. Author manuscript; available in PMC: 2012 Jul 15.
Published in final edited form as:Cancer. 2011 Jan 24;117(14):3242–3251. doi:
10.1002/cncr.25854 Abstract BackgroundThis is the first study to use the linked National Longitudinal Mortality Studies (NLMS) and Surveillance, Epidemiology and End Results (SEER) data to determine the effects of individual-level socioeconomic factors (health insurance, education, income, and poverty status) on racial disparities in receiving treatment and in survival.
MethodsThis study included 13,234 cases diagnosed with eight most common types of cancer (female breast, colorectal, prostate, lung and bronchus, uterine cervix, ovarian, melanoma, and urinary bladder) at age ≥25, identified from the NLMS-SEER data in 1973–2003. Kaplan-Meier methods and Cox regression models were used for survival analysis.
ResultsThree-year all-cause observed survival for cases diagnosed with local-stage cancers of the eight leading tumors combined was ≥82% regardless of race/ethnicity. More favorable survival was associated with higher socioeconomic status. Compared to whites, blacks were less likely to receive first-course cancer-directed surgery, perhaps reflecting a less favorable stage distribution at diagnosis. Hazard ratio for cancer-specific mortality was significantly higher among blacks compared to whites (hazard ratio=1.2, 95%CI=1.1–1.3) after adjusting for age, sex and tumor stage, but not after further controlling for socioeconomic factors and treatment (1.0, 95%CI=0.9–1.1). Hazard ratios for all-cause mortality among patients with breast cancer and for cancer-specific mortality in patients with prostate cancer were significantly higher for blacks compared to whites after adjusting for socioeconomic factors, treatment, patient and tumor characteristics.
ConclusionsFavorable survival was associated with higher socioeconomic status. Racial disparities in survival persisted after adjusting for individual-level socioeconomic factors and treatment for patients with breast and prostate cancer.
Keywords: Cancer, socioeconomic status, racial disparities, treatment, survival
IntroductionRacial/Ethnic disparities in health care and outcomes have been evident for almost all cancer sites as indicated by the National Cancer Institute’s Surveillance, Epidemiology and End Results (SEER) Annual Cancer Statistics Review,1 cancer facts and figures presented by the American Cancer Society,2 and other studies.3–20 For example, the annual review of the 1973–2007 data showed that blacks had a higher mortality for breast, colorectal, lung, prostate and many other common tumors than whites.1 The increased mortality in blacks with cancer can be attributed to more aggressive cancers and more advanced stage-at-diagnosis,1,2,20 differences in treatment,10–13 socioeconomic factors,8,10–13 physician characteristics,9 and personal beliefs.21 There have been numerous original studies and meta-analyses on racial disparities in survival, and the results are not consistent.11–13 Some studies demonstrated that if patients had equal access to quality health care, the outcomes would be similar among different racial groups.10–13 However, other studies showed that racial disparities still existed even after controlling for socioeconomic factors and for access to equitable care and treatment.10–13
Many of these studies examined one or several specific tumor sites and few studies reported all or multiple tumor sites from the same cohorts of population on racial disparities in survival, treatment and socioeconomic factors. This study presents the recently linked data between the 30 cohorts of the National Longitudinal Mortality Studies (NLMS) and SEER cancer registries. We aimed to determine the effect of socioeconomic factors at the individual level (i.e., health insurance, education, income and poverty) on racial disparities in receiving treatment and in survival among patients diagnosed with cancer. This study examined racial disparities on 8 specific types of tumor and also on all 8 tumors combined. We hypothesized that patients with no insurance or with lower socioeconomic status were less likely to receive the recommended therapy compared to those with private health insurance and those with higher socioeconomic status, and that racial disparities in treatment were largely explained by differences in health insurance status and socioeconomic factors. We also hypothesized that patients with no insurance or with lower socioeconomic status would experience less favorable survival (all-cause and cancer-specific) compared to those with private health insurance and higher socioeconomic status and that racial disparities in survival were largely explained by differences in health insurance, socioeconomic status, and treatment rendered.
Materials and Methods Data Sources and Study PopulationThis study utilized the SEER-NLMS linked data for cases in 15 participating SEER registries between 1973 and 2003. The detailed methods for this data linkage were described elsewhere.20,22,23 In brief, the SEER-NLMS linkage was conducted by the Census Bureau and the linked dataset is maintained by the Census Bureau in compliance with registry and federal requirements to protect health information of human research subjects.
The SEER registries ascertain all newly diagnosed incident invasive cancer cases from multiple reporting sources.24 Information includes tumor location, stage, size and grade; demographic characteristics such as age, gender, race and marital status; and types of first course treatment provided within four to six months after the date of diagnosis.26 The NLMS is an ongoing mortality follow-up study of selected cohorts from the Census Bureau’s Current Population Survey (CPS) respondents and Census sample. To date, the 30 cohorts in the NLMS were sampled and surveyed in March 1973, 1979, and 1981 through 2003, with additional surveys in February 1978, April, August and December 1980, September 1985, as well as the 1980 Census E Sample. We studied the eight most common tumors (breast, colorectal, prostate, lung and bronchus, cervix, ovarian, melanoma of the skin, and urinary bladder cancer) because of relatively large numbers of cases which provided informative results for most stratified analyses by race/ethnicity and socioeconomic status (SES).
