 1920, (f) 1925, (g) 1930, (h) 1935, (i) 1940, (j) 1945, (k) 1950, (l) 1955, (m) 1960, (n) 1965, (o) 1970, (p) 1975, (q) 1980, and (r) 1985. Shaded regions show 95% confidence intervals.
Figure 3.
Breast cancer mortality rates per 100,000 women according to year of death, age and race/ethnicity, 1969 to 2019, United States for (a) 1969, (b) 1974, (c) 1979, (d) 19834 (e) 1989, (f) 1994, (g) 1999, (h) 2004, (i) 2009, (j) 2014, and (k) 2019. Shaded regions show 95% confidence intervals.
Table 1.
Breast cancer mortality rates per 100,000 women by age, race/ethnicity, and calendar year of death for 1979, 1999, and 2019
| Age, Race and Ethnicity | 1979 | 1999 | 2019 | |||
|---|---|---|---|---|---|---|
| Rate | (95% CI) | Rate | (95% CI) | Rate | (95% CI) | |
| Age 30 | ||||||
| American Indian/Alaska Nativea | 1 | (1–1) | 1 | (1–1) | 1 | (1–1) |
| Asian/Pacific Islandera | 1 | (1–2) | 1 | (1–1) | 1 | (1–1) |
| Hispanicb | 2 | (2–3) | 2 | (2–2) | 1 | (1–2) |
| Non-Hispanic Black | 5 | (5–6) | 4 | (4–5) | 3 | (3–3) |
| Non-Hispanic White | 3 | (3–3) | 2 | (2–2) | 1 | (1–1) |
| Age 40 | ||||||
| American Indian/Alaska Native | 9 | (8–11) | 6 | (5–6) | 5 | (5–5) |
| Asian/Pacific Islander | 9 | (8–10) | 7 | (7–7) | 6 | (6–6) |
| Hispanic | 12 | (12–13) | 9 | (9–9) | 7 | (7–7) |
| Non-Hispanic Black | 26 | (25–27) | 24 | (23–24) | 17 | (16–18) |
| Non-Hispanic White | 19 | (18–19) | 12 | (11–12) | 9 | (8–9) |
| Age 50 | ||||||
| American Indian/Alaska Native | 27 | (23–32) | 18 | (16–19) | 13 | (12–14) |
| Asian/Pacific Islander | 26 | (24–28) | 22 | (22–23) | 18 | (17–18) |
| Hispanic | 31 | (29–32) | 26 | (25–27) | 18 | (17–18) |
| Non-Hispanic Black | 58 | (56–60) | 59 | (58–60) | 39 | (38–40) |
| Non-Hispanic White | 51 | (50–52) | 35 | (34–35) | 22 | (22–23) |
| Age 60 | ||||||
| American Indian/Alaska Native | 49 | (42–59) | 36 | (34–39) | 22 | (21–24) |
| Asian/Pacific Islander | 35 | (33–38) | 35 | (33–36) | 28 | (27–29) |
| Hispanic | 45 | (42–47) | 40 | (39–41) | 30 | (29–31) |
| Non-Hispanic Black | 80 | (77–83) | 85 | (83–87) | 63 | (61–65) |
| Non-Hispanic White | 86 | (85–88) | 64 | (63–65) | 40 | (40–41) |
| Age 70 | ||||||
| American Indian/Alaska Native | 65 | (54–78) | 58 | (54–62) | 38 | (36–40) |
| Asian/Pacific Islander | 32 | (30–35) | 40 | (39–42) | 37 | (35–38) |
| Hispanic | 48 | (46–51) | 54 | (53–56) | 47 | (46–48) |
| Non-Hispanic Black | 90 | (86–93) | 109 | (107–111) | 90 | (88–93) |
| Non-Hispanic White | 104 | (102–106) | 93 | (92–94) | 65 | (64–66) |
| Age 80 | ||||||
| American Indian/Alaska Native | 75 | (62–91) | 85 | (78–92) | 63 | (58–67) |
| Asian/Pacific Islander | 33 | (30–36) | 50 | (47–52) | 50 | (48–52) |
| Hispanic | 56 | (53–59) | 76 | (74–78) | 69 | (67–71) |
| Non-Hispanic Black | 103 | (99–107) | 147 | (143–150) | 126 | (123–130) |
| Non-Hispanic White | 125 | (120–129) | 135 | (134–137) | 104 | (103–106) |
Due to smaller numbers, American Indian/Alaska Native and Asian/Pacific Islander include both Hispanic and non-Hispanic
Includes all races
Abbreviation: CI, confidence interval
After age-adjustment, the percent of deaths due to breast cancer in 2019 was highest for Asian/Pacific Islander women (5.6%) followed by non-Hispanic Black (4.5%), Hispanic (4.4%), non-Hispanic White (4.1%), and American Indian/Alaska Native women (2.6%) (Figure 4). The fact that non-Hispanic White women have a lower percent of deaths due to breast cancer than Hispanic, non-Hispanic Black, and Asian/Pacific Islander women is a recent phenomenon, with the fraction of deaths due to breast cancer for non-Hispanic White women first surpassed by Asian/Pacific Islander women in 2005, followed by non-Hispanic Black women in 2012 and then ultimately Hispanic women in 2015.
