Racial disparities and treatment trends in a large cohort of elderly black and white patients with nonsmall cell lung cancer

Authors


Abstract

BACKGROUND:

This study investigated whether there was a significant gap in receipt of treatment for nonsmall cell lung cancer (NSCLC) between blacks and whites, and whether the gap or disparity changed during the past 12 years from 1991 to 2002.

METHODS:

The study population consisted of 83,101 patients including 75,141 (90.4%) whites and 7960 (9.6%) blacks aged ≥65 years who were diagnosed with American Joint Committee on Cancer (AJCC) stages I-IV NSCLC identified from the Surveillance, Epidemiology, and End Results (SEER) program's Medicare database. Age-adjusted and sex-adjusted rates and crude and adjusted odds ratios for receiving appropriate stage-specific treatment of NSCLC were reported.

RESULTS:

For stages I-II NSCLC combined, blacks were 37% less likely (OR, 0.63; 95% confidence interval [CI], 0.55-0.73) to receive surgery, 42% less likely (OR, 0.58; 95% CI, 0.36-0.92) to receive chemotherapy, and for stages III-IV combined, 57% less likely (OR, 0.43; 95% CI, 0.30-0.61) to receive chemotherapy compared with whites. Older patients, women, and those in lower socioeconomic quartiles had greater disparities in receipt of treatment compared with the highest income quartile. Disparity trends were not significantly narrowed during the past 12 years between blacks and whites for receipt of the above treatments.

CONCLUSIONS:

There have been substantial disparities in receiving recommended treatments between blacks and whites, and these disparities have been relatively stable without a significant trend of narrowing during the past 12 years. Efforts should focus on providing appropriate quality treatment and educating blacks on the value of having these treatments to reduce these disparities in receipt of treatment for NSCLC. Cancer 2009. © 2009 American Cancer Society.

Lung cancer is the leading cause of death from all types of cancers in the United States with an estimated 161,840 deaths projected for 2008.1 Although crude death rates are similar for blacks and whites, blacks have a higher age-adjusted death rate of 95.8 versus 72.6 per 100,000 population and a 1.3 death rate ratio.2

Currently, there is no effective screening tool available for lung cancer, although a large trial is underway to determine whether a computed tomography scan can be effective for screening for higher risk populations of smokers.3-5 Early findings from this study suggest an 80% early detection for stage I lung cancer. There are no specific recommendations for lung cancer screening for the general population; however, smokers are encouraged to stop smoking and have routine evaluations with their primary care physicians on an annual basis.6 As a result of the lack of effective screening tests and the finding that symptoms for lung cancer generally do not appear until the disease has progressed, most cases of lung cancer are diagnosed in the later stages,1, 2 which is typical for blacks who also tend to receive less aggressive treatment that can potentially result in poorer survival.7, 8

The National Comprehensive Cancer Network (NCCN) recommends surgical resection and adjuvant treatment with chemotherapy or chemoradiation for patients with stage I-II lung cancer. For stage IIIA, surgery with chemoradiation and chemotherapy or chemotherapy and mediastinal radiation therapy are recommended when margins are negative or positive, respectively. With further metastasis typical of distant spread of lung cancer, as in stages IIIB-IV, radiation and/or chemotherapy, or chemoradiation is usually recommended. Patients who are expected to undergo treatment are counseled on smoking cessation. For patients with negative margins, sequential chemotherapy or radiation therapy is given. Alternatively, for patients with positive margins, concurrent chemotherapy and/or radiation therapy is the treatment of choice.6

Several studies have identified that the disproportionate disadvantages that blacks experience in receipt of treatment for nonsmall cell lung cancer (NSCLC) compared with whites may be responsible for substantial racial differences in lung cancer death rates.9-12 Surgical resection is the optimal treatment for early stage lung cancer. Blacks tend to have unequal access to high-quality treatment and lower utilization of treatment options for surgery and chemotherapy compared with whites.9-11, 13-15 Moreover, blacks with advanced stage lung cancer are more likely to receive single-agent chemotherapy rather than platinum-based combination therapy, which is associated with improved survival.12