All cases were diagnosed with a primary malignant cancer in one of 15 SEER registries. Years of cancer diagnosis varied across registries. In the SEER 9 registries (San Francisco Bay Area, Connecticut, Metropolitan Detroit, Hawaii, Iowa, New Mexico, Seattle (Puget Sound), Utah, and Metropolitan Atlanta) cases were diagnosed from 1979 through 2003. In the San Jose-Monterey and Los Angeles registries cases were diagnosed from 1992 through 2003. In the Greater California, Kentucky, Louisiana, and New Jersey registries cases were diagnosed from 2000 through 2003. A total of 13,620 cases met National Longitudinal Mortality Study requirements for matching to the National Death Index and were 25 years of age or older at the time of their Current Population Survey. Of these cases, 386 were excluded from analysis due to incomplete racial and ethnicity data, resulting in an analytic dataset of 13,234 cases.
Socio-demographicsRace/ethnicity was classified into non-Hispanic white, non-Hispanic black, non-Hispanic Asian or Pacific Islander, Hispanic, and American-Indian or Alaskan-Natives. Age at diagnosis was categorized in 10 year intervals. Socioeconomic variables included health insurance, years of education, family income, and poverty status. Health insurance was originally categorized into employer health care, government, Medicare, private company, Medicaid, uninsured, and unknown or missing. The year of education was classified into less than high school (<12 years), high school graduate (12 years), and some post high school education (≥13 years). There were 3 cases with missing information on education that were not reported in the results. Family income referred to total combined income of all family members during the 12 months preceding the survey20 and the dollar amount or the median value of the category of income was adjusted to the year 1990 dollars by the appropriate CPI value for inflation in individuals from various NLMS cohorts. We categorized the family income as <$10,000, $10,000–$34,999, ≥$35,000, and unknown/missing. Poverty status was measured as of the 1990 census in terms of the ratio of the family income to the poverty threshold for a four-person family20 and grouped into ≤100% (lowest), 100–400%, ≥400%, and unknown or missing.
Tumor characteristics and TreatmentTumor characteristics and treatment variables were obtained from the SEER data. The 8 common tumor sites included breast (female), colorectal, prostate, lung and bronchus, cervix, ovarian, melanoma of the skin and urinary bladder cancer. Because we included cases with cancer from 1979 through 2003, SEER historic tumor stages were analyzed rather than the American Joint Committee on Cancer Stage which was available since 1988. First-course cancer-directed surgery was defined according to the SEER surgery codes, mostly those with codes of >10. Specific codes for common tumor sites included: breast (10–90), colorectal (30–90), prostate (20–90), lung and bronchus (30–90), cervical (20–90), ovarian (10–90), melanoma of the skin (20–90), and urinary bladder (10–90). First-course radiation therapy was defined from the SEER variable ‘radiation therapy’ as yes or no.
SurvivalThe observed survival time in months was calculated from the date of cancer diagnosis to the date of death or to the date of last follow-up (December 31, 2003). All-cause mortality was defined as death from any cause as provided in the SEER registry data. Patients still alive at the last date of follow-up were censored. The cancer-specific mortality was defined if cancer (of any type) was the underlying cause of death. In this specific analysis, patients who died of causes other than cancer or were still alive at the date of last follow-up were censored. The 3-year observed survival rate was calculated as the proportion of patients who survived for at least three years among those cases who were followed up for at least 3 years after the date of cancer diagnosis using the Kaplan-Meier product limit methods.
AnalysisAll analyses were weighted according to data size, number of cohorts, and U.S. populations during study period. Differences in the distribution of baseline characteristics among the racial/ethnic groups were tested using the chi-square statistic. Multivariable logistic regression analyses were used to assess the odds ratio of receiving various therapies in association with race/ethnicity while adjusting for age, sex, and tumor stage, and by additionally adjusting for socioeconomic variables (education, family income and poverty status) and health insurance in those without missing data in health insurance. In the Cox proportional hazard regression analyses of survival, the hazard ratio of all-cause or cancer-specific mortality was presented with race/ethnicity in the models while adjusting for age, sex, and tumor stage, and by additionally adjusting for socioeconomic variables (education, family income and poverty status), treatment (cancer-directed surgery and radiation therapy), and health insurance in those with usable data in health insurance.
ResultsTable-1 presents the distribution of patient age, sex, tumor stage, and SES (health insurance, years of education, family income and poverty status) among racial/ethnic groups of patients diagnosed with 8 most common types of cancer. A higher proportion of cases were diagnosed at age <45 among Hispanics (5.9%) compared to non-Hispanic whites (4.0%). A slightly higher proportion of cases were men than women. A larger proportion of black cases were diagnosed with distant stage cancer, while greater percentages of cases were diagnosed with localized stage among whites and among Asians and Pacific Islanders. A greater proportion of non-Hispanic blacks, American-Indians and Alaska-Natives, and Hispanics had Medicaid coverage or no health insurance, while higher percentages of Asian and white patients had employer or private health care. Larger proportions of black and Hispanic cases were in the lowest categories of educational attainment, family income and poverty status compared to whites. For example, 44.5% of Blacks and 52.4% of Hispanics had less than a high school diploma compared to 23.6% of whites. The differences in the distribution of the above factors between ethnic groups were all statistically significant.