Figure 4.
Age-adjusted percent of all deaths due to breast cancer according to year and race/ethnicity, 1968–2019.
Plots for the proportion of all deaths due to breast cancer within strata of age and race/ethnicity reflect the patterns in the age-adjusted figure with proportions generally decreasing over the past three decades for non-Hispanic White and American Indian/Alaska Native women and increasing or staying level for other three groups (Figure 5). For example, between 1979 and 2019 for 50-year-old women (Table 2), the percent of all deaths due to breast cancer increased for Asian/Pacific Islanders (10.2 to 14.1%), Hispanic women (9.2 to 9.5%) and non-Hispanic Black women (7.1 to 8.7%), but decreased for American Indian/Alaska Native women (5.0 to 3.7%) and non-Hispanic White women (12.8 to 6.9%).
Figure 5.
Proportion of deaths due to breast cancer in women by race/ethnicity according to year of death, 1968 to 2019, United States for (a) 1969, (b) 1974, (c) 1979, (d) 1984, (e) 1989, (f) 1994, (g) 1999, (h) 2004, (i) 2009, (j) 2014, and (k) 2019. Shaded regions show corresponding 95% confidence intervals.
Table 2.
The proportion (%) of all deaths due to breast cancer in women according to age, race/ethnicity, and calendar year of death for 1979, 1999, and 2019
| Age, Race and Ethnicity | 1979 | 1999 | 2019 | |||
|---|---|---|---|---|---|---|
| % | (95% CI) | % | (95% CI) | % | (95% CI) | |
| Age 30 | ||||||
| American Indian/Alaska Nativea | 1.5 | (1.2–1.8) | 1.1 | (1.0–1.2) | 0.7 | (0.7–0.8) |
| Asian/Pacific Islandera | 3.9 | (3.5–4.4) | 3.8 | (3.5–4.2) | 3.7 | (3.4–4.1) |
| Hispanicb | 4.6 | (4.3–4.9) | 3.8 | (3.7–4.0) | 3.2 | (3.0–3.4) |
| Non-Hispanic Black | 3.4 | (3.3–3.6) | 3.6 | (3.5–3.7) | 3.1 | (2.8–3.3) |
| Non-Hispanic White | 5.5 | (5.4–5.6) | 3.1 | (3.1–3.2) | 1.7 | (1.6–1.8) |
| Age 40 | ||||||
| American Indian/Alaska Native | 4.2 | (3.5–4.9) | 3.3 | (3.1–3.6) | 2.6 | (2.4–2.8) |
| Asian/Pacific Islander | 10.2 | (9.4–11.0) | 11.2 | (10.7–11.7) | 11.9 | (11.4–12.4) |
| Hispanic | 9.8 | (9.4–10.3) | 9.3 | (9.1–9.6) | 8.6 | (8.3–8.8) |
| Non-Hispanic Black | 7.3 | (7.1–7.6) | 7.9 | (7.7–8.1) | 7.8 | (7.5–8.0) |
| Non-Hispanic White | 12.6 | (12.4–12.8) | 8.6 | (8.5–8.7) | 5.4 | (5.2–5.5) |
| Age 50 | ||||||
| American Indian/Alaska Native | 5.0 | (4.2–5.9) | 5.1 | (4.8–5.5) | 3.7 | (3.5–4.0) |
| Asian/Pacific Islander | 10.2 | (9.5–11.0) | 14.1 | (13.6–14.6) | 14.1 | (13.6–14.5) |
| Hispanic | 9.2 | (8.8–9.6) | 11.0 | (10.8–11.3) | 9.5 | (9.3–9.7) |
