Abstract
Background
The end‐of‐life period is a crucial time in lung cancer care. To have a better understanding of the racial‐ethnic disparities in health care expenditures, access, and quality, we evaluated these disparities specifically in the end‐of‐life period for patients with lung cancer in the U.S.
Materials and Methods
We used the Surveillance, Epidemiology, and End Results (SEER)‐Medicare database to analyze characteristics of lung cancer care among those diagnosed between the years 2000 and 2011. Linear and logistic regression models were constructed to measure racial‐ethnic disparities in end‐of‐life care cost and utilization among non‐Hispanic (NH) Asian, NH black, Hispanic, and NH white patients while controlling for other risk factors such as age, sex, and SEER geographic region.
Results
Total costs and hospital utilization were, on average, greater among racial‐ethnic minorities compared with NH white patients in the last month of life. Among patients with NSCLC, the relative total costs were 1.27 (95% confidence interval [CI], 1.21–1.33) for NH black patients, 1.36 (95% CI, 1.25–1.49) for NH Asian patients, and 1.21 (95% CI, 1.07–1.38) for Hispanic patients. Additionally, the odds of being admitted to a hospital for NH black, NH Asian, and Hispanic patients were 1.22 (95% CI, 1.15–1.30), 1.47 (95% CI, 1.32–1.63), and 1.18 (95% CI, 1.01–1.38) times that of NH white patients, respectively. Similar results were found for patients with SCLC.
Conclusion
Minority patients with lung cancer have significantly higher end‐of‐life medical expenditures than NH white patients, which may be explained by a greater intensity of care in the end‐of‐life period.
Implications for Practice
This study investigated racial‐ethnic disparities in the cost and utilization of medical care among lung cancer patients during the end‐of‐life period. Compared with non‐Hispanic white patients, racial‐ethnic minority patients were more likely to receive intensive care in their final month of life and had statistically significantly higher end‐of‐life care costs. The findings of this study may lead to a better understanding of the racial‐ethnic disparities in end‐of‐life care, which can better inform future end‐of‐life interventions and help health care providers develop less intensive and more equitable care, such as culturally competent advanced care planning programs, for all patients.
Keywords: Racial‐ethnic disparity • Lung cancer • End‐of‐life care • Cost and utilization
Short abstract
This article evaluates racial‐ethnic disparities in cancer care for patients with lung cancer in the United States, focusing on cost and utilization of medical care in the end‐of‐life stage.
Introduction
Lung cancer accounts for more than 25% of total cancer deaths estimated for 2018 in the U.S., greatly exceeding cancers of any other site 1. Additionally, over 600 patients with lung cancer were estimated to be diagnosed per day in 2018, making lung cancer the second most commonly diagnosed cancer in the U.S., behind breast cancer. Both lung cancer incidence and mortality rates have been shown to be affected by race and ethnicity, with attributions for these disparities ranging broadly. Socioeconomic status, lack of access to diagnostic and treatment facilities, low health literacy, and unfamiliarity with providers, among numerous other factors, have been linked to differences in incidence and mortality 1, 2, 3, 4. Despite the laudable improvements in both early detection via lung cancer screening and treatment using advanced immunotherapy and targeted therapy, disparities in the appropriate use of these resources jeopardize their utility across racial and ethnic (racial‐ethnic) groups.
Several studies have documented the inequity between racial‐ethnic groups in the timeliness of lung cancer care and the use of effective treatment strategies 5, 6, 7, 8, 9. A particularly sensitive time in the course of care for patients with lung cancer is the end‐of‐life period, commonly considered the last month of a patient's life 10, 11. End‐of‐life care is typically high intensity, as the severity of lung cancer and related complications grow more serious and life‐threatening. Although the high‐risk nature of end‐of‐life care makes this period a crucial time for disease and symptom management, it also has significant consequences for the cost of care 12. Given the importance of end‐of‐life care to both the health and financial outcomes of lung cancer treatment, whether the aforementioned racial‐ethnic inequities in care persist in this period has been brought into question 13, 14, 15.
A recent study by Karanth and colleagues identified difference in end‐of‐life care costs and the most probable contributing factors 14, finding that non‐Hispanic (NH) black patients had higher end‐of‐life care costs compared with NH white patients. Although this study provided valuable insight into the broader determinants of end‐of‐life care disparities, an analysis of racial‐ethnic differences in the specific components of end‐of‐life care has yet to be performed. Studying the different pieces that make up end‐of‐life care costs and the difference in their utilization would provide additional understanding of how these racial‐ethnic disparities manifest in actual practice. For this study, we investigated potential disparities in medical care during the end‐of‐life period both generally—by examining the cost and utilization of hospital inpatient, hospital outpatient, hospice, and physician services—and more granularly—by delving into differences in costs and utilization of specific treatment strategies during inpatient stays, which account for most of the end‐of‐life care costs. We aim to determine how different types of end‐of‐life care are utilized by each racial‐ethnic group and identify specific areas for potential improvements in equitable care.
Materials and Methods
Data Source and Inclusion and Exclusion Criteria
We used the Surveillance, Epidemiology, and End Results (SEER)‐Medicare database for the years 2000–2013 for our analysis, which is composed of the SEER data set (including clinical, demographic, and cause of death information for persons with cancer from 18 U.S. cancer registries, representing about 34% of the population) linked to Medicare claims files (including inpatient, outpatient, and physician services claims) 16. In the U.S., 97% of persons aged 65 and over are eligible for Medicare coverage 17, which is composed of four parts. Part A coverage includes hospital, skilled nursing facility, hospice, and some home health care. Nearly all Medicare beneficiaries have Part A coverage. Part B coverage includes physician and outpatients services; 96% of Part A beneficiaries choose to enroll in Part B 18. Part C coverage (or Medicare Advantage) includes plans that are offered via private insurance companies but are approved by Medicare. Finally, Medicare Part D, which began in 2006, covers prescription drug expenses. Among persons aged 65 years and older in the SEER files, 93% were linked with the Medicare enrollment file 19.
We included patients aged 65 and older who were diagnosed with non‐small cell lung cancer (NSCLC) or small cell lung cancer (SCLC) between January 1, 2000, and December 31, 2011, in the analysis. Additionally, only patients who were continuously enrolled in both Part A and Part B coverage from 15 months before their cancer diagnosis to death or the end of 2013 were included to ensure that we captured complete claims data for health services and for the calculation of the Charlson comorbidity score. Because there are no detailed claims submitted from health maintenance organizations, individuals who were enrolled in managed care at any time during the study period were excluded. Individuals were also excluded if they had any postdeath costs, the date of claim was unknown, the month of cancer diagnosis was unknown, diagnosis was made at autopsy, or there were more than 3 months between the date of death recorded in the Medicare database and the date of death recorded in the SEER database. All people included in the analysis were deceased and survived at least 30 days after diagnosis to ensure that all outcomes and costs had the same potential to be observed. Detailed information about each inclusion and exclusion step is shown in a CONSORT diagram (supplemental online Fig. 1).