Although racial disparities in lung cancer have been documented for surgical resection for early stages of NSCLC,9-11, 13-15 it is not clear whether these racial disparities still exist for receipt of surgery and whether or not they also exist for receipt of chemotherapy and radiation therapy. More interestingly, there is no study that addresses whether racial disparities have changed over time for these standard therapies. Therefore, this study was conducted to first determine whether there is a difference in the receipt of treatment for NSCLC for stage-specific treatment between blacks and whites. The second aim of this study was to examine whether these racial disparities have changed during the past 12 years from 1991 to 2002. We hypothesized that blacks are more likely to receive less treatment compared with whites, but that the treatment disparities between blacks and whites would have narrowed during the past 12 years.

MATERIALS AND METHODS

Data Sources

Data from the Surveillance, Epidemiology, and End Results (SEER) program was linked to the Medicare database. The SEER program collects and reports data on cancer incidence and survival in 17 SEER registries that gather data on 26% of the US population. In this study, we chose to combine Atlanta and rural Georgia as 1 SEER registry, and all 4 registries in California as another SEER registry.16 Medicare is the official federal health insurance for persons aged ≥65 years and those with disabilities. This insurance offers various health plans, namely, Medicare parts A, B, C, and D with designated covered healthcare benefits.16, 17 Patients who did not have full coverage for Medicare parts A and B, which pays for hospital expenses and physicians visits, or who were members of a health maintenance organization (HMO) in the year of their diagnosis were excluded from the study. The University of Texas Health Science Center at Houston (Texas) reviewed and approved the study protocol.

Study Population

The study's target population comprised 89,834 Medicare patients aged ≥65years who were diagnosed with American Joint Committee on Cancer (AJCC) stages I-IV NSCLC from January 1991 to December 2002. The study sample consisted of 83,101 patients (90.4% whites and 9.6% blacks) diagnosed with NSCLC from 1991 to 2002. Patients with missing or unknown information (n = 11,510) and cases with unknown SES values (n = 852) were excluded from the study; thus a total of 70,901 patients were included in the final multivariate analyses (Tables 4 and 5).

Study Variables

The outcome variables were receipt of surgery, chemotherapy, and radiation therapy. Patients were stratified into 2 groups according to AJCC stage as follows: early (stages I-II) and late (stages III-IV) disease. Patients were defined as having received definitive surgery for stages I-II when their SEER codes were 10 to 70, or 90, which include local surgical excision, partial wedge/segmental resection, partial lobectomy, resection of the whole lung, or at least 1 lobe, but less than the whole lung or when patients had made Medicare claims of resection identified by the International Classification of disease (ICD)-9 procedure codes 32.29-32.99,18 and Common Procedure Terminology (CPT) codes 32440-32525, 32657, or 32663.19 Information on chemotherapy20, 21 and radiation therapy identified from SEER22 or from Medicare claims23 are described elsewhere.

The exposure variable, ethnicity, was categorized as whites (non-Hispanic) and blacks (non-Hispanic African Americans and blacks from other countries). Covariates used in adjusted analyses were age, sex, marital status, and histology. A comorbidity score was calculated by using the Deyo adaptation of the Charlson comorbidity index,24-26 with several procedure codes that reflected the Romano revision,27 and categorized into 0, 1, 2, 3, ≥4. The percentage of individuals living below the poverty line at the census-tract level was used to define socioeconomic status (SES) in quartiles.28 In further adjusted models, tumor grade and tumor size were added to alternate models to assess changes-in-effect estimates.

Statistical Analysis

Analyses were conducted by using STATA, version 10.0 (StataCorp, College Station, Tex). Descriptive statistics for each ethnic/racial group with number and percentage in subgroups for all covariates were performed by using Pearson chi-square tests with a P value of .05 as the cutoff point (Table 1). Similar descriptive statistics were calculated for receipt of treatment for stages I-II and stages III-IV (Table 2). Age-adjusted and sex-adjusted rates, computed by using the SEER 2002 total population as the standard, were compared between ethnic groups, and the disparity for these rates was calculated (Table 3). In addition, graphs were constructed to test linear trend in age-adjusted and sex-adjusted rates of receiving treatment by race over time from 1991 to 2002 for surgery and chemotherapy (Figs. 1 and 2). Trend tests were performed on disparities in treatments from 1991 to 2002 by using normalized log-transformed age-adjusted and sex-adjusted rates in linear regression.