| Non-Hispanic Black | 7.1 | (6.9–7.3) | 9.4 | (9.2–9.6) | 8.7 | (8.4–8.9) |
| Non-Hispanic White | 12.8 | (12.6–13.0) | 11.6 | (11.5–11.7) | 6.9 | (6.8–7.1) |
| Age 60 | ||||||
| American Indian/Alaska Native | 4.0 | (3.4–4.8) | 4.1 | (3.9–4.4) | 3.4 | (3.2–3.6) |
| Asian/Pacific Islander | 6.0 | (5.6–6.5) | 8.4 | (8.1–8.7) | 9.4 | (9.1–9.7) |
| Hispanic | 5.7 | (5.4–6.0) | 6.9 | (6.7–7.0) | 6.6 | (6.5–6.7) |
| Non-Hispanic Black | 4.9 | (4.8–5.1) | 6.4 | (6.3–6.5) | 6.1 | (6.0–6.3) |
| Non-Hispanic White | 8.7 | (8.6–8.8) | 8.3 | (8.2–8.4) | 5.7 | (5.6–5.8) |
| Age 70 | ||||||
| American Indian/Alaska Native | 2.6 | (2.2–3.1) | 2.8 | (2.6–3.0) | 3.0 | (2.8–3.1) |
| Asian/Pacific Islander | 2.4 | (2.3–2.6) | 3.6 | (3.5–3.8) | 5.2 | (5.0–5.4) |
| Hispanic | 2.8 | (2.7–2.9) | 3.6 | (3.5–3.6) | 4.4 | (4.3–4.5) |
| Non-Hispanic Black | 2.9 | (2.8–3.0) | 3.8 | (3.7–3.9) | 4.5 | (4.4–4.6) |
| Non-Hispanic White | 4.9 | (4.8–4.9) | 4.9 | (4.8–4.9) | 4.5 | (4.4–4.5) |
| Age 80 | ||||||
| American Indian/Alaska Native | 1.5 | (1.3–1.8) | 1.9 | (1.7–2.0) | 1.9 | (1.8–2.1) |
| Asian/Pacific Islander | 0.9 | (0.8–1.0) | 1.5 | (1.5–1.6) | 2.2 | (2.1–2.3) |
| Hispanic | 1.3 | (1.2–1.4) | 1.9 | (1.8–1.9) | 2.3 | (2.3–2.4) |
| Non-Hispanic Black | 1.8 | (1.7–1.9) | 2.4 | (2.4–2.5) | 2.8 | (2.7–2.9) |
| Non-Hispanic White | 2.4 | (2.3–2.5) | 2.6 | (2.6–2.7) | 2.6 | (2.5–2.6) |
Due to smaller numbers, Asian/Pacific Islander includes both Hispanic and non-Hispanic
Includes all races
Abbreviation: CI, confidence interval
Across all years and ages, the proportion of all deaths attributable to breast cancer for non-Hispanic White women peaked at 14.3% for women aged 44 in 1989 (Table 3). The peak occurred at the same age (44) for non-Hispanic Black women in 1991 (10.1%). Breast cancer as a percentage of all deaths peaked at age 43 in 1990 for Hispanic women at 12.7% and at age 45 for Asian/Pacific Islander women in 1992 (14.9%). This peak percentage occurred at the oldest age (49) and the lowest level (5.8%) for American Indian/Alaska Native women in 1991. By 2019 non-Hispanic White women had a lower fraction of deaths due to breast cancer than non-Hispanic Black women at all ages, Hispanic women under age 68, and Asian/Pacific Islander women under age 75. In 2019, American Indian/Alaska Native women has a lower fraction of deaths due to breast cancer than all other groups for ages under 84—the age at which Asian Pacific Islander had the lowest fraction of all groups through age 99.