Measures
Total cost of care during the month preceding death for patients with lung cancer who died between 2000 and 2013 was calculated by combining Medicare payments, coinsurance payments, and deductibles and copayments billed to patients. We decomposed total cost into four mutually exclusive costs: (a) inpatient hospital or skilled nursing facility claims costs (referred to collectively as inpatient costs), (b) outpatient claims costs, (c) hospice claims costs, (d) noninstitutional claims costs for physician and supplier services (referred to collectively as physician services costs), and (e) other costs related to home health services and durable medical equipment. Other costs were not analyzed because their contribution to total cost was minimal. The costs from all inpatient hospital or skilled nursing facility claims refer to charge amounts for specific medical procedures or services provided during a beneficiary's inpatient stay. These include the beneficiary's liability and payment made on behalf of the beneficiary by Medicare and other payers. Outpatient, hospice, and physician services claims costs also include both Medicare payment and patient's liability. The patient's liability included any coinsurance payments, deductibles, and copayments that were billed to patients.
Costs were adjusted to constant 2018 U.S. dollars using the Centers for Medicare and Medicaid Services (CMS) Prospective Payment System Hospital Price Index for Part A claims and Medicare economic index for Part B claims. Part D costs were excluded from the analysis. All costs are presented in inflation‐adjusted 2018 U.S. dollars.
The utilization rates of inpatient care, outpatient care, hospice, and physician services were quantified using Medicare claims data. In most cases, a single record in the Medicare claims files reflects a summary of all care provided during a hospital or hospice stay or outpatient service. However, if the stay was long, there may be more than one claim per service. To avoid duplicates, we combined the claims from the same service for each patient. The utilization rates of specific medical procedures or services within inpatient care and the length of stay in intensive care unit (ICU) and coronary care unit (CCU) were also measured.
Patient race and ethnicity was coded as NH Asian, NH black, Hispanic, NH white, and other (includes Native Americans and those labeled as “other” or “unknown” in the SEER‐Medicare database). We considered several clinical and demographic characteristics that may be associated with end‐of‐life costs, including age and calendar year at death, sex, urban status (large metropolitan area and less urban or rural area), Medicaid (yes, no), marital status, American Joint Committee on Cancer (AJCC) cancer stages at diagnosis (I, II, III, IV), Charlson comorbidity score (0, 1, 2+) prior to cancer diagnosis, and SEER geographic region 14, 15. We extracted comorbidity information from Medicare inpatient, outpatient, and physician claims from 13 months before cancer diagnosis to the diagnosis date and then calculated the Deyo adaptation of the Charlson comorbidity index 20, 21, 22. We created a Medicaid variable to reflect socioeconomic status using the state buy‐in variables from SEER‐Medicare. We defined a patient as having participated in this program if they were enrolled for at least 1 month during the year prior to their cancer diagnosis. Geographic region was defined by aggregating the 18 SEER cancer registries Midwest, Northeast, South, and West according to the U.S. Census Bureau's definition of these four statistical regions.
Statistical Analysis
Linear Regression Analysis
Multivariable linear regression models were constructed to examine the association between log‐transformed total end‐of‐life cost and race and ethnicity, controlling for potential confounders mentioned above. Additional models were developed for component of inpatient cost. These included costs for ICU services, CCU services, pharmacy and drug therapies (pharmacy/drug), laboratory testing, emergency room services, blood administration, inhalation therapy services (e.g. oxygen tanks, inhalers, and respiratory therapists), magnetic resonance imaging, operating room services, physical therapy, occupational therapy, speech pathology, cardiology, anesthesia, and radiology. Multivariable linear regression models were fitted with charges for these specific cost types as the dependent variables and risk factors included in the linear regression for total inpatient cost as the independent variables. The cost estimates are calculated using linear regression with the age variable set to the median age at death (72 years), the year variable set to 2018, and all other variables fixed at their reference levels. In all linear regression models, NH white patients served as the reference category.
Logistic Regression Analysis
Logistic regression models were fit using whether inpatient claim, outpatient claim, hospice claim, or physician services claim incurred in the last month of a patient's life as the dependent variable and with the same set of independent variables as in the multivariable linear regression. Additional logistic regression models were developed using inpatient procedures (i.e., ICU admission, CCU admission, pharmacy/drug use, and inhalation therapy). The odds ratio estimates are calculated using logistic regression with the age variable set to the median age at death (72 years), the year variable set to 2018, and all other variables fixed at their reference levels.
The data analysis for this study was generated using SAS software.
Results
Sample Characteristics
Table 1 summarizes the demographic characteristics of our patient sample. There were 63,375 patients with NSCLC and 26,819 patients with SCLC. NH white patients made up the largest percentage (85.4% in NSCLC; 87.3% in SCLC), followed by NH black (9.0% in NSCLC; 7.9% in SCLC), NH Asian (2.8% in NSCLC; 2.2% in SCLC), and Hispanic (1.1% in both NSCLC and SCLC) patients. Patients in the Other category composed 1.7% and 1.5% of the NSCLC and SCLC group, respectively. Based on the results of the multivariable regression models, other risk factors—younger age, earlier calendar year, sex (being male), urban status (living in an urban area), participating in the state‐level Medicaid program, marital status (being married), AJCC cancer stages at diagnosis (later stages), comorbidity level (more comorbidities), and geographic region (being in the West or Northeast)—were all statistically significantly positively associated with total end‐of‐life care costs (supplemental online Table 1 and Table 2).
Table 1.