Table 1.Comparison of demographic and tumor characteristics by race/ethnicity
Characteristics NH-White NH-Black NH-API Hispanic AI/AN n % n % n % n % n % Age (years) <45 425 4.0 30 3.3 37 5.5 49 5.9 - - 45–54 1220 11.3 119 12.9 75 11.1 109 13.0 - - 55–64 2330 21.7 266 28.9 138 20.4 194 23.2 - - 65–74 3451 32.1 294 31.9 217 32.0 292 34.9 - - 75–84 2581 24.0 175 19.0 169 24.9 151 18.1 - - 85+ 742 6.9 38 4.1 42 6.2 41 4.9 - - Sex Male 5767 53.7 507 55.0 346 51.0 456 54.5 25 51.0 Female 4982 46.3 415 45.0 332 49.0 380 45.5 24 49.0 Tumor stage Localized* 3645 33.9 212 23.0 249 36.7 239 28.6 - - Regional 2398 22.3 191 20.7 164 24.2 191 22.8 - - Distant 1659 15.4 193 20.9 115 17.0 135 16.1 - - Localized/regional (Prostate) 2238 20.8 246 26.7 127 18.7 219 26.2 - - Unstaged/Missing 809 7.5 80 8.7 23 3.4 52 6.2 - - Health Insurance Employer/Medicare/Private/government 4604 42.8 343 34.2 329 48.5 373 44.6 24 49.0 Medicaid/Not Insured 469 4.4 66 7.2 43 6.3 108 12.9 - - Unknown 5676 52.8 513 55.6 306 45.1 355 42.5 - - Years of Education <12 2538 23.6 410 44.5 196 28.9 438 52.4 25 51.0 12 4050 37.7 263 28.5 250 36.9 239 28.6 - - ≥13 4158 38.7 249 27.0 232 34.2 159 19.0 - - Family income quintile <$10,000 878 8.2 193 20.9 55 8.1 144 17.2 18 36.7 $10,000–$34,999 3919 36.5 373 40.5 233 34.4 381 45.6 - - ≥$35000 5365 49.9 302 32.8 363 53.5 270 32.3 - - Unknown 587 5.5 54 5.9 27 4.0 41 4.9 - - Poverty status quintile ≤100% 618 5.7 162 17.6 58 8.6 152 18.2 17 34.7 100–400% 4833 45.0 467 50.7 317 46.8 456 54.5 17 34.7 ≥400% 4297 40.0 211 22.9 286 42.2 168 20.1 - - Unknown 1001 9.3 82 8.9 17 2.5 60 7.2 - - Total 10749 100 922 100 678 100 836 100 49 100Table-2 presents the percentage of cases receiving first course of cancer-directed surgery and radiation therapy by tumor stage, ethnicity and SES. A slightly larger proportion of white and Asian and Pacific Islanders cases with local stage tumors received cancer-directed surgery compared to Hispanics and blacks. Receipt of surgery declined for all racial and ethnic groups among those with distant stage cancer. The receipt of cancer-directed surgery and radiotherapy were generally higher among those with employer or private insurance or Medicare and those with more education, and lower among those with Medicaid or no insurance or those with less education. Receipt of surgery and radiation varied across sites, reflecting the unique clinical feature of each tumor site.
Table 2.Percentage of cases receiving surgery and radiotherapy by race and SES, stratified by tumor stage
Race/ethnicity Local Regional Local/regional (Prostate) Distant Surgery Radiation Surgery Radiation Surgery Radiation Surgery Radiation NH-White 80.7 20.0 67.3 33.9 33.4 35.1 29.1 32.1 NH-Black 77.8 22.6 58.6 38.7 33.3 30.5 23.8 31.1 NH-API 82.3 25.3 73.2 36.6 26.8 47.2 31.3 32.2 Hispanic 77.0 25.9 73.3 36.1 38.4 26.6 40.7 25.9 AI/AN 76.9 38.5 60.0 33.3 18.2 45.5 50.0 12.5 Health Insurance Employer/Medicare/Private 78.9 22.7 65.8 35.1 31.2 35.1 27.6 31.4 Government 72.2 5.6 73.9 17.4 33.3 27.8 14.3 50 Medicaid/Not-insured 74.4 22.8 63.7 34.7 29.3 33.3 25.9 36.7 Unknown 82.5 19.1 69.0 34.2 35.7 34.2 31.7 30.7 Years of Education† <12 77.6 17.1 61.3 31.2 25.4 30.6 27.7 30.6 12 80.3 19.5 66.1 36.8 31.8 37.2 29.1 36.9 ≥13 82.2 24.3 74.5 34.5 38.7 34.8 32.2 34.7 Tumor Sites Breast 88.8 40.4 91.3 38.1 52.4 37.4 Colorectal 70.0 6.0 91.1 15.5 64.5 13.2 Prostate* 33.4 34.6 4.1 18.7 Lung and bronchus 46.5 21.7 20.6 49.2 3.5 49.7 Cervix 78.6 30.8 36.4 92.7 16.7 55.6 Ovarian 94.0 0.0 81.8 0.0 71.4 1.5 Melanoma of the skin 87.6 0.8 74.5 3.9 45.0 25 Urinary bladder 85.0 1.8 84.5 24.2 61.5 23.1Table-3 presents survival rates by racial/ethnic groups, SES and tumor sites. The overall observed 3-year survival for all cases with local stage tumor was 85% for whites, 82% for blacks, 85% for Asian and Pacific Islanders, 90% for Hispanics, and 86% for American-Indians/Alaska-Natives. As expected, the survival rate was lower for advanced stage tumors, but the general patterns by ethnicity were similar for all-cause and for cancer-specific survival. More favorable survival was experienced by patients with health insurance and higher education compared to cases without health insurance and lower education. Percent of three-year observed survival was relatively high among those with melanoma, breast or prostate cancer, and low among cases with lung cancer. The survival rate was high for those with early stage cervical and ovarian cancer, but low among those diagnosed with late stage cancer of these sites.