Table 3.
Ages and years that breast cancer peaked as a proportion of all causes of death by race/ethnicity among females, 1968–2019
| Year | Race/Ethnicitya | Age at Peak | Proportion at Peak | 95% CI |
|---|---|---|---|---|
| 1979 | American Indian/Alaska Native | 50 | 5.0 | (4.2–5.9) |
| Asian/Pacific Islander | 43 | 10.7 | (9.9–11.6) | |
| Hispanic | 42 | 10.0 | (9.6–10.5) | |
| Non-Hispanic Black | 42 | 7.5 | (7.2–7.8) | |
| Non-Hispanic White | 43 | 13.1 | (12.9–13.4) | |
| 1999 | American Indian/Alaska Native | 50 | 5.1 | (4.8–5.5) |
| Asian/Pacific Islander | 49 | 14.2 | (13.7–14.7) | |
| Hispanic | 49 | 11.1 | (10.8–11.3) | |
| Non-Hispanic Black | 49 | 9.4 | (9.2–9.6) | |
| Non-Hispanic White | 50 | 11.6 | (11.5–11.7) | |
| 2019 | American Indian/Alaska Native | 50 | 3.7 | (3.5–4.0) |
| Asian/Pacific Islander | 49 | 14.2 | (13.8–14.7) | |
| Hispanic | 49 | 9.6 | (9.4–9.9) | |
| Non-Hispanic Black | 49 | 8.7 | (8.5–9.0) | |
| Non-Hispanic White | 50 | 6.9 | (6.8–7.1) | |
| 1968–2019 | American Indian/Alaska Native | 49 (in 1991) | 5.8 | (5.2–6.5) |
| Asian/Pacific Islander | 45 (in 1992) | 14.9 | (14.2–15.6) | |
| Hispanic | 43 (in 1990) | 12.7 | (12.3–13.1) | |
| Non-Hispanic Black | 44 (in 1991) | 10.1 | (9.8–10.4) | |
| Non-Hispanic White | 44 (in 1989) | 14.3 | (14.1–14.5) |
Due to smaller numbers, Asian/Pacific Islander includes both Hispanic and non-Hispanic.
Hispanic includes all races.
Abbreviation: CI, confidence interval
Discussion
Our results agree and extend findings by Harper4 that relative racial and ethnic disparities for breast cancer mortality have increased over the past several decades. Mortality rates on an absolute scale including the rank-order by race and ethnicity have maintained a similar pattern for over 50 years, with breast cancer mortality rates for non-Hispanic Black and White women across all ages exceeding breast cancer mortality rates for Hispanic women and Asian/Pacific Islanders. Conversely, relative disparities have re-ordered dramatically with the proportion of non-Hispanic White women dying from breast cancer going from the largest proportion before 2005 to the smallest proportion after 2015. Harper used population-weighted summary measures while we leveraged an age-period-cohort model for all races combined to stabilize the estimates for the racial/ethnic subgroups and to capture the inter-dependent trends represented by birth cohort, age, and year of death. We included women of all ages (<120 y) in our analysis, whereas Harper4 limited their study to women aged ≥50 y due to their focus on mammography screening. Since the average age at breast cancer diagnosis for Black, Hispanic and Asian/Pacific Islander breast cancer cases tends to be younger than for non-Hispanic White women, a wide age range is important to consider when making comparisons across racial and ethnic groups. This point is underscored by the observation that older (≥70 y) non-Hispanic White women have benefited from decreasing breast cancer mortality rates but rates have actually increased for older Asian/Pacific Islander, Hispanic, and non-Hispanic Black women. These trends are occurring in the context of declining mortality rates for all causes combined as well as for circulatory diseases, and increasing life expectancy, for all racial and ethnic groups.