Figure 1.

Percentage of patients who received surgery for stages I-II nonsmall cell lung cancer in 1991–2002, stratified by race/ethnicity.

Figure 2.

Percentage of patients who received chemotherapy for stages III-IV nonsmall cell lung cancer in 1991–2002, stratified by race/ethnicity.

Table 1. Distribution of Patient Characteristics With Nonsmall Cell Lung Cancer Stratified by Race, 1991–2002
CharacteristicWhites (n=75,141)Blacks (n=7960)P
No.%No.%
  1. P values were calculated by using Pearson chi-square tests of hypothesis for independence to assess statistical significance of relationships between ethnic groups.

Age, y    <.0001
 65-6917,52523.3244830.8 
 70-7420,79227.7233929.4 
 75-7918,89025.2174521.9 
 80-8412,32116.4101012.7 
 ≥8556137.54185.3 
Sex    <.0001
 Men41,57355.3476859.9 
 Women33,56844.7319240.1 
Socioeconomic status    <.0001
 Highest quartile20,78627.73504.4 
 2nd quartile20,26827.05537.0 
 3rd quartile19,30725.7124715.7 
 Lowest (poorest) quartile13,98918.6574972.2 
 Unknown7911.1610.8 
Year of diagnosis    .605
 199151766.95496.9 
 199252827.05827.3 
 199352547.05216.6 
 199452407.05526.9 
 199551016.85617.1 
 199650046.75026.3 
 199747586.35136.4 
 199845846.14956.2 
 199944575.95066.4 
 2000991813.2102012.8 
 200110,17413.5105513.3 
 200210,19313.5110413.9 
SEER registry area    <.0001
 Connecticut955112.74115.2 
 Detroit988213.2264933.3 
 Hawaii5060.750.1 
 Iowa930812.41181.5 
 New Mexico25463.4370.5 
 Seattle748810.01802.3 
 Utah18092.4160.2 
 Atlanta/Rural Georgia35674.897412.2 
 Kentucky39265.22373.0 
 Louisiana23313.16548.2 
 New Jersey47826.45667.1 
 California19,44525.9211326.6 
Comorbid score    <.0001
 026,98935.9255632.1 
 126,76735.6255932.2 
 211,77215.7132216.6 
 349836.66568.2 
 ≥446306.286710.9 
Tumor stage    <.0001
 Stage-I15,74121.0144218.1 
 Stage-II24923.31762.2 
 Stage-IIIA762310.183810.5 
 Stage-IIIB13,60518.1169921.3 
 Stage-IV25,15333.5282235.5 
 Unknown10,52714.098312.3 
Histology    <.0001
 Carcinoma, nonspecific22,99430.6234829.5 
 Large cell51726.95567.0 
 Squamous cell19,35225.8260032.7 
 Adenocarcinoma24,08332.1217627.3 
 Bronchus25793.41792.3 
 Adenosquamous cell9411.31011.3 
Tumor grade    <.0001
 Well27233.62142.7 
 Moderate10,43013.9105313.2 
 Poor21,97529.3232029.2 
 Undifferentiated38435.13204.0 
 Missing/Unknown36,17048.1405350.9 
Tumor size, cm    <.0001
 <16830.9700.9 
 1-240525.43564.5 
 2-3839011.27409.3 
 3-4860811.580810.2 
 ≥422,97130.6268933.8 
 Missing30,43740.5329741.4 
Surgery    <.0001
 No58,47578.7673885.3 
 Yes15,87521.4116114.7 
Chemotherapy    <.0001
 No54,48772.5596775.0 
 Yes20,65428.0199325.0 
Radiation therapy    .006
 No39,49952.6405550.9 
 Yes35,64247.4390549.1 
Table 2. Comparison of Treatment for Nonsmall Cell Lung Cancer Received by Whites and Blacks 1991–2002, Stratified by Tumor Stage
Receipt of TherapyWhitesBlacksP