Table 3.Observed 3-year survival rate by race and SES, stratified by tumor stage
Observed 3-year survival rate (95% CI) Race/ethnicity and SES Local Stage Regional Stage Distant Stage Unstaged/Missing All-cause Cancer-specific All-cause Cancer-specific All-cause Cancer-specific All-cause Cancer-specific Race/ethnicity NH-White 85(84–86) 92(91–93) 49(47–52) 55(53–57) 13(11–15) 14(12–16) 36(33–39) 42(39–46) NH-Black 82(79–86) 88(85–91) 42(35–49) 47(40–54) 16(11–20) 17(12–22) 37(27–47) 44(33–55) NH-API 85(80–89) 91(87–95) 59(49–68) 64(54–74) 20(11–30) 23(13–33) 34(12–56) 45(18–72) Hispanic 90(86–93) 94(91–97) 54(45–63) 62(52–71) 14(7–22) 15(8–23) 39(23–55) 43(26–60) AI/AN 86(68–100) 94(82–100) 26(3–50) 28(3–53) 0 0 21(0–100) 21(0–100) Health Insurance Employer/Medicare/Private 85(83–86) 92(91–93) 50(47–53) 55(52–58) 14(12–16) 15(12–17) 38(32–43) 44(38–50) Government 83(69–97) 84(71–98) 42(19–65) 51(25–76) 0 0 0 0 Medicaid/Not-insured 76(70–81) 85(80–90) 47(40–55) 51(43–59) 8(3–12) 9(4–14) 29(14–44) 33(16–50) Unknown 85(84–87) 93(92–93) 49(46–52) 55(52–58) 14(12–16) 15(13–18) 36(32–40) 43(38–47) Years of Education† <12 74(72–77) 85(83–87) 39(35–42) 44(40–47) 10(7–12) 11(8–13) 31(27–36) 38(32–43) 12 85(84–87) 92(91–93) 49(46–52) 55(52–58) 13(11–16) 14(11–16) 40(34–45) 46(40–52) ≥13 90(89–91) 95(95–96) 59(56–62) 65(61–68) 18(15–21) 20(16–23) 39(33–45) 46(40–53) Tumor Sites Breast 90(89–92) 96(95–97) 82(80–85) 88(86–90) 29(22–36) 32(24–40) 48(37–59) 63(51–76) Colorectal 81(78–84) 91(88–93) 60(57–64) 67(63–70) 11(8–14) 12(9–15) 33(24–42) 40(30–50) Prostate* 89(88–91) 96(95–96) - - 37(30–44) 41(34–48) 63(57–68) 76(70–81) Lung and bronchus 43(38–48) 49(44–55) 15(13–17) 17(14–19) 3(2–4) 3(2–4) 10(7–13) 11(8–15) Cervix 89(81–97) 93(87–100) 59(45–74) 63(51–78) 12(0–26) 12(0–26) 50(14–86) 56(19–94) Ovarian 88(80–96) 94(88–100) 74(54–95) 74(49–77) 34(28–39) 37(30–43) 22(4–39) 24(6–43) Melanoma of the skin 91(89–94) 96(94–98) 49(34–65) 56(39–73) 8(0–22) 9(0–23) 90(75–100) 93(81–100) Urinary bladder 78(75–81) 89(86–92) 28(21–35) 37(28–46) 9(0–20) 10(0–22) 65(48–83) 74(56–91)Table-4 presents results of the multivariate logistic regression analysis for receiving cancer-directed surgery and radiation therapy by socioeconomic attributes and by tumor site. Black cases were less likely to receive cancer-directed surgery than whites, even after controlling for education, income and poverty status at the individual level in addition to other patient and tumor characteristics (odds ratios=0.7, 95% CI=0.6–0.8), and further adjusting for health insurance (odds ratios=0.5, 95% CI=0.4–0.6). There were no significant differences among ethnic groups for receiving radiation therapy after adjusting for socioeconomic factors. The results were similar in tumor site-specific analyses, but the confidence intervals were wider due to smaller samples.