Summary statistics of sample characteristics by race and ethnicity
| Characteristics | NSCLC | SCLC | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| All | NH white | NH black | NH Asian | Hispanic | Other | All | NH white | NH black | NH Asian | Hispanic | Other | |
| Age at death (SD) | 77.5 (6.7) | 77.6 (6.7) | 76.4 (6.9) | 79.2 (6.4) | 78.4 (6.5) | 77.0 (7.0) | 76.0 (6.1) | 76.0 (6.1) | 75.0 (6.2) | 78.1 (6.1) | 77.3 (5.7) | 75.9 (6.3) |
| Total, n (%) | 65,375 (100.0) | 55,849 (85.4) | 5,891 (9.0) | 1,795 (2.8) | 732 (1.1) | 1,108 (1.7) | 26,819 (100.0) | 23,403 (87.3) | 2,130 (7.9) | 599 (2.2) | 292 (1.1) | 395 (1.5) |
| Sex, n (%) | ||||||||||||
| Male | 34,434 (52.7) | 29,221 (52.3) | 3,071 (52.1) | 1,053 (58.7) | 448 (61.2) | 641 (57.9) | 13,699 (51.1) | 11,860 (50.7) | 1,050 (49.3) | 380 (63.4) | 189 (64.7) | 220 (55.7) |
| Female | 30,941 (47.3) | 26,628 (47.7) | 2,820 (47.9) | 742 (41.3) | 284 (38.8) | 467 (42.2) | 13,120 (48.9) | 11,543 (49.3) | 1,080 (50.7) | 219 (36.6) | 103 (35.3) | 175 (44.3) |
| Year at death, n (%) | ||||||||||||
| 2000–2004 | 23,365 (35.7) | 19,932 (35.7) | 2,204 (37.4) | 567 (31.6) | 272 (37.2) | 390 (35.2) | 9,569 (35.7) | 8,405 (35.9) | 734 (34.5) | 186 (31.1) | 109 (37.3) | 135 (34.2) |
| 2004–2008 | 20,966 (32.1) | 17,967 (32.2) | 1,860 (31.6) | 588 (32.8) | 219 (29.9) | 332 (30.0) | 10,238 (38.2) | 8,896 (38.0) | 837 (39.3) | 240 (40.1) | 109 (37.3) | 156 (39.5) |
| 2008–2013 | 21,044 (32.2) | 17,950 (32.1) | 1,827 (31.0) | 640 (35.7) | 241 (32.9) | 386 (34.8) | 7,012 (26.1) | 6,102 (26.1) | 559 (26.2) | 173 (28.9) | 74 (25.3) | 104 (26.3) |
| Urban status, n (%) | ||||||||||||
| Large metropolitan | 33,531 (51.3) | 27,638 (49.5) | 3,504 (59.5) | 1,358 (75.7) | 422 (57.7) | 609 (55.0) | 13,098 (48.8) | 11,008 (47.0) | 1,305 (61.3) | 438 (73.1) | 162 (55.5) | 185 (46.8) |
| Less urban or rural | 31,844 (48.7) | 28,211 (50.5) | 2,387 (40.5) | 437 (24.4) | 310 (42.4) | 499 (45.0) | 13,721 (51.2) | 12,395 (53.0) | 825 (38.7) | 161 (26.9) | 130 (44.5) | 210 (53.2) |
| Medicaid, n (%) | ||||||||||||
| Yes | 12,661 (19.4) | 7,994 (14.3) | 2,496 (42.4) | 1,241 (69.1) | 527 (72.0) | 403 (36.4) | 4,823 (18.0) | 3,242 (13.9) | 879 (41.3) | 404 (67.5) | 199 (68.2) | 99 (25.1) |
| No | 52,714 (80.6) | 47,855 (85.7) | 3,395 (57.6) | 554 (30.9) | 205 (28.0) | 705 (63.6) | 21,996 (82.0) | 20,161 (86.2) | 1,251 (58.7) | 195 (32.6) | 93 (31.9) | 296 (74.9) |
| Marital status, n (%) | ||||||||||||
| Currently married | 31,253 (47.8) | 25,942 (46.5) | 3,802 (64.5) | 659 (36.7) | 378 (51.6) | 472 (42.6) | 12,478 (46.5) | 10,602 (45.3) | 1,356 (63.7) | 199a (>33.3) | 146a (>50.0) | 168 (42.5) |
| Not married | 2,493 (3.8) | 2,147 (3.8) | 268 (4.6) | 29 (1.6) | 24 (3.3) | 25 (2.3) | 893 (3.3) | 778 (3.3) | 87 (4.1) | <11a (<1.8) | <11a (<3.8) | 13 (3.3) |
| Unknown | 31,629 (48.4) | 27,760 (49.7) | 1,821 (30.9) | 1,107 (61.7) | 330 (45.1) | 611 (55.1) | 13,448 (50.1) | 12,023 (51.4) | 687 (32.3) | 389 (64.9) | 135 (46.2) | 214 (54.2) |
| Stage at diagnosis, n (%) | ||||||||||||
| I | 12,870 (19.7) | 11,340 (20.3) | 991 (16.8) | 254 (14.2) | 120 (16.4) | 165 (14.9) | 2,352 (8.8) | 2,058 (8.8) | 193 (9.1) | 49 (8.2) | 19a (>6.5) | 23a (>5.8) |
| II | 2,812 (4.3) | 2,473 (4.4) | 208 (3.5) | 67 (3.7) | 25 (3.4) | 39 (3.5) | 632 (2.4) | 578 (2.5) | 29 (1.4) | 13 (2.2) | <11a (<3.8) | <11a (<2.8) |
| III | 19,847 (30.4) | 16,817 (30.1) | 1,886 (32.0) | 578 (32.2) | 223 (30.5) | 343 (31.0) | 8,427 (31.4) | 7,305 (31.2) | 720 (33.8) | 183 (30.6) | 95 (32.5) | 124 (31.4) |
| IV | 29,846 (45.7) | 25,219 (45.2) | 2,806 (47.6) | 896 (49.9) | 364 (49.7) | 561 (50.6) | 15,408 (57.5) | 13,462 (57.5) | 1,188 (55.8) | 354 (59.1) | 167 (57.2) | 237 (60.0) |
| Charlson score, n (%) | ||||||||||||
| 0 | 28,673 (43.9) | 24,740 (44.3) | 2,261 (38.4) | 869 (48.4) | 275 (37.6) | 528 (47.7) | 11,746 (43.8) | 10,322 (44.1) | 838 (39.3) | 297 (49.6) | 108 (37.0) | 181 (45.8) |
| 1 | 17,834 (27.3) | 15,329 (27.5) | 1,522 (25.8) | 500 (27.9) | 184 (25.1) | 299 (27.0) | 7,642 (28.5) | 6,709 (28.7) | 583 (27.4) | 160 (26.7) | 85 (29.1) | 105 (26.6) |
| 2+ | 18,868 (28.9) | 15,780 (28.3) | 2,108 (35.8) | 426 (23.7) | 273 (37.3) | 281 (25.4) | 7,431 (27.7) | 6,372 (27.2) | 709 (33.3) | 142 (23.7) | 99 (33.9) | 109 (27.6) |
| Region, n (%) | ||||||||||||
| Northeast | 13,012 (19.9) | 11,767 (21.1) | 982 (16.7) | 94 (5.2) | 106 (14.5) | 63 (5.7) | 5,338 (19.9) | 4,851 (20.7) | 373 (17.5) | 27 (4.5) | 54 (18.5) | 33 (8.4) |
| South | 20,083 (30.7) | 17,272 (30.9) | 2,695 (45.8) | 40 (2.2) | 15 (2.1) | 61 (5.5) | 8,570 (32.0) | 7,560 (32.3) | 955 (44.8) | 16 (2.7) | <11a (<3.8) | 31 (7.9) |
| Midwest | 9,642 (14.7) | 8,410 (15.1) | 1,145 (19.4) | 22 (1.2) | 12 (1.6) | 53 (4.8) | 3,714 (13.8) | 3,285 (14.0) | 389 (18.3) | 11 (1.8) | <11a (<3.8) | 21 (5.3) |
| West | 22,638 (34.6) | 18,400 (33.0) | 1,069 (18.2) | 1,639 (91.3) | 599 (81.8) | 931 (84.0) | 9,197 (34.3) | 7,707 (32.9) | 413 (19.4) | 545 (91.0) | 222 (76.0) | 310 (78.5) |
Data masked to comply with SEER‐Medicare policy for groups with less than 11 patients.