No.%No.%
 Stage I
Surgery    <.0001
 No634140.778154.5 
 Yes923759.365345.5 
Chemotherapy    .007
 No13,53886.9121084.4 
 Yes204013.122415.6 
Radiation therapy    <.0001
 No10,66868.590763.3 
 Yes491031.552736.8 
Stage II
Surgery    <.0001
 No59924.27341.7 
 Yes187275.810258.3 
Chemotherapy    .332
 No182473.813577.1 
 Yes64726.24022.9 
Radiation therapy    .147
 No1,12845.77040.0 
 Yes134354.410560.0 
Stage III
Surgery    <.0001
 No41,80684.3497888.9 
 Yes411615.734311.1 
Chemotherapy    <.0001
 No30,06263.8376568.7 
 Yes15,86036.2155631.3 
Radiation therapy    .838
 Yes20,35143.1243443.3 
 No25,57156.9288756.7 
Stage IV
Surgery    .003
 No20,35196.7243497.8 
 Yes25,5713.328872.3 
Chemotherapy    <.0001
 No16,65266.9203672.6 
 Yes824433.176727.4 
Radiation therapy    .009
 No11,29445.4134448.0 
 Yes13,60254.6145952.1 
Table 3. Age-adjusted and Sex-adjusted Percentage of Whites and Blacks Who Received Treatment From 1991 to 2002 for Nonsmall Cell Lung Cancer Stratified by Tumor Stage Combined
Year of DiagnosisRates by Race Adjusted by Age and Sex
WhitesBlacksDisparityWhitesBlacksDisparity
  Stages I-II  
Surgery (n=19,658)Chemotherapy (n=19,658)
199163.241.821.48.013.8−5.9
199260.243.416.810.018.2−8.2
199359.843.816.08.68.8−0.2
199458.449.49.011.515.0−3.6
199559.149.69.510.618.2−7.5
199660.739.521.213.217.9−4.7
199761.849.612.215.313.12.2
199858.943.915.016.915.71.2
199959.559.8−0.317.89.97.9
200059.041.117.918.420.0−1.5
200160.243.916.317.721.3−3.6
200263.653.010.616.618.3−1.7
 Stages III-IV
 Chemotherapy (n=51,780)Radiation Therapy (n=51,780)
199121.617.24.460.562.4−1.9
199223.423.6−0.158.255.03.2
199321.418.52.959.659.20.4
199425.019.25.855.855.70.1
199527.120.07.157.054.03.0
199629.521.08.655.952.63.2
199733.133.7−0.656.251.64.6
199836.429.96.555.257.5−2.4
199937.335.81.552.951.51.4
200040.033.26.852.354.5−2.2
200139.033.06.051.349.12.2
200240.735.84.950.949.41.4

Univariate and multivariate logistic regression analyses were performed to assess the relative odds of receiving treatment for blacks compared with whites. Crude and multivariate odds ratios for receipt of treatment stratified by years of diagnosis for 1991-1995, 1996-1999, and 2000-2002 are reported in Table 4. Interaction was assessed by identifying effect modifiers for race/ethnicity with other confounding variables by using the likelihood ratio test with a P-value cutoff point of .05. These were included along with single-order covariates in the model.

Table 4. Crude and Multivariate Adjusted Odds Ratios and 95% Confidence Intervals of Receiving Treatment for Nonsmall Cell Lung Cancer by Race, 1991-2005 Stratified by Tumor Stage Combined
 Stages I-IIStages III-IV
Odds Ratios (95% CI)
Surgery (n=19,658)Chemotherapy (n=19,658)Chemotherapy (n=51,243)Radiation Therapy (n=51,243)
  • CI indicates confidence intervals; SES, socioeconomic status.