Table 4.Receipt of cancer-directed surgery and radiation-therapy, by tumor site and race
Cancer Site Race/Ethnicity Cancer-directed Surgery Radiotherapy Model-1 (n=13,234) Model-2 (n=6,367) Model-1 (n=13,234) Model-2 (n=6,367) No. OR(95% CI)* No. OR(95% CI) OR(95% CI) OR(95% CI) All 8 cancers combined NH-White 10749 1.0(Ref)* 5073 1.0(Ref) 1.0(Ref) 1.0(Ref) NH-Black 922 0.7(0.6–0.8) 409 0.5(0.4–0.6) 1.0(0.8–1.2) 1.1(0.9–1.5) NH-API 678 1.1(0.9–1.4) 372 1.2(0.9–1.7) 1.1(0.9–1.4) 1.2(0.9–1.6) Hispanic 836 1.1(0.9–1.3) 481 1.1(0.8–1.3) 1.0(0.8–1.2) 1.1(0.8–1.4) AI/AN 49 0.9(0.4–1.9) 32 0.9(0.4–2.2) 0.8(0.3–1.7) 0.8(0.3–2.1) Breast NH-White 2492 1.0(Ref) 1169 1.0(Ref) 1.0(Ref) 1.0(Ref) NH-Black 214 0.7(0.5–1.2) 86 0.7(0.3–1.3) 0.7(0.5–1.0) 0.8(0.4–1.3) NH-API 191 1.3(0.6–2.7) 115 2.0(0.8–5.1) 1.0(0.7–1.5) 1.1(0.7–1.8) Hispanic 216 0.9(0.5–1.5) 129 1.5(0.8–2.7) 1.2(0.9–1.7) 1.2(0.8–2.0) Colorectal NH-White 1791 1.0(Ref) 791 1.0(Ref) 1.0(Ref) 1.0(Ref) NH-Black 140 0.7(0.5–1.1) 56 0.5(0.3–1.0) 0.9(0.5–1.5) 0.6(0.2–1.5) NH-API 145 1.0(0.6–1.7) 75 1.2(0.6–2.3) 1.1(0.6–2.1) 1.3(0.6–3.4) Hispanic 166 0.5(0.3–0.8) 98 0.6(0.3–1.0) 1.7(1.0–3.0) 1.7(0.7–3.9) Prostate NH-White 2652 1.0(Ref) 1232 1.0(Ref) 1.0(Ref) 1.0(Ref) NH-Black 311 0.8(0.6–1.1) 137 0.7(0.4–1.1) 0.8(0.6–1.1) 1.0(0.6–1.6) NH-API 152 0.8(0.5–1.3) 80 0.8(0.4–1.5) 1.4(0.9–2.2) 1.5(0.9–2.8) Hispanic 252 1.3(0.9–1.9) 132 0.9(0.5–1.5) 0.7(0.5–1.0) 1.1(0.7–1.8) Lung and bronchus NH-White 2241 1.0(Ref) 1057 1.0(Ref) 1.0(Ref) 1.0(Ref) NH-Black 207 0.8(0.5–1.3) 101 0.5(0.2–1.1) 1.0(0.7–1.3) 1.4(0.8–2.2) NH-API 126 2.0(1.1–3.6) 69 1.8(0.8–4.0) 0.9(0.6–1.5) 1.0(0.6–1.9) Hispanic 125 1.5(0.8–3.5) 69 1.0(0.4–2.2) 0.7(0.6–1.5) 0.7(0.4–1.3) Ovarian NH-White 318 1.0(Ref) 143 1.0(Ref) 1.0(Ref) 1.0(Ref) NH-Black 29 0.9(0.3–2.4) <20 0.6(0.1–2.8) u/d u/d NH-API 17 5.3(0.8–35.0) <20 5.1(0.4–72.4) 10.6(1.3–87.1) u/d Hispanic 38 2.4(0.8–6.8) 21 15.3(1.0–225) 5.5(0.2–198.0) 4.0(0.1–291.9) Urinary bladder NH-White 795 1.0(Ref) 395 1.0(Ref) 1.0(Ref) 1.0(Ref) NH-Black 29 0.5(0.2–1.3) <20 0.7(0.1–4.2) 0.5(0.1–2.1) 1.6(0.3–10.2) NH-API 38 2.0(0.8–5.2) <20 2.2(0.5–9.1) 0.30(0.1–1.3) u/d Hispanic 34 1.9(0.4–9.3) 20 1.0(0.2–5.1) 0.4(0.1–1.6) 0.6(0.1–3.8)Table-5 presents the effect of race/ethnicity, socioeconomic factors and treatment on all-cause and cancer-specific mortality while adjusting for other patient and tumor factors. In an initial model that compared to whites, blacks were significantly more likely to die of cancer after adjusting for age, sex and tumor stage (hazard ratio=1.2, 95% CI=1.1–1.3). The statistical significance of the hazard ratio of cancer-specific mortality was no longer elevated for blacks compared to whites (hazard ratio=1.0, 95% CI=0.9–1.1) after additionally controlling for socioeconomic factors (education, income and poverty status) and treatment. Hazard ratios for cancer-specific mortality among blacks were similar to those for all-cause mortality. Asian and Pacific Islanders and Hispanics appeared to have lower risk of all-cause and disease-specific mortality, whereas mortality among American-Indians and Alaska-Natives was elevated but not significantly different from that in whites after controlling for patient and tumor characteristics, treatment, education, income and health insurance. We also added the registry variable (15 locations) to the final model and found that the hazard ratio of all-cause mortality was among blacks compared to whites was essentially unchanged (hazard ratio=1.03, 95% CI=0.94–1.11).