Abbreviations: NH, non‐Hispanic; NSCLC, non‐small cell lung cancer; SCLC, small cell lung cancer.
Table 2.
Association between patients’ last month costs and race and ethnicity groups from linear regression
| Race‐ethnicity | NSCLC | SCLC | ||||
|---|---|---|---|---|---|---|
| n | Relative cost (95% CI) | p value | n | Relative cost (95% CI) | p value | |
| Total | ||||||
| NH white | 55,849 | 1 | 23,403 | 1 | ||
| NH black | 5,891 | 1.27 (1.21–1.33) | <.001 a | 2,130 | 1.17 (1.09–1.26) | <.001 a |
| NH Asian | 1,795 | 1.36 (1.25–1.49) | <.001 a | 599 | 1.26 (1.09–1.45) | .001 a |
| Hispanic | 732 | 1.21 (1.07–1.38) | .003 a | 292 | 1.24 (1.03–1.51) | .027 a |
| Others | 1,108 | 1.23 (1.11–1.37) | <.001 a | 395 | 1.18 (1.00–1.39) | .052 |
| Hospital inpatient | ||||||
| NH white | 34,079 | 1 | 14,741 | 1 | ||
| NH black | 3,933 | 1.14 (1.10–1.18) | <.001 a | 1,429 | 1.14 (1.08–1.21) | <.001 a |
| NH Asian | 1,248 | 1.22 (1.14–1.30) | <.001 a | 422 | 1.12 (1.01–1.25) | .031 a |
| Hispanic | 473 | 1.05 (0.96–1.16) | .281 | 203 | 0.98 (0.85–1.13) | .784 |
| Others | 723 | 1.07 (0.99–1.16) | .086 | 260 | 1.03 (0.91–1.17) | .620 |
| Hospital outpatient | ||||||
| NH white | 27,549 | 1 | 12,096 | 1 | ||
| NH black | 2,794 | 1.16 (1.09–1.24) | <.001 a | 1,068 | 1.07 (0.97–1.18) | .182 |
| NH Asian | 745 | 1.04 (0.93–1.17) | .475 | 296 | 1.10 (0.91–1.32) | .322 |
| Hispanic | 315 | 0.93 (0.78–1.10) | .382 | 145 | 1.17 (0.91–1.50) | .234 |
| Others | 540 | 1.04 (0.91–1.19) | .522 | 197 | 1.13 (0.91–1.41) | .255 |
| Hospice | ||||||
| NH white | 25,729 | 1 | 11,324 | 1 | ||
| NH black | 2,393 | 1.04 (1.00–1.09) | .043 a | 936 | 1.08 (1.02–1.16) | .014 a |
| NH Asian | 560 | 1.05 (0.97–1.14) | .239 | 194 | 0.87 (0.76–1.00) | .044 a |
| Hispanic | 295 | 1.03 (0.92–1.15) | .618 | 118 | 1.26 (1.06–1.50) | .009 a |
| Others | 412 | 1.06 (0.97–1.16) | .221 | 157 | 1.09 (0.94–1.27) | .252 |
| Physician services | ||||||
| NH white | 51,653 | 1 | 21,831 | 1 | ||
| NH black | 5,551 | 1.12 (1.08–1.16) | <.001 a | 2,007 | 1.04 (0.98–1.10) | .184 |
| NH Asian | 1,666 | 1.22 (1.14–1.30) | <.001 a | 563 | 1.14 (1.03–1.27) | .013 a |
| Hispanic | 680 | 1.12 (1.02–1.23) | .022 a | 276 | 1.11 (0.96–1.29) | .165 |
| Others | 1,035 | 1.09 (1.01–1.18) | .029 a | 363 | 1.13 (1.00–1.29) | .055 |
Indicates significant difference when compared with NH white patients at p < .05.
Bold text indicates statistically significant with a p < .05.
Abbreviations: CI, confidence interval; NH, non‐Hispanic; NSCLC, non‐small cell lung cancer; SCLC, small cell lung cancer.
Racial‐Ethnic Disparities in End‐of‐Life Care Costs
Disparities were detected in an analysis of total cost, as well as each component of total cost (hospital inpatient, hospital outpatient, hospice, and physician services costs; Table 2). Total end‐of‐life costs were significantly higher for all racial‐ethnic groups relative to NH white patients, whose average total costs and patient liability costs in the last month of life were estimated to be $9,113 (95% confidence interval [CI], $8,547–$9,717) and $797 (95% CI, $751–$845), respectively, among patients with NSCLC (supplemental online Table 3). Among patients with NSCLC, the relative costs were 1.27 (95% CI, 1.21–1.33) for NH black patients, 1.36 (95% CI, 1.25–1.49) for NH Asian patients, and 1.21 (95% CI, 1.07–1.38) for Hispanic patients. Similar results were found for patients with SCLC. Inpatient care costs, the largest component of total end‐of‐life care costs (Fig. 1), were significantly higher for both NH black (1.14; 95% CI, 1.10–1.18) and NH Asian (1.22; 95% CI, 1.14–1.30) patients with NSCLC, as well as for NH black (1.14; 95% CI, 1.08–1.21) and NH Asian (1.12; 95% CI, 1.01–1.25) patients with SCLC. Outpatient care costs, hospice care costs, and physician services costs were also significantly higher for some racial‐ethnic minority groups (Table 2). These disparities persisted whether or not patients enrolled in hospice (supplemental online Table 4).
Figure 1.
Total end‐of‐life care costs breakdown. This plot shows the proportion of each claim cost to the total cost in patients’ last month of life among patients with non‐small cell lung cancer (NSCLC) and patients with small cell lung cancer (SCLC).
Racial‐Ethnic Disparity in End‐of‐Life Care Utilization
As shown in Table 3, among patients NSCLC, the hospital admission rates during the last month of life were significantly higher for all minority groups compared with NH white patients. The odds of being admitted to a hospital for NH black, NH Asian, and Hispanic patients were 1.22 (95% CI, 1.15–1.30), 1.47 (95% CI, 1.32–1.63), and 1.18 (95% CI, 1.01–1.38) times that of NH white patients, respectively. In terms of outpatient and hospice services, minority groups had lower utilization rates than NH white patients. The odds of receiving outpatient services for NH Asian patients were 0.85 (95% CI, 0.77–0.94) times that of NH white patients. The odds of being enrolled in hospice for NH black and NH Asian patients were statistically significantly lower than that of NH white patients, with odds ratios of 0.81 (95% CI, 0.76–0.86) and 0.62 (95% CI, 0.55–0.69), respectively. The odds of receiving physician services among these patients with NSCLC were only statistically significantly higher for NH black patients compared with NH white patients, with an odds ratio of 1.24 (95% CI, 1.09–1.40). Among patients with SCLC, NH Asian patients also had a significantly higher hospital admission rate during the last month of life compared with NH white patients. NH black and NH Asian patients with SCLC also had a significantly lower hospice enrollment rate than NH white patients with SCLC (Table 3).