While calculations would be trivial to forecast breast cancer mortality rates into the future using the APC models developed in this study, these predictions would be surrounded by large error for at least two reasons. First, the coronavirus disease 2019 (COVID-19) pandemic has dramatically increased deaths from flu-like illness worldwide. Deaths reported to the US Centers for Disease Control by May 23, 2020 due to COVID-19 among females (44,823) exceeded the expected total number of deaths due to breast cancer among females in the US for the entire year (42,170);1,31 racial and ethnic groups have not experienced illness and deaths due to COVID-19 equally, either from the virus itself or its impacts on the prevention, diagnosis, and treatment of other health conditions including breast cancer. Second, and conversely, emerging advances in breast cancer screening and treatment are expected to continue to improve survival after a breast cancer diagnosis. Increased efforts will be necessary to ensure that new approaches, for example, abbreviated magnetic resonance imaging for breast cancer screening and molecularly-targeted breast cancer therapies, equally benefit women of all racial and ethnic groups.
This study extends work by our group and others that describe breast cancer mortality rates by age, period and cohort, for the purpose of facilitating identification of etiologic factors and healthcare interventions influencing disease burden.20,32 Unlike these previous studies, the current work describes patterns by race/ethnicity rather than for all racial and ethnic groups combined. We have also focused on the percent of all deaths due to breast cancer, which is affected by uneven progress in reducing mortality from non-breast cancer health conditions for women of different races and Hispanic ethnicity. Our analysis is limited by the data available for women from disenfranchised populations, which are subject to political, social, and cultural changes in definitions of racial and ethnicity categories and data reporting practices.33 Because of the smaller number of deaths that would make confidence intervals very wide and rate estimates uninformative and due to data unavailability, we did not estimate values for Hispanic Black women, Asian/Pacific Islanders according to Hispanic ethnicity, or for Asian and Hispanic women according to country of origin. Black, Hispanic, and indigenous women are more likely to be under-counted and misclassified by race and ethnicity.34–36 Breast cancer incidence and mortality for women in the U.S. varies according to specific Asian, Central American, and South American countries of origin,37,38 reflecting the need to better characterize the breast cancer mortality burden by subgroups within these larger racial and ethnic categories. Our study is also limited since we did not examine differences in stages at diagnosis, patterns of breast cancer care, or molecular subtypes of breast cancer across racial and ethnic subgroups; distributions of breast cancer subtypes vary across racial/ethnic subgroups and are associated with different mortality patterns due to the availability of targeted treatments.
Racial and ethnicity comparisons are critical for identifying population subgroups that experience inappropriate variation in care or comparatively poor health outcomes, which are often the result of structural racism.15 Strategies to eliminate racial disparities in breast cancer outcomes are more likely to be effective if they consider multiple stakeholders with influence over structural factors including clinicians, community organizations, public health departments, and healthcare institutions.39 Furthermore, interventions and health system that redesign efforts to address structural racism and health disparities need to jointly consider both race and ethnicity.40 Differences in health outcomes between race and ethnic groups, for example, the high percentage of triple negative breast cancer and worse survival after breast cancer among Black as compared with White women, should not be assumed to simply reflect inherent biological distinctions or divert resources away from addressing the structural reasons for these differences.41
Declines in overall breast cancer mortality rates reflect that improvements in breast cancer screening and treatment have impacted the population-level burden of breast cancer, but that racial and ethnic minority women have been excluded from benefiting at the same pace or degree as non-Hispanic White women especially among the oldest age groups. These results also underscore that progress in reducing the burden of breast cancer depends on how this burden is measured, since a relative measure of breast cancer mortality described by the proportion of all deaths attributable to breast cancer show that this proportion is now greater for non-Hispanic Black, Asian/Pacific Islander, and Hispanic women than for non-Hispanic White women for the first time in 50 years.
Supplementary Material
eAppendix 1. Technical appendix that describes estimation for all-cause mortality, breast cancer mortality, and mortality for other causes for all race/ethnicity groups combined and separately (Hispanic women [all races], White women, non-Hispanic White women, Black women, non-Hispanic Black women, American Indian/Alaska Native women, and Asian/Pacific Islander women).
eFigure 1. Age-adjusted all-cause mortality rates per 100,000 women according to year of death and race/ethnicity, 1968–2019, United States.
eFigure 2. Age-adjusted circulatory disease mortality rates per 100,000 women according to year of death and race/ethnicity, 1968–2019, United States.
eFigure 3. Age-adjusted mortality rates for all cancer types other than breast cancer per 100,000 women according to year of death and race/ethnicity, 1968–2019, United States.
eFigure 4. Age-adjusted respiratory disease mortality rates per 100,000 women according to year of death and race/ethnicity, 1968–2019, United States.
eTable 1. All-cause mortality rates per 100,000 women by age, race/ethnicity, and calendar year of death for 1979, 1999, and 2019.