Table 5.Hazard ratios of mortality by race/ethnicity among patients with invasive tumors
Attributes Hazard ratio (95% CI) of All-Cause Mortality Hazard ratio (95% CI) of Cancer-Specific Mortality No. Model-1 No. Model-2 Model-1 Model-2 Race/ethnicity NH-White 10749 1.0(Referent) 5073 1.0(Referent) 1.0(Referent) 1.0(Referent) NH-Black 922 0.9(0.9–1.0) 409 1.0(0.9–1.1) 1.0(0.9–1.1) 1.0(0.9–1.2) NH-API 678 0.8(0.7–0.9) 372 0.8(0.7–1.0) 0.8(0.6–0.9) 0.7(0.6–1.0) Hispanic 836 0.8(0.7–0.9) 481 0.8(0.7–1.0) 0.8(0.7–1.0) 0.8(0.6–1.0) AI/AN 49 1.5(1.0–2.3) 32 1.5(0.8–2.5) 1.9(1.1–3.1) 1.7(1.0–3.2) Health Insurance Employer/Medicare/Private - - 5595 1.0(Referent) - 1.0(Referent) Government - - 78 1.5(1.1–2.1) - 1.6(1.1–2.3) Medicaid/Not-insured - - 694 1.4(1.2–1.5) - 1.3(1.1–1.4) Years of Education <12 3607 1.3(1.2–1.4) 1605 1.2(1.1–1.4) 1.4(1.3–1.5) 1.3(1.1–1.4) 12 4814 1.2(1.1–1.2) 2306 1.1(1.0–1.2) 1.2(1.1–1.3) 1.2(1.1–1.3) ≥13 4810 1.0(Referent) 2456 1.0(Referent) 1.0(Referent) 1.0(Referent) Annual Family Income <$10,000 1288 1.3(1.2–1.5) 646 1.3(1.1–1.6) 1.2(1.1–1.4) 1.2(1.0–1.5) $10,000–$34,999 4921 1.2(1.1–1.3) 2404 1.2(1.0–1.3) 1.1(1.0–1.2) 1.1(1.0–1.3) ≥$35,000 6314 1.0(Referent) 3024 1.0(Referent) 1.0(Referent) 1.0(Referent) Unknown 711 1.1(1.0–1.2) 293 0.2(0.0–2.2) 1.0(0.9–1.2) 0.3(0.0–2.5) Family poverty status(1990 threshold) ≤100% 1007 1.0(0.9–1.2) 473 0.9(0.7–1.1) 1.0(0.8–1.2) 0.9(0.7–1.1) 100–400% 6090 1.0(1.0–1.1) 1248 1.0(0.9–1.2) 1.1(1.0–1.2) 1.1(0.9–1.2) ≥400% 5982 1.0(Referent) 1721 1.0(Referent) 1.0(Referent) 1.0(Referent) Unknown 155 1.1(1.0–1.2) 2925 4.9(0.5–43.7) 1.0(0.9–1.2) 3.8(0.4–34.5) Surgery(cancer-directed) Yes 7261 1.0(Referent) 3415 1.0(Referent) 1.0(Referent) 1.0(Referent) No 5973 2.2(2.1–2.3) 2952 2.4(2.2–2.6) 2.6(2.4–2.8) 2.8(2.5–3.1) Radiotherapy Yes 3761 1.0(Referent) 1854 1.0(Referent) 1.0(Referent) 1.0(Referent) No 9473 1.1(1.0–1.1) 4513 1.1(1.0–1.2) 1.0(0.9–1.1) 1.0(0.9–1.1)Compared to cases with employer, Medicare, or private health insurance, hazard ratios were significantly elevated among cases with either Medicaid or no health insurance. For example, among those with Medicaid or no insurance, after adjusting for socioeconomic attributes and first course therapy, the hazard ratio of all-cause mortality was 1.4 with 95% CI of 1.2–1.5 and 1.3 with 95% CI of 1.1–1.4 for cancer-specific mortality. There was also a more favorable prognosis associated with higher SES based on education and income, but there was no change in risk of mortality associated with family poverty status (Table-5). Even in a further sensitivity analysis to break down those with 100–400% into <100%, 100–<200%, 200–400%, poverty status was not strongly associated with both all-cause and cancer specific mortality. Cases who did not receive cancer-directed surgery had less favorable outcome compared to those who did, whereas cases without radiotherapy experienced similar mortality to those receiving this therapy. There were no significant interactions between race/ethnicity and socioeconomic factors (including health insurance) on the risk of mortality.
Table-6 presents the racial/ethnic disparities in all-cause and cancer-specific mortality by tumor sites. For breast cancer, the hazard ratio for all-cause mortality was significantly higher among blacks compared to whites in models adjusted for patient and tumor characteristics (Model-1) socioeconomic factors (education and income, poverty) and treatment (Model-2), and after additional adjustment for health insurance (Model-3). There was no significant difference in breast cancer mortality among other ethnic groups. Among cases with colorectal cancer, blacks were significantly more likely than whites to die of all causes and cancer-specific causes even after controlling for education, income, poverty and treatment (Model-1 and Model-2), but the association was no longer significant after adjusting for health insurance (Model-3). Among men with prostate cancer, hazard ratio of cancer-specific mortality for blacks compared to whites was significantly higher even after controlling for socioeconomic factors, treatment, and health insurance. Among cases with urinary bladder cancer, hazard ratios of both all-cause and cancer-specific mortality were nearly twice as high among blacks, but were not significantly different from whites among other racial and ethnic groups after controlling for socioeconomic factors and treatment. Among cases with lung and bronchus, cervical, ovarian cancer and melanoma, there were no significant differences in hazard ratios for blacks or any other racial and ethnic groups compared to whites after fully adjusting for socioeconomic factors and treatment.