Table 3.
Association between patients’ last month treatment utilization rate and race and ethnicity groups from logistic regression
| Race‐ethnicity | NSCLC | SCLC | ||||
|---|---|---|---|---|---|---|
| Estimated percentage | Estimated odds ratio (95% CI) | p value | Estimated percentage | Estimated odds ratio (95% CI) | p value | |
| Inpatient | ||||||
| NH white | 60.2 | 1 | 60.5 | 1 | ||
| NH black | 64.9 | 1.22 (1.15–1.30) | <.001 a | 63.6 | 1.14 (1.04–1.26) | .008 a |
| NH Asian | 68.9 | 1.47 (1.32–1.63) | <.001 a | 67.6 | 1.36 (1.13–1.64) | .001 a |
| Hispanic | 64.1 | 1.18 (1.01–1.38) | .040 a | 66.1 | 1.27 (0.99–1.65) | .063 |
| Others | 64.6 | 1.21 (1.07–1.37) | .003 a | 63.3 | 1.13 (0.91–1.39) | .268 |
| Outpatient | ||||||
| NH white | 63.7 | 1 | 65.7 | 1 | ||
| NH black | 62.9 | 0.97 (0.91–1.02) | .245 | 65.9 | 1.01 (0.92–1.10) | .888 |
| NH Asian | 59.9 | 0.85 (0.77–0.94) | .002 a | 67.3 | 1.07 (0.90–1.27) | .433 |
| Hispanic | 60.1 | 0.86 (0.74–1.00) | .051 | 66.2 | 1.02 (0.81–1.29) | .866 |
| Others | 63.8 | 1.01 (0.89–1.14) | .912 | 64.7 | 0.96 (0.78–1.17) | .667 |
| Hospice | ||||||
| NH white | 58.6 | 1 | 58.4 | 1 | ||
| NH black | 53.4 | 0.81 (0.76–0.86) | <.001 a | 53.9 | 0.83 (0.76–0.92) | <.001 a |
| NH Asian | 46.6 | 0.62 (0.55–0.69) | <.001 a | 46.2 | 0.61 (0.51–0.73) | <.001 a |
| Hispanic | 58.6 | 1.00 (0.86–1.16) | .987 | 56.7 | 0.93 (0.73–1.19) | .558 |
| Others | 52.5 | 0.78 (0.69–0.89) | <.001 a | 53.3 | 0.81 (0.66–1.00) | .049 a |
| Physician services | ||||||
| NH white | 91.0 | 1 | 91.6 | 1 | ||
| NH black | 92.6 | 1.24 (1.09–1.40) | .001 a | 92.6 | 1.15 (0.93–1.41) | .198 |
| NH Asian | 91.3 | 1.04 (0.85–1.28) | .704 | 93.1 | 1.25 (0.85–1.83) | .262 |
| Hispanic | 90.7 | 0.97 (0.71–1.30) | .815 | 93.0 | 1.23 (0.71–2.12) | .467 |
| Others | 92.3 | 1.20 (0.92–1.55) | .176 | 89.9 | 0.83 (0.56–1.21) | .323 |
Indicates significant difference when compared with NH white patients at p < .05.
Bold text indicates statistically significant with a p < .05.
The percentage estimates are calculated using logistic regression with the age variable set to the median age at death (72 years), the year variable set to 2018, and all other variables fixed at their reference levels.
Abbreviations: CI, confidence interval; NH, non‐Hispanic; NSCLC, non‐small cell lung cancer; SCLC, small cell lung cancer.
Specific Hospital Inpatient Costs
Given that inpatient care costs made up 65.5% and 65.1% of the total cost of end‐of‐life care among patients with NSCLC and patients with SCLC (Fig. 1), respectively, we further analyzed racial‐ethnic disparities within subcategories of inpatient costs. Several specific inpatient cost estimates for the last month of life were statistically significantly greater for some minority groups compared with NH white patients. Supplemental online Table 4 provides cost estimates and significance measures for each end‐of‐life inpatient cost studied for each race group.
ICU costs made up the largest portion of total inpatient cost estimates. Differences compared with NH white patients were statistically significant for NH black and NH Asian patients with NSCLC and for NH black patients with SCLC (Fig. 2). Among patients with NSCLC, ICU costs for NH black and NH Asian patients were estimated to be $28,292 (95% CI, $25,534–$31,347) and $28,099 (95% CI, $24,839–$31,787), respectively, compared with $24,345 (95% CI, $22,392–$26,469) for NH white patients (supplemental online Table 5). CCU costs were the second largest component of total inpatient cost estimates, and, again, NH black patients with NSCLC, as well as NH black patients with SCLC, had statistically significantly greater cost estimates than NH white patients (Fig. 2). These specific estimates can be found in supplemental online Table 5.
Figure 2.
Specific hospital inpatient cost estimates by racial‐ethnic group. Among all specific hospital inpatient claims, intensive care unit, coronary care unit, laboratory, pharmacy and drugs, and inhalation therapy are the five highest cost services. (A) and (B) show the costs of specific hospital inpatient services among patients with non‐small cell lung cancer (NSCLC) and small cell lung cancer (SCLC), respectively. The cost estimates are calculated using linear regression with the age variable set to the median age at death (72 years), the year variable set to 2018, and all other variables fixed at their reference levels. Asterisk (*) indicates that the cost difference between the racial‐ethnic group and white group is statistically significant. The white lines across bar charts represent the cost of non‐Hispanic (NH) white patients for each specific hospital inpatient service.
Costs for laboratory tests were statistically significantly greater for all racial‐ethnic minority patient groups, as compared with NH white patients, among both patients with NSCLC and those with SCLC, except for Hispanic patients with SCLC (Fig. 2). Among patients with NSCLC, laboratory costs were estimated to be $14,357 (95% CI, $13,415–$15,364) for NH black patients, $13,983 (95% CI, $12,831–$15,239) for NH Asian patients, and $14,711 (95% CI, $13,059–$16,571) for Hispanic patients compared with $12,261 (95% CI, $11,608–$12,950) for NH white patients. Similarly, differences in cost estimates for pharmacy/drug were statistically significant for NH black and NH Asian patients as compared with NH white patients among patients NSCLC, although these differences were not significant for patients with SCLC (Fig. 2). Laboratory and pharmacy/drug cost estimates can be found in supplemental online Table 5.
Finally, although inhalation therapy cost estimates were relatively low in magnitude, statistically significant differences were observed in NH Asian patients with NSCLC and in NH black patients with SCLC (Fig. 2). For NH Asian patients with NSCLC, inhalation therapy cost estimates were $7,119 (95% CI, $6,238–$8,125) compared with $5,772 (95% CI, $5,307–$6,279) for NH white patients. Minority patients were found to have statistically significantly greater cost estimates for several other specific hospital inpatient costs (supplemental online Table 5), although these were less substantial than the aforementioned inpatient cost types.