Acknowledgements:
The authors acknowledge Julie McGregor for study support and John Hampton for assistance with data at the University of Wisconsin-Madison, and Dr. Jeanne Mandelblatt at Georgetown University for her coordination of the CISNET Breast Working Group.
Sources of financial support:
This work was supported by grants from the National Cancer Institute at the National Institutes of Health including support of the Cancer Intervention and Simulation Modeling Network (CISNET) (grant numbers U01 CA253911, U01 CA199218, U01 CA199218-S1, P30 CA014520, P30 CA046592); and faculty start-up funds from the University of Michigan to Dr. Christina Chapman. The funders had no role in the design of the study, the collection, analysis, or interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication.
Footnotes
Conflicts of interest: The authors declare no conflict of interest.
Ethics approval: Only publicly available anonymous data were used; this study was exempt from review as determined by the University of Wisconsin Health Sciences Institutional Review Board.
Process of obtaining code and data: Data are available by download from websites described in references 22 (CDC WONDER Online Databases, http://wonder.cdc.gov) and 24 (Berkely Mortality Database, http://u.demog.berkeley.edu/~bmd/states.html). Model code and output is available by download at https://github.com/rgangnon/NonBreastCancerMortality_by_Race
References
- 1.Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin 2020;70(1):7–30. [DOI] [PubMed] [Google Scholar]
- 2.DeSantis CE, Ma J, Goding Sauer A, Newman LA, Jemal A. Breast cancer statistics, 2017, racial disparity in mortality by state. CA Cancer J Clin 2017;67(6):439–448. [DOI] [PubMed] [Google Scholar]
- 3.Cronin KA, Lake AJ, Scott S, Sherman RL, Noone AM, Howlader N, Henley SJ, Anderson RN, Firth AU, Ma J, Kohler BA, Jemal A. Annual Report to the Nation on the Status of Cancer, part I: National cancer statistics. Cancer 2018;124(13):2785–2800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Harper S, Lynch J, Meersman SC, Breen N, Davis WW, Reichman MC. Trends in area-socioeconomic and race-ethnic disparities in breast cancer incidence, stage at diagnosis, screening, mortality, and survival among women ages 50 years and over (1987–2005). Cancer Epidemiol Biomarkers Prev 2009;18(1):121–31. [DOI] [PubMed] [Google Scholar]
- 5.Ademuyiwa FO, Gao F, Hao L, Morgensztern D, Aft RL, Ma CX, Ellis MJ. US breast cancer mortality trends in young women according to race. Cancer 2015;121(9):1469–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Yasmeen F, Hyndman RJ, Erbas B. Forecasting age-related changes in breast cancer mortality among white and black US women: a functional data approach. Cancer Epidemiol 2010;34(5):542–9. [DOI] [PubMed] [Google Scholar]
- 7.Davis Lynn BC, Rosenberg PS, Anderson WF, Gierach GL. Black-White Breast Cancer Incidence Trends: Effects of Ethnicity. J Natl Cancer Inst 2018;110(11):1270–1272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hall IJ, Tangka FKL, Sabatino SA, Thompson TD, Graubard BI, Breen N. Patterns and Trends in Cancer Screening in the United States. Prev Chronic Dis 2018;15:E97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Press R, Carrasquillo O, Sciacca RR, Giardina EG. Racial/ethnic disparities in time to follow-up after an abnormal mammogram. J Womens Health (Larchmt) 2008;17(6):923–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Jones BA, Dailey A, Calvocoressi L, Reams K, Kasl SV, Lee C, Hsu H. Inadequate follow-up of abnormal screening mammograms: findings from the race differences in screening mammography process study (United States). Cancer Causes Control 2005;16(7):809–21. [DOI] [PubMed] [Google Scholar]