  • Models were adjusted for age, sex, socioeconomic status level, diagnosis year, marital status, histology, comorbidity score, and SEER 17 areas. Chemotherapy and radiation therapy were added to the adjusted model for surgery at stages I-II; surgery and radiation therapy were added to model for chemotherapy at stages I-II and stages III-IV; surgery and chemotherapy were added to the adjusted model for radiation therapy for stages III-IV.

  • *

    Statistically significant odds ratios (P value < .05).

Crude Model    
 Whites1.001.001.001.00
 Blacks0.55 (0.50-0.61)*1.12 (0.97-1.28)0.78 (0.74-0.83)*0.93 (0.84-1.02)
Adjusted Model    
Race/Ethnicity    
 Whites1.001.001.001.00
 Blacks0.63 (0.55-0.73)*0.58 (0.36-0.92)*0.43 (0.30-0.61)*1.02 (0.95-1.10)
Age, y    
 65-691.001.001.001.00
 70-740.75 (0.68-0.83)*0.89 (0.80-0.99)*0.81 (0.77-0.85)*0.83 (0.79-0.88)*
 75-790.45 (0.44-0.50)*0.61 (0.55-0.69)*0.54 (0.51-0.57)*0.68 (0.64-0.74)*
 80-840.22 (0.20-0.25)*0.37 (0.32-0.44)*0.30 (0.28-0.32)*0.51 (0.48-0.54)*
 ≥850.10 (0.08-0.12)*0.24 (0.19-0.31)*0.16 (0.14-0.18)*0.36 (0.34-0.40)*
Sex    
 Men1.001.001.001.00
 Women0.95 (0.87-1.03)0.54 (0.39-0.76)*0.72 (0.62-0.84)*1.03 (0.99-1.07)
SES    
 Highest quartile1.001.001.001.00
 2nd quartile0.92 (0.84-1.14)0.91 (0.81-1.02)0.87 (0.78-0.96)*1.01 (0.96-1.07)
 3rd quartile0.78 (0.75-1.03)0.96 (0.85-1.09)0.76 (0.63-0.90)*0.93 (0.88-0.99)*
 Lowest quartile0.68 (0.70-0.99)*0.85 (0.74-0.98)*0.60 (0.45-0.79)*0.88 (0.82-0.93)*
Year of diagnosis    
 19911.001.001.001.00
 19920.81 (0.65-1.20)1.29 (0.17-1.72)1.24 (1.10-1.45)*0.96 (0.86-1.07)
 19930.91 (0.74-1.12)1.12 (0.84-1.50)1.08 (0.95-1.22)1.04 (0.94-1.17)
 19940.85 (0.69-1.05)1.59 (1.21-2.09)*1.40 (1.23-1.58)*0.88 (0.79-0.98)*
 19950.86 (0.70-1.06)1.62 (1.23-2.14)*1.52 (1.35-1.72)*0.89 (0.80-0.99)*
 19960.92 (0.74-1.13)2.02 (1.54-2.66)*1.80 (1.60-2.09)*0.86 (0.77-0.96)*
 19971.01 (0.81-1.25)2.41 (1.83-3.16)*2.32 (2.05-2.62)*0.85 (0.76-0.95)*
 19980.84 (0.68-1.05)2.98 (2.27-3.93)2.66 (2.35-3.00)*0.81 (0.73-0.91)*
 19990.96 (0.77-1.20)2.98 (2.27-3.92)*3.05 (2.70-3.44)*0.72 (0.65-0.81)*
 20000.89 (0.73-1.08)3.62 (2.81-4.65)*3.47 (3.11-3.88)*0.71 (0.64-0.78)*
 20010.98 (0.80-1.19)3.43 (2.67-4.40)*3.39 (3.04-3.78)*0.69 (0.62-0.76)*
 20021.13 (0.93-1.38)3.33 (2.58-4.29)*3.76 (3.37-4.19)*0.66 (0.60-0.73)*

RESULTS

Characteristics of the Study Population

As seen in Table 1, a higher proportion of whites were diagnosed at age ≥80 years, whereas more blacks were diagnosed at younger ages; however, median age at diagnosis was similar. A larger proportion of white women (44.7%) were diagnosed with NSCLC compared with 40.1% in black women (P < .001). Most blacks (72.2%) lived in the poorest quartile of SES compared with 18.6% of whites. Most whites (52.4%) were married at the time of diagnosis compared with 35.9% of blacks. A higher percentage of black patients were diagnosed at later stages. Blacks also had higher comorbidity scores and were more likely to have a diagnosis of squamous cell cancer compared with whites, who were more likely to be diagnosed with adenocarcinoma.