Racial‐Ethnic Disparity in End‐of‐Life Inpatient Care Utilization
We continued our investigation of end‐of‐life inpatient care by examining racial‐ethnic disparities in the utilization rates of specific inpatient services. Table 4 shows estimated odds ratios for being admitted to the ICU in the last month of life. Substantial differences in ICU admission rates by race and ethnicity groups were observed among patients with lung cancer. Among patients with NSCLC, the odds ratios of being admitted to the ICU for NH black, NH Asian, and Hispanic patients, when compared with NH whites, were 1.09 (95% CI, 1.02–1.18), 1.30 (95% CI, 1.15–1.47), and 1.42 (95% CI, 1.17–1.73), respectively. Furthermore, the length of stay in the ICU per patient with NSCLC was significantly longer for NH Asian patients (5.50 days; 95% CI, 5.00–6.01), when compared with NH white patients (5.18 days; 95% CI, 4.77–5.60; supplemental online Table 6).
Table 4.
Association between patients’ last month inpatient care utilization rate and race and ethnicity groups from logistic regression
| Race‐ethnicity | NSCLC | SCLC | ||||
|---|---|---|---|---|---|---|
| Estimated percentage | Estimated odds ratio (95% CI) | p value | Estimated percentage | Estimated odds ratio (95% CI) | p value | |
| ICU | ||||||
| NH white | 54.5 | 1 | 53.8 | 1 | ||
| NH black | 56.7 | 1.09 (1.02–1.18) | <.001 a | 56.8 | 1.13 (1.00–1.27) | .046 a |
| NH Asian | 60.8 | 1.30 (1.15–1.47) | <.001 a | 54.9 | 1.04 (0.84–1.29) | .703 |
| Hispanic | 63.0 | 1.42 (1.17–1.73) | <.001 a | 58.8 | 1.23 (0.91–1.65) | .175 |
| Others | 56.3 | 1.08 (0.92–1.26) | .367 | 57.1 | 1.14 (0.88–1.48) | .326 |
| CCU | ||||||
| NH white | 18.4 | 1 | 19.5 | 1 | ||
| NH black | 20.1 | 1.11 (1.00–1.23) | .041 a | 21.4 | 1.13 (0.95–1.33) | .169 |
| NH Asian | 18.6 | 1.01 (0.85–1.20) | .883 | 16.8 | 0.83 (0.61–1.15) | .260 |
| Hispanic | 19.3 | 1.06 (0.82–1.38) | .660 | 24.3 | 1.33 (0.91–1.95) | .147 |
| Others | 18.3 | 0.99 (0.80–1.23) | .934 | 19.2 | 0.98 (0.68–1.42) | .930 |
| Lab | ||||||
| NH white | 98.7 | 1 | 98.0 | 1 | ||
| NH black | 98.9 | 1.21 (0.97–1.51) | .090 | 98.0 | 0.99 (0.68–1.45) | .969 |
| NH Asian | 99.1 | 1.52 (1.03–2.26) | .037 a | 97.0 | 0.66 (0.39–1.11) | .116 |
| Hispanic | 98.7 | 1.00 (0.62–1.61) | .997 | 99.1 | 2.29 (0.72–7.27) | .162 |
| Others | 98.9 | 1.23 (0.77–1.97) | .379 | 98.2 | 1.08 (0.50–2.31) | .853 |
| Pharmacy and drug | ||||||
| NH white | 99.9 | 1 | 99.5 | 1 | ||
| NH black | 99.8 | 0.84 (0.53–1.34) | .462 | 99.1 | 0.57 (0.27–1.21) | .145 |
| NH Asian | 99.9 | 0.90 (0.45–1.83) | .774 | 98.5 | 0.35 (0.13–0.92) | .033 a |
| Hispanic | 99.8 | 0.64 (0.27–1.51) | .310 | 100.0 | >999.99 (<0.01 to >999.99) | .983 |
| Others | 99.8 | 0.58 (0.26–1.27) | .170 | 99.5 | 1.19 (0.16–8.81) | .865 |
| Inhalation therapy | ||||||
| NH white | 79.3 | 1 | 79.6 | 1 | ||
| NH black | 77.1 | 0.88 (0.81–0.96) | .004 a | 79.2 | 0.98 (0.85–1.13) | .761 |
| NH Asian | 79.8 | 1.03 (0.89–1.20) | .668 | 78.2 | 0.92 (0.72–1.19) | .531 |
| Hispanic | 78.6 | 0.96 (0.76–1.20) | .708 | 82.8 | 1.23 (0.85–1.80) | .274 |
| Others | 80.6 | 1.09 (0.90–1.31) | .398 | 76.3 | 0.83 (0.62–1.11) | .201 |
Indicates significant difference when compared with NH white patients at p < .05.
Bold text indicates statistically significant with a p < .05.
The percentage estimates are calculated using logistic regression with the age variable set to the median age at death (72 years), the year variable set to 2018, and all other variables fixed at their reference levels.
Abbreviations: CI, confidence interval; CCU, coronary care unit; ICU, intensive care unit; NH, non‐Hispanic; NSCLC, non‐small cell lung cancer; SCLC, small cell lung cancer.
We also examined the association of CCU admission rates and the use of laboratory tests, pharmacy/drug services, and inhalation therapy with racial‐ethnic status (Table 4). There were few significant differences in the utilization of these services between patient groups. NH Asian patients with NSCLC received more lab tests than NH white patients with NSCLC (odds ratio [OR], 1.52; 95% CI, 1.03–2.26). NH black patients with NSCLC received less inhalation therapy compared with NH white patients with NSCLC (OR, 0.88; 95% CI, 0.81–0.96). In addition, NH Asian patients with SCLC tended to receive less pharmacy/drug services than NH white patients with SCLC (OR, 0.35; 95% CI, 0.13–0.92). Although NH Asian patients with SCLC did receive statistically significantly fewer pharmacy/drug services according to the estimated odds ratio, close to 100% of patients in all racial‐ethnic patient groups received pharmacy/drug services.
Discussion
Using the SEER‐Medicare database, we studied the cost and utilization of care for patients with lung cancer during the final month of life. We found that the total cost of end‐of‐life care differed significantly between each racial‐ethnic minority patient group. NH black and NH Asian patients had significantly higher inpatient end‐of‐life care costs when compared with NH white patients. All minority patient groups were significantly more likely to be admitted to inpatient care. The greatest differences in inpatient cost estimates for certain minority patient groups were found in those covering ICU care, CCU care, laboratory services, pharmacy/drugs, and inhalation therapy. Additionally, NH black, NH Asian, and Hispanic patients were more likely to be admitted to the ICU than were NH white patients in the final month of life, potentially contributing to the higher end‐of‐life inpatient cost, because ICU utilization is the largest of the specific cost types that make up overall hospital inpatient costs. Significant differences in the cost and utilization of end‐of‐life care were also found for outpatient, hospice, and physician services, although these cost estimates were decidedly lower than inpatient cost estimates and, thus, contributed much less to the overall end‐of‐life cost disparities.