- 11.Nguyen KH, Pasick RJ, Stewart SL, Kerlikowske K, Karliner LS. Disparities in abnormal mammogram follow-up time for Asian women compared with non-Hispanic white women and between Asian ethnic groups. Cancer 2017;123(18):3468–3475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chavez-MacGregor M, Clarke CA, Lichtensztajn DY, Giordano SH. Delayed Initiation of Adjuvant Chemotherapy Among Patients With Breast Cancer. JAMA Oncol 2016;2(3):322–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Gold HT, Thwin SS, Buist DS, Field TS, Wei F, Yood MU, Lash TL, Quinn VP, Geiger AM, Silliman RA. Delayed radiotherapy for breast cancer patients in integrated delivery systems. Am J Manag Care 2009;15(11):785–9. [PMC free article] [PubMed] [Google Scholar]
- 14.Chen L, Li CI. Racial disparities in breast cancer diagnosis and treatment by hormone receptor and HER2 status. Cancer Epidemiol Biomarkers Prev 2015;24(11):1666–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bailey ZD, Krieger N, Agenor M, Graves J, Linos N, Bassett MT. Structural racism and health inequities in the USA: evidence and interventions. Lancet 2017;389(10077):1453–1463. [DOI] [PubMed] [Google Scholar]
- 16.Brawley OW. Is race really a negative prognostic factor for cancer? J Natl Cancer Inst 2009;101(14):970–1. [DOI] [PubMed] [Google Scholar]
- 17.Bonham VL, Green ED, Pérez-Stable EJ. Examining How Race, Ethnicity, and Ancestry Data Are Used in Biomedical Research. JAMA 2018;320(15):1533–1534. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Henley SJ, Ward EM, Scott S, Ma J, Anderson RN, Firth AU, Thomas CC, Islami F, Weir HK, Lewis DR, Sherman RL, Wu M, Benard VB, Richardson LC, Jemal A, Cronin K, Kohler BA. Annual report to the nation on the status of cancer, part I: National cancer statistics. Cancer 2020;126(10):2225–2249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Harper S, Lynch J. Methods for Measuring Cancer Disparities: Using Data Relevant to Healthy People 2010 Cancer-Related Objectives. NCI Cancer Surveillance Monograph Series, Number 6. Bethesda, MD, 2005. NIH Publication; No. 05–5777. [Google Scholar]
- 20.Gangnon RE, Stout NK, Alagoz O, Hampton JM, Sprague BL, Trentham-Dietz A. Contribution of Breast Cancer to Overall Mortality for US Women. Med Decis Making 2018;38(1_suppl):24S–31S. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Rosenberg MA. The impact of mammography and adjuvant therapy on US breast cancer mortality (1975–2000): collective results from the Cancer Intervention and Surveillance Modeling Network. Competing risks to breast cancer mortality. J Natl Cancer Inst Monogr 2006;36:15–19. [DOI] [PubMed] [Google Scholar]
- 22.Centers for Disease Control and Prevention, U.S. Department of Health & Human Services. WONDER Online Databases, Underlying Cause of Death. http://wonder.cdc.gov, AccessedDecember 1, 2016.
- 23.Friede A, Reid JA, Ory HW. CDC WONDER: a comprehensive on-line public health information system of the Centers for Disease Control and Prevention. Am J Public Health 1993;83(9):1289–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Berkely Mortality Database. Data for the United States. http://u.demog.berkeley.edu/~bmd/states.html, December 1, 2016.
- 25.Carstensen B Age-Period-Cohort models for the Lexis Diagram. Stat Med 2007;26:3018–3045. [DOI] [PubMed] [Google Scholar]
- 26.Wood SN. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J Royal Stat Soc (B) 2011;73(1):3–36. [Google Scholar]
- 27.Wood SN. Thin-plate regression splines. J Royal Stat Soc (B) 2003;65(1):95–114. [Google Scholar]
- 28.Craven P, Wahba G. Smoothing noisy data with spline functions: estimating the correct degree of smoothing by the method of generalized cross-validation. Numer Math 1979;31:377–403. [Google Scholar]
- 29.Wickham H ggplot2: Elegant Graphics for Data Analysis. New York: Springer-Verlag, 2009. [Google Scholar]
- 30.R Core Team. R: A language and environment for statistical computing. Vienna, Austria. http://www.R-project.org/R Foundation for Statistical Computing, 2020. AccessedJuly 1, 2020. [Google Scholar]
- 31.National Center for Health Statistics. Provisional COVID-19 death counts by sex, age, and week. Accessed6/24/2020ed, 2020. Available at: https://www.cdc.gov/nchs/nvss/vsrr/covid_weekly/.