Comparison of Treatments

Table 2 presents a summary of treatments rendered, which were stratified by ethnic/racial groups and tumor stage. At stages I and II, a greater proportion of whites (59.3% and 75.8%, respectively) received surgical resection compared with blacks (45.5% and 58.3%, respectively). There were also significant differences at stage I for receipt of chemotherapy for blacks versus whites (15.6% vs 13.1%) and for radiation therapy for blacks versus whites (36.8% vs 31.5%). There were no significant differences between the racial groups for receipt of these therapies at stage II.

The majority of black and white patients did not receive surgery at stages III and IV (88.9% vs 83.3%). This trend was similar for receipt of chemotherapy (68.7% vs 63.8%) for blacks versus whites. There were no significant differences for receipt of radiation therapy at stage III, but at stage IV, there was a small difference in receipt of this therapy for blacks versus whites (52.1% vs 54.6%).

Age-Sex–adjusted Rates in Receiving Treatment

Table 3 presents the age-sex–adjusted rate differences in receiving treatment. For stages I-II, blacks appeared to have lower rates for receipt of surgery compared with whites. From 1991 to 2002, the percentage of whites who received surgery remained relatively stable, from 63.2% to 63.6%. On the contrary, blacks had a greater increase in receiving surgery from 41.8% in 1991 to 53.0% in 2002. Figure 1 is a graphical view of this disparity. From 1991-1998 and 2000-2002, there were noticeable disparities in receipt of surgery between blacks and whites, but in 1999, there was no disparity in receipt of surgery between these 2 groups. The overall disparity trend was not statistically significant (P = .275). For chemotherapy, there was a small disparity for stages I-II between whites and blacks, although this disparity increased as much as 8% in 1992 and 1999 (Table 3). From 1991-1992, 1994-1996, and 2000-2002, more blacks were given chemotherapy. In addition, from 1997-1999, the receipt of chemotherapy was lower for blacks compared with whites but was not statistically significantly (P = .089).

Although receipt of chemotherapy increased for both groups over time for stages III-IV, there was a disparity between blacks and whites (Table 3 and Fig. 2). The greatest disparity was observed between 1992 and 1996. The disparity decreased noticeably in succeeding years, but was still apparent in 2002. Overall trend was not statistically significant (P = .498). Receipt of radiation therapy for stages III-IV showed a small disparity between whites and blacks over time for some years, but trend was not significant (P = .992).

Crude and Multivariate Adjusted Analyses

Table 4 presents results of both univariate and adjusted odds ratios that controlled for potential confounding variables such as age and comorbidity. For stages I-II, blacks were 37% (OR, 0.63; 95% confidence interval [CI], 0.55-0.73) less likely to receive surgery than whites. Similarly, blacks with stages I-II were 42% (OR, 0.58; 95% CI, 0.36-0.92) less likely to receive chemotherapy than whites, and for stages III-IV, blacks were 57% (OR, 0.43; 95% CI, 0.30-0.61) less likely than whites to receive chemotherapy. Adjusted analyses for separate stages agreed with combined stage analyses. Blacks were 34% (OR, 0.66; 95% CI, 0.56-0.78) less likely and 52% (OR, 0.48; 95% CI, 0.31-0.74) less likely to receive surgery at stage I and II; 40% (OR, 0.60; 95% CI, 0.36-0.99) less likely, and 0.58% (OR, 0.42; 0.44-0.78) less likely to receive chemotherapy at stage I and stage IV. There were no significant differences for receipt of chemotherapy at stage III.