Implications of Specific Care Types
ICU and CCU care costs were estimated to be higher for NH black and NH Asian patients than for NH white patients. These differences, along with higher ICU admission rates and greater lengths of stay, suggest that these patients may have received a greater intensity of care in their final month of life. Additionally, although nearly the same proportion of racial‐ethnic minority patients and NH white patients received laboratory and pharmacy/drug services, patients in all three racial‐ethnic minority groups accumulated higher estimated costs for laboratory services and pharmacy/drugs. These results may indicate that among patients who received drug therapies and laboratory tests, minority patients received a larger number or more expensive drug therapies and had more laboratory tests run during their final month of life. The differences for each of these specific inpatient cost estimates were substantial, generally approaching or exceeding $2,000 per patient. These results identify specific points of care in which increased attention could be focused on developing solutions to reduce such wide cost disparities. Future research should investigate why disparities manifest so greatly in these contexts.
Moreover, these points of care, being indicators of greater intensity of care, are made more critical in that they likely affect quality of life during the patient's final month. A recent study showed that increasing hospice care was associated with less aggressive medical treatment and lower costs 23. Our results provide evidence that racial‐ethnic minority patients enrolled in hospice much less frequently than NH white patients. Indeed, it is known that NH white patients are more likely to receive less aggressive end‐of‐life care outside of the hospital 24. This pattern of care not only suggests inequities in cost burden between racial‐ethnic groups but also raises the concern that racial‐ethnic minority patients may experience a lower quality of life by receiving less palliation before death 25, 26, underscoring the need for attending to these subcategories of inpatient care.
Other Contributing Factors
As was previously mentioned, other studies in the literature have looked to identify overarching factors that could lead to the greater costs and intensity of care seen in our results, including geographic differences, comorbidity level, ineffective communication with providers, and culturally‐driven preferences for more aggressive treatment 14, 24, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36. Although geographic differences may contribute to disparities in end‐of‐life care cost 14, these racial‐ethnic disparities persisted even when controlling for a patient's region. Our results supported the notion that greater comorbidity levels and poor management of comorbidities among racial‐ethnic minorities could lead to higher expenditures 14. We observed an association between comorbidity level and total end‐of‐life cost, as well as a significant amount of spending associated with the CCU and inhalation therapy. Comprehensive care that considers comorbidities in addition to lung cancer could potentially reduce these hospitalizations and related costs. Additionally, insufficient communication could cause providers to not fully understand patients’ needs and order unnecessary treatments or tests 31, 32, 33, 34, which may help explain the results of our analysis of pharmacy/drug and laboratory costs. Regarding the inequality of treatment among minority groups, Belisomo (2018) pointed out that structural racism existed in end‐of‐life care and that providing advance care planning that is culturally attentive would help to alleviate this issue 37.
Limitations
Our study is subject to several limitations inherent in claims data analysis. First, our study population is limited to patients over age 65 who are Medicare beneficiaries and patients from SEER cancer registry regions, meaning that our results may not be generalizable to younger populations. Our study population is also limited to patients who had continuous Medicare enrollment during the whole study period, which could potentially bias the analysis by excluding patients who had noncontinuous Medicare enrollment from the year of diagnosis to one month prior to their date of death. Second, the costs in our analysis include both Medicare payments and the patient's liability. We are not able to determine the patient's out‐of‐pocket expense, which could affect a patient's choice of end‐of‐life care. Finally, the SEER‐Medicare data used only includes claims up to 2013, before the introduction of immunotherapy treatments for stage IV lung cancer in 2015 38. Immunotherapy treatments are expensive and would likely increase the overall cost of end‐of‐life care. The utilization and cost differences of immunotherapy treatment by racial‐ethnic group would be an appropriate focus for future research.
Conclusion
Racial‐ethnic minority patients with lung cancer experience significantly higher end‐of‐life care costs and more intense care compared with NH white patients, with the greatest differences apparent in costs and utilization related to inpatient care services. Although several broader factors are likely to contribute to the disparity in end‐of‐life care cost and utilization, it is worthwhile to focus in on what types of services, specifically, reveal the greatest cost differences. In this way, care providers can work to better understand this issue at the hospital level and develop more actionable solutions to provide equitable, culturally competent care.
Author Contributions
Conception/design: Yufan Chen, Chung Yin Kong
Provision of study material or patients: Andrew Eckel, Angela C. Tramontano, Chung Yin Kong
Collection and/or assembly of data: Yufan Chen, Andrew Eckel, Angela C. Tramontano
Data analysis and interpretation: Yufan Chen, Steven D. Criss, Nathaniel D. Mercaldo, Chung Yin Kong
Manuscript writing: Yufan Chen, Steven D. Criss, Tina R. Watson, Andrew Eckel, Lauren Palazzo, Angela C. Tramontano, Ying Wang, Nathaniel D. Mercaldo, Chung Yin Kong
Final approval of manuscript: Yufan Chen, Steven D. Criss, Tina R. Watson, Andrew Eckel, Lauren Palazzo, Angela C. Tramontano, Ying Wang, Nathaniel D. Mercaldo, Chung Yin Kong
Disclosures
The authors indicated no financial relationships.
Supporting information
See http://www.TheOncologist.com for supplemental material available online.
Supplemental Figure
Supplemental Tables
Acknowledgments
This study used the linked SEER‐Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. This work was supported by grants from the National Institutes of Health (U01CA199284 R01CA202956). The authors acknowledge the efforts of the National Cancer Institute; the Office of Research, Development and Information, CMS; Information Management Services, Inc.; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER‐Medicare database.
This work was supported by the National Institutes of Health (U01CA199284 R01CA202956).
Disclosures of potential conflicts of interest may be found at the end of this article.