- 32.Holford TR, Chen HS, Annett D, Krapcho M, Dorogaeva A, Feuer EJ. CP*Trends: An Online Tool for Comparing Cohort and Period Trends Across Cancer Sites. Am J Epidemiol 2019;188(7):1361–1370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Pratt BM, Hixson L, Jones NA. Measuring race and ethnicity aross the decades, 1790–2010. U.S. Department of Commerce, Economics and Statistics Administration, U.S. Census Bureau Available at: https://www.census.gov/data-tools/demo/race/MREAD_1790_2010.html AccessedDecember 10, 2018. [Google Scholar]
- 34.Robinson JG, Ahmed B, Dasgupta P, Woodrow KA. Estimation of Population Coverage in the 1990 United-States Census Based on Demographic-Analysis. Journal of the American Statistical Association 1993;88(423):1061–1071. [PubMed] [Google Scholar]
- 35.Adlakha AL, Robinson JG, West KK, Bruce A. Assessment of Consistency of Census Data with Demographic Benchmarks at the Subnational Level. Census 2000 Evaluation Series O.20 Washington, D.C.: U.S. Census Bureau. Available at: https://www.census.gov/pred/www/rpts/O.20.pdf, 2003. [Google Scholar]
- 36.Davis PP, Mulligan J. Census Coverage Measurement Estimation Report: Net Coverage for the Household Population in the United States. DSSD 2010 Census Coverage Measurement Memorandum Series #2010-G-03. Washington, D.C.: U.S. Census Bureau. Available at: https://www.census.gov/coverage_measurement/pdfs/g03.pdf, May22, 2012. [Google Scholar]
- 37.Gomez SL, Quach T, Horn-Ross PL, Pham JT, Cockburn M, Chang ET, Keegan TH, Glaser SL, Clarke CA. Hidden breast cancer disparities in Asian women: disaggregating incidence rates by ethnicity and migrant status. Am J Public Health 2010;100Suppl 1:S125–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Hunt BR. Breast Cancer Prevalence and Mortality among Hispanic Subgroups in the United States, 2009–2013. J Cancer Epidemiol 2016;2016:8784040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Pallok K, De Maio F, Ansell DA. Structural Racism - A 60-Year-Old Black Woman with Breast Cancer. N Engl J Med 2019;380(16):1489–1493. [DOI] [PubMed] [Google Scholar]
- 40.LaVeist-Ramos TA, Galarraga J, Thorpe RJ Jr., Bell CN, Austin CJ. Are black Hispanics black or Hispanic? Exploring disparities at the intersection of race and ethnicity. J Epidemiol Community Health 2012;66(7):e21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Brawley OW. Some thoughts on health surveillance data, race, and population categorization. CA Cancer J Clin 2016;66(3):179–81. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eAppendix 1. Technical appendix that describes estimation for all-cause mortality, breast cancer mortality, and mortality for other causes for all race/ethnicity groups combined and separately (Hispanic women [all races], White women, non-Hispanic White women, Black women, non-Hispanic Black women, American Indian/Alaska Native women, and Asian/Pacific Islander women).
eFigure 1. Age-adjusted all-cause mortality rates per 100,000 women according to year of death and race/ethnicity, 1968–2019, United States.
eFigure 2. Age-adjusted circulatory disease mortality rates per 100,000 women according to year of death and race/ethnicity, 1968–2019, United States.
eFigure 3. Age-adjusted mortality rates for all cancer types other than breast cancer per 100,000 women according to year of death and race/ethnicity, 1968–2019, United States.
eFigure 4. Age-adjusted respiratory disease mortality rates per 100,000 women according to year of death and race/ethnicity, 1968–2019, United States.
eTable 1. All-cause mortality rates per 100,000 women by age, race/ethnicity, and calendar year of death for 1979, 1999, and 2019.