In further stage combined adjusted analyses (stages I-II and III-IV), older patients were less likely to receive surgery and chemotherapy. Women were also less likely to receive chemotherapy than men, but this association was not statistically significant for surgery. Patients at the lowest SES quartile were less likely to receive surgery (OR, 0.68; 95% CI, 0.70-0.99) and chemotherapy (OR, 0.85; 95% CI, 0.74-0.98) compared with the highest SES quartile. Patients diagnosed in 1994-2002 were more likely to receive chemotherapy for stages I-II. Blacks with stages I-II who did not receive surgery were 47% (OR, 0.53; 95% CI, 0.29-0.95) less likely to be given chemotherapy compared with whites. There were no significant differences in receipt of chemotherapy between blacks and whites when surgery was given. For stages III-IV, black patients were 57% (OR, 0.43; 95% CI, 0.30-0.61) less likely to receive chemotherapy than whites. Furthermore, blacks were less likely to receive chemotherapy regardless of whether radiation therapy was given.

Receipt of radiation therapy was not significantly different between blacks and whites with stages III-IV. Analyses at separate stages agreed with the stage-combined analyses. In stage III-IV–combined analyses older patients, women, and those with poor SES status were less likely to receive chemotherapy and radiation therapy. These results in disparity did not change considerably between blacks and whites after further controlling for tumor grade and tumor size in adjusted models as follows: for stages I-II, receipt of surgery (OR, 0.66; 95% CI, 0.56-0.79) and chemotherapy (OR, 0.56, 95% CI, 0.35-0.90); for stages III-IV receipt of chemotherapy (OR, 0.43, 95% CI, 0.30-0.61) and radiation therapy (OR, 1.01; 95% CI, 0.94-1.08).

The disparities in receiving treatment between blacks and whites were stratified by years of diagnosis for 1991-1995, 1996-1999, and 2000-2002 (Table 5). Blacks had lower odds of receiving surgery for all tratified diagnosis years for both crude and adjusted analyses. There were no significant differences in receipt of treatment between blacks and whites for chemotherapy at stages I-II. However, for stages III-IV in 2000-2002, blacks were 76% (OR, 0.24; 95% CI, 0.14-0.40) less likely to receive chemotherapy than whites. Receipt of radiation therapy was not significantly different between ethnic groups for all stratified time periods.

Table 5. Crude and Multivariate Odds Ratios and 95% Confidence Intervals for Disparities of Treatment by Race Stratified by Years of Diagnosis, 1991–2002
Treatment1991–19951996–19992000-2002
 Odds Ratios (95% CI)
  • CI indicates confidence intervals.

  • See Table 4 for adjusted variables in models.

  • *

    Statistically significant odds ratios (P < .05).

Surgery, stages I-II   
Crude model (blacks vs whites)0.55 (0.46-0.65)*0.56 (0.46-0.69)*0.55 (0.47-0.65)*
Adjusted Model   
Race/Ethnicity   
 Whites1.001.001.00
 Blacks0.68 (0.53-0.87)*0.56 (0.41-0.75)*0.61 (0.49-0.77)*
Sex   
 Men1.001.001.00
 Women0.89 (0.77-1.02)0.93 (0.79-1.09)1.00 (0.89-1.14)
Chemotherapy, stages I-II   
Crude model (blacks vs whites)1.36 (1.06-1.74)*0.89 (0.67-1.18)1.17 (0.95-1.43)
Adjusted Model   
Race/Ethnicity   
 Whites1.001.001.00
 Blacks0.78 (0.35-1.77)0.34 (0.12-0.91)*0.58 (0.29-1.19)
Sex   
 Men1.001.001.00
 Women0.61 (0.33-1.13)0.38 (0.19-0.75)*0.59 (0.35-0.98)*
Chemotherapy, stages III-IV   
Crude model (blacks vs whites)0.78 (0.69-0.89)*0.79 (0.70-0.89)*0.77 (0.70-0.84)*
Adjusted Model   
Race/Ethnicity   
 Whites1.001.001.00
 Blacks0.74 (0.35-1.54)0.76 (0.39-1.47)0.24 (0.14-0.40)*
Sex   
 Men1.001.001.00
 Women1.01 (0.75-1.38)0.66 (0.49-0.90)*0.64 (0.51-0.80)*
Radiation therapy, stages III-IV   
Crude model (blacks vs whites)0.92 (0.83-1.02)0.92 (0.82-1.03)0.98 (0.89-1.06)
Adjusted Model   
Race/Ethnicity   
 Whites1.001.001.00
 Blacks0.93 (0.82-1.06)1.05 (0.91-1.21)1.03 (0.93-1.14)
Sex   
 Men1.001.001.00
 Women1.04 (0.97-1.12)1.01 (0.93-1.09)1.03 (0.97-1.10)