References
- 1. Howlader N, Noone AM, Krapcho M et al. Seer cancer statistics review, 1975‐2016. Bethesda, MD: National Cancer Institute, 2019. [Google Scholar]
- 2. Albano JD, Ward E, Jemal A et al. Cancer mortality in the United States by education level and race. J Natl Cancer Inst 2007;99:1384–1394. [DOI] [PubMed] [Google Scholar]
- 3. Rubin MS, Clouston S, Link BG. A fundamental cause approach to the study of disparities in lung cancer and pancreatic cancer mortality in the United States. Soc Sci Med 2014;100:54–61. [DOI] [PubMed] [Google Scholar]
- 4. Coughlin SS, Matthews‐Juarez P, Juarez PD et al. Opportunities to address lung cancer disparities among African Americans. Cancer Med 2014;3:1467–1476. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Shugarman LR, Mack K, Sorbero ME et al. Race and sex differences in the receipt of timely and appropriate lung cancer treatment. Med Care 2009;47:774–781. [DOI] [PubMed] [Google Scholar]
- 6. Hardy D, Liu CC, Xia R et al. Racial disparities and treatment trends in a large cohort of elderly black and white patients with nonsmall cell lung cancer. Cancer 2009;115:2199–2211. [DOI] [PubMed] [Google Scholar]
- 7. Nadpara PA, Madhavan SS, Tworek C. Disparities in lung cancer care and outcomes among elderly in a medically underserved state population‐a cancer registry‐linked database study. Popul Health Manag 2016;19:109–119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Farjah F, Wood DE, Yanez ND, 3rd et al. Racial disparities among patients with lung cancer who were recommended operative therapy. Arch Surg 2009;144:14–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Gross CP, Smith BD, Wolf E, et al. Racial disparities in cancer therapy: Did the gap narrow between 1992 and 2002? Cancer 2008;112:900–908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Saito AM, Landrum MB, Neville BA et al. Hospice care and survival among elderly patients with lung cancer. J Palliat Med 2011;14:929–939. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Earle CC, Neville BA, Landrum MB et al. Trends in the aggressiveness of cancer care near the end of life. J Clin Oncol 2004;22:315–321. [DOI] [PubMed] [Google Scholar]
- 12. Brown ML, Riley GF, Schussler N et al. Estimating health care costs related to cancer treatment from SEER‐Medicare data. Med Care 2002;40(suppl):IV104–IV117. [DOI] [PubMed] [Google Scholar]
- 13. Institute of Medicine . Dying in America: Improving quality and honoring individual preferences near the end of life. Washington, DC: The National Academies Press, 2015. [PubMed] [Google Scholar]
- 14. Karanth S, Rajan SS, Revere FL et al. Factors affecting racial disparities in end‐of‐life care costs among lung cancer patients: A SEER‐Medicare‐based study. Am J Clin Oncol 2019;42:143–153. [DOI] [PubMed] [Google Scholar]
- 15. Karanth S, Rajan SS, Sharma G et al. Racial‐ethnic disparities in end‐of‐life care quality among lung cancer patients: A SEER‐Medicare‐based study. J Thorac Oncol 2018;13:1083–1093. [DOI] [PubMed] [Google Scholar]
- 16. National Cancer Institute . Overview of the SEER program. 2017. Available at https://seer.cancer.gov/about/overview.html. Accessed January 4, 2019.
- 17. De Lew N. Medicare: 35 years of service. Health Care Financ Rev 2000;22:75–103. [PMC free article] [PubMed] [Google Scholar]
- 18. National Cancer Institute . SEER‐Medicare: Medicare enrollment & claims data. 2018. Available at https://healthcaredelivery.cancer.gov/seermedicare/medicare/. Accessed January 7, 2019.
- 19. National Cancer Institute . SEER‐Medicare: How the SEER & Medicare data are linked. 2017. Available at https://healthcaredelivery.cancer.gov/seermedicare/overview/linked.html. Accessed January 7, 2019.
- 20. Charlson ME, Pompei P, Ales KL et al. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987;40:373–383. [DOI] [PubMed] [Google Scholar]
- 21. Deyo RA, Cherkin DC, Ciol M. Adapting a clinical comorbidity index for use with ICD‐9‐CM administrative databases. J Clin Epidemiol 1992;45:613–619. [DOI] [PubMed] [Google Scholar]
- 22. Romano PS, Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD‐9‐CM administrative data: Differing perspectives. J Clin Epidemiol 1993;46:1075–1079. [DOI] [PubMed] [Google Scholar]
- 23. Mor V, Wagner TH, Levy C et al. Association of expanded VA hospice care with aggressive care and cost for veterans with advanced lung cancer. JAMA Oncol 2019;5:810–816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Degenholtz HB, Thomas SB, Miller MJ. Race and the intensive care unit: Disparities and preferences for end‐of‐life care. Crit Care Med 2003;31(suppl):S373–S378. [DOI] [PubMed] [Google Scholar]
- 25. Kinoshita H, Maeda I, Morita T et al. Place of death and the differences in patient quality of death and dying and caregiver burden. J Clin Oncol 2015;33:357–363. [DOI] [PubMed] [Google Scholar]
- 26. Wright AA, Zhang B, Ray A et al. Associations between end‐of‐life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA 2008;300:1665–1673. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Fennell ML. Racial disparities in care: Looking beyond the clinical encounter. Health Serv Res 2005;40:1713–1721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Nelson A. Unequal treatment: Confronting racial and ethnic disparities in health care. J Natl Med Assoc 2002;94:666–668. [PMC free article] [PubMed] [Google Scholar]
- 29. Oster A and Bindman AB. Emergency department visits for ambulatory care sensitive conditions: Insights into preventable hospitalizations. Med Care 2003;41:198–207. [DOI] [PubMed] [Google Scholar]
- 30. Doty MM, Holmgren AL. Health care disconnect: Gaps in coverage and care for minority adults. Findings from the Commonwealth Fund Biennial Health Insurance Survey (2005). Issue Brief (Commonw Fund) 2006;21:1–12. [PubMed] [Google Scholar]
- 31. Jacobs E, Chen AH, Karliner LS et al. The need for more research on language barriers in health care: A proposed research agenda. Milbank Q 2006;84:111–133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Anderson KO, Richman SP, Hurley J et al. Cancer pain management among underserved minority outpatients: Perceived needs and barriers to optimal control. Cancer 2002;94:2295–2304. [DOI] [PubMed] [Google Scholar]
- 33. Muni S, Engelberg RA, Treece PD et al. The influence of race/ethnicity and socioeconomic status on end‐of‐life care in the ICU. Chest 2011;139:1025–1033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Dallo FJ, Borrell LN, Williams SL. Nativity status and patient perceptions of the patient‐physician encounter: Results from the Commonwealth Fund 2001 survey on disparities in quality of health care. Med Care 2008;46:185–191. [DOI] [PubMed] [Google Scholar]
- 35. LoPresti MA, Dement F, Gold HT. End‐of‐life care for people with cancer from ethnic minority groups: A systematic review. Am J Hosp Palliat Care 2016;33:291–305. [DOI] [PubMed] [Google Scholar]
- 36. Garrett JM, Harris RP, Norburn JK et al. Life‐sustaining treatments during terminal illness: Who wants what? J Gen Intern Med 1993;8:361–368. [DOI] [PubMed] [Google Scholar]
- 37. Belisomo R. Reversing racial inequities at the end of life: A call for health systems to create culturally competent advance care planning programs within African American communities. J Racial Ethn Health Disparities 2018;5:213–220. [DOI] [PubMed] [Google Scholar]
- 38. Kazandjian D, Suzman DL, Blumenthal G et al. FDA approval summary: Nivolumab for the treatment of metastatic non‐small cell lung cancer with progression on or after platinum‐based chemotherapy. The Oncologist 2016;21:634–642. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
See http://www.TheOncologist.com for supplemental material available online.
Supplemental Figure
Supplemental Tables