Table 6.Mortality in association with race/ethnicity among patients with invasive tumors, by tumor site
Tumor site and Race/ethnicity Hazard ratio (95% CI) of All-Cause Mortality Hazard ratio (95% CI) of Cancer-Specific Mortality Model-1 Model-2 Model-3 Model-1 Model-2 Model-3 Breast NH-Black 1.4(1.2–1.8) 1.2(1.0–1.6) 1.7(1.1–2.5) 1.2(0.9–1.6) 1.1(0.8–1.5) 1.4(0.8–2.4) NH-API 0.6(0.4–0.8) 0.5(0.4–0.9) 0.6(0.3–1.1) 0.6(0.3–1.0) 0.6(0.3–1.1) 0.8(0.4–1.7) Hispanic 1.0(0.7–1.4) 0.9(0.7–1.2) 1.1(0.7–1.7) 0.8(0.5–1.3) 0.7(0.5–1.2) 0.8(0.4–1.5) AI/AN 0.1(0.0–30.2) 0.1(0.0–29.3.3) undefined 0.1(0.0 –45.2) 0.1(0.0–55.5) undefined Colorectal NH-Black 1.4(1.1–1.7) 1.2(1.0–1.5) 1.2(0.9–1.7) 1.4(1.1–1.8) 1.2(1.0–1.6) 1.2(0.8–1.8) NH-API 0.9(0.7–1.1) 0.9(0.7–1.1) 0.9(0.6–1.3) 0.8(0.6–1.2) 0.9(0.6–1.2) 0.9(0.6–1.5) Hispanic 1.1(0.8–1.4) 1.0(0.8–1.3) 1.0(0.7–1.5) 1.1(0.8–1.5) 1.0(0.7–1.3) 1.1(0.7–1.6) AI/AN 4.4(1.8–10.9) 4.8(1.9–11.9) 4.9(1.9–12.7) 6.1(2.5–15.1) 7.0(2.8–17.4) 6.6(2.6–17.2) Prostate NH-Black 1.2(1.1–1.5) 1.0(0.9–1.2) 1.1(0.9–1.4) 1.5(1.2–1.8) 1.3(1.0–1.6) 1.7(1.2–2.3) NH-API 0.8(0.6–1.1) 0.8(0.6–1.1) 0.7(0.5–1.2) 0.5(0.3–0.9) 0.5(0.3–0.9) 0.4(0.2–0.9) Hispanic 0.8(0.6–1.1) 0.7(0.5–0.9) 0.4(0.3–0.8) 0.9(0.6–1.4) 0.8(0.5–1.2) 0.4(0.2–0.9) AI/AN 2.0(0.7–5.3) 1.8(0.7–4.8) 2.2(0.7–6.9) 1.6(0.3–8.2) 1.5(0.3–7.4) 1.7(0.3–9.7) Lung and bronchus NH-Black 1.1(1.0–1.3) 1.0(0.9–1.2) 1.1(0.9–1.3) 1.1(1.0–1.3) 1.1(0.9–1.2) 1.1(0.9–1.3) NH-API 0.8(0.7–1.0) 0.8(0.7–1.1) 0.9(0.6–1.2) 0.8(0.6–1.0) 0.8(0.6–1.0) 0.8(0.5–1.1) Hispanic 1.0(0.8–1.2) 0.9(0.7–1.1) 0.8(0.6–1.1) 1.0(0.8–1.3) 0.9(0.7–1.2) 0.8(0.6–1.2) AI/AN 2.8(1.3–5.8) 2.1(1.0–4.4) 1.7(0.7–4.3) 3.1(1.4–6.7) 2.4(1.1–5.2) 2.0(0.8–5.2) Cervical NH-Black 0.6(0.2–1.4) 0.4(0.2–1.2) 0.1(0.0–0.5) 0.4(0.2–1.3) 0.4(0.1–1.4) 0.7(0.1–5.8) NH-API 1.6(0.6–4.3) 1.6(0.6–4.7) 2.3(0.3–19.7) 2.3(0.8–6.4) 2.1(0.7–6.8) 3.5(0.4–30.4) Hispanic 1.1(0.5–2.3) 1.0(0.4–2.3) 2.7(0.6–12.6) 1.3(0.6–3.0) 1.2(0.4–3.0) 1.4(0.2–17.1) AI/AN 2.2(0.2–21.3) 1.1(0.1–13.8) 5.6(0.0–>99) 2.4(0.1–69.0) 3.3(0.1–>99) 408.7(6.6–>999) Ovarian NH-Black 1.2(0.8–1.9) 1.2(0.8–1.9) 1.5(0.8–3.0) 1.1(0.7–1.9) 1.2(0.7–2.0) 1.3(0.6–2.6) NH-API 0.3(0.1–0.9) 0.3(0.1–1.0) 0.2(0.0–1.5) 0.4(0.1–1.1) 0.4(0.1–1.1) 0.3(0.0–1.7) Hispanic 0.9(0.5–1.7) 0.9(0.5–1.6) 1.3(0.6–3.2) 0.8(0.4–1.5) 0.7(0.3–1.4) 0.9(0.3–2.7) AI/AN 1.4(0.3–7.3) 2.2(0.4–12.5) undefined 1.6(0.3–8.1) 3.0(0.5–17.5) undefined Melanoma of the skin NH-Black 1.7(0.3–8.3) 1.6(0.3–8.2) undefined 3.6(0.7–18.3) 3.5(0.7–18.4) undefined NH-API 4.3(1.5–11.9) 4.7(1.6–13.4) 2.5(0.6–11.1) 4.9(1.5–15.6) 5.3(1.6–17.3) 0.8(0.1–5.3) Hispanic 3.4(1.2–9.9) 4.2(1.5–12.4) 1.5(0.3–7.7) 4.5(1.5–13.7) 6.5(2.2–19.6) 1.5 (0.3–9.2) AI/AN 5.8(0.6–59.0) 3.5(0.3–42.1) 6.8(0.4–103.4) undefined undefined undefined Urinary Bladder NH-Black 1.5(1.0–2.4) 1.3(0.8–2.0) 2.0(1.0–4.4) 1.8(1.0–3.1) 1.7(0.9–3.0) 2.3(0.9–6.0) NH-API 0.9(0.5–1.5) 0.9(0.5–1.6) 1.3(0.6–2.7) 0.9(0.4–2.1) 1.0(0.4–2.2) 1.2(0.4–3.9) Hispanic 0.7(0.3–1.5) 0.6(0.3–1.4) 0.6(0.2–1.8) 0.7(0.2–2.1) 0.7(0.2–2.1) 0.7(0.2–2.9) AI/AN 0.5(0.1–1.8) 0.7(0.2–2.5) 0.5(0.1–2.6) 0.3(0.0–3.0) 0.5(0.1–4.7) undefined DiscussionThis study examined the effects of socioeconomic factors at the individual level (health insurance, education, income and poverty status) and treatment on racial disparities in survival in large cohorts of cases that were diagnosed with 8 most common types of cancer at age ≥25 years. Noteworthy findings included that even after controlling for socioeconomic factors and patient and tumor characteristics, blacks were significantly less likely to receive cancer-directed surgery compared to whites, possibly due to a less favorable stage distribution at diagnosis. Hazard ratios for all-cause and cancer-specific mortality for 8 common tumors combined were no longer significantly higher among blacks after controlling for treatment and socioeconomic factors (education, income and poverty), or after further adjusting for health insurance. However, substantially higher hazard ratios persisted for all-cause mortality among black women with breast cancer and for cancer-specific mortality among black men with prostate cancer compared to whites. These associations were not observed among other racial/ethnic groups. Future studies should assess the role of treatment and local area effects as well as individual level factors in minority disparities in cancer survival.
The differences in survival between black and white cases have been attributed to numerous factors.7–20 Although racial/ethnic differences were likely multi-factorial, access to quality care and socioeconomic factors are prominent among these factors.10–24 Several studies demonstrated that if cases have equal access to quality health care, the outcomes are similar among different racial groups.10–13 Other studies showed the racial disparities still existed even after controlling for socioeconomic factors and for access to equitable care and treatment.10–13 These studies had variable quality data on SES, and the majority of studies relied on socioeconomic data at the area level (i.e., zip code or census tract) rather than at the individual level. Our study included several socioeconomic variables at the individual level, including health insurance, years of education, family income and poverty status. These unique data allow for a more complete adjustment for confounding by socioeconomic factors.
Differences in mortality between black and white cases were substantially reduced and were only marginally significant after adjusting for socioeconomic factors and treatment. This indicated that socioeconomic differences and treatment may play a major role in achieving equal outcomes for cases with cancer. These factors are generally modifiable. With improvement in these underlying factors, it may be more likely that national goals can be achieved25 such as Healthy People 2010 and 2020 objectives to eliminate racial disparities in health care and outcomes.
Several limitations should be noted. Socioeconomic status (health insurance, education income and poverty status) was obtained from participants at the time of survey rather than at time of cancer diagnosis. It was possible that socioeconomic status might have changed from the time of survey to the time of diagnosis, and therefore might not be a true socioeconomic status that mattered most at time of diagnosis, leading to certain degree of exposure misclassification. However, since the study population were selected for aged ≥25 and were later diagnosed with cancer, they were less likely to change SES substantially, particularly on education level. The models used in this analysis did not adjust for area-level factors such as economic status and environmental or physical conditions, which could therefore contribute to residual confounding, since several studies indicated that neighborhood socioeconomic status were independent predictors of health outcomes7,8,10–12 albeit often less influential than individual-level socioeconomic factors.7,8 In addition, although information on cancer-directed surgery and radiation from SEER data was available, the lack of data on chemotherapy and hormone therapy was a limitation, as was the absence of data on occupational exposures which may contribute to urinary bladder cancer.
In summary, survival time from diagnosis was significantly decreased among cancer cases with low compared to those with high socioeconomic status. Hazard ratios for all-cause and cancer-specific mortality among blacks compared to whites for 8 leading tumors combined lost statistical significance after adjusting for socioeconomic factors and treatment. Blacks had unfavorable prognoses compared to whites even after adjustment for socioeconomic status and treatment for several leading tumor sites such as breast, colorectal and urinary bladder cancer as did hazard ratios for disease-specific survival following diagnosis of colorectal and prostate cancer. Future population-based studies examining racial/ethnic disparities of cancer care and outcomes should include detailed measure of all treatment modalities rendered and account for both individual level and area socioeconomic factors.
AcknowledgmentsWe acknowledge efforts to the National Longitudinal Mortality Study (NLMS) by the U.S. Bureau of the Census, the National Cancer Institute, and SEER tumor registries to create this database. Data analysis was supported by Interagency Agreement Y1-PC-9021 between the National Cancer Institute and the Census Bureau. We thank Loreta Ajagba for her efforts to prepare the tables presented in this report.
ReferencesRetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.3