In a separate analysis that used only 12 SEER areas and that excluded those SEER areas that joined SEER from 2000 and beyond (other areas of California, Kentucky, Louisiana, and New Jersey), the effect estimate for receipt of surgery at stages I-II was similar between blacks and whites when the 17 SEER areas were in the model (OR, 0.62; 95% CI, 0.54-0.72). Likewise, for receipt of chemotherapy at stages I-II, the effect estimate for blacks compared with whites remained unchanged for the 12 SEER areas. The results were similar in racial disparities for receipt of chemotherapy and radiation therapy for stages III-IV.

Two-way interactions identified were race with sex for chemotherapy for stages I-II; race with sex, and race with poverty level for chemotherapy for stages III-IV. The interaction between race and histology was identified for radiation therapy for advanced stages, but this did not change the direction of the effect estimate for race/ethnicity, so the simple model without interaction terms were chosen over the model with interaction terms. There were no interactions identified for surgery for stages I-II.

DISCUSSION

This study of a large cohort of elderly black and white patients with NSCLC had several important findings. Concerning receipt of treatment, surgery, the optimal care for stages I-II that results in a substantial cure, was lower for blacks compared with whites. Blacks who did not receive surgery for stages I-II were also less likely to be given chemotherapy. However, there was no difference in receipt of chemotherapy when surgery was performed. Blacks had lower receipt of chemotherapy overall (stages I-IV). Blacks were also less likely to receive chemotherapy regardless of whether or not radiation therapy was given, but the disparity in chemotherapy was much more increased when radiation therapy was not given. Older patients, women, and persons in lower SES quartiles had greater disparities in receipt of treatment. Disparities in receipt of surgery were constant over time when stratified by years of diagnosis for 1991-1995, 1996-1999, and 2000-2002. Disparities for receipt of chemotherapy were found for periods 1996-1997 for stages I-II and for 2000-2002 for stages III-IV. Trends in racial disparities in treatments were not significantly decreased over time from 1991-2002.

Disparities in receipt of treatment have also been reported by other investigators for NSCLC in SEER. Lathan et al29 found that black patients were 45% less likely to have surgery for early stage NSCLC compared with whites (OR, 0.55; 95% CI, 0.47-0.64). Bach et al9 found the rate of surgery for blacks to be 12 points lower compared with whites (64.0 vs 76.7 [P < .001]). In addition, Greenwald et al10 found that blacks were 23% less likely to receive surgery for early stage lung cancer compared with whites and that median family income was a strong predictor for receipt of surgery. Neighbors et al11 study performed on non-SEER patients found that black patients had lower receipt of surgery at low-volume hospitals (OR,0.45; 95% CI, 0.34-0.58). For receipt of chemotherapy for advanced stages of lung cancer, Earle et al14 found that African Americans were 30% less likely to receive chemotherapy for metastasis of lung cancer (OR, 0.70; 95% CI, 0.55-0.88). Hayman et al30 reported similar findings in his study on the use of radiation therapy with elderly patients aged ≥80 years with metastatic NSCLC. Our study found comparable findings with these other studies in that receipt of surgery for early stage NSCLC and chemotherapy for both early stage and late-stage NSCLC was less for blacks compared with whites. To our knowledge, no other study has addressed these changes in receipt of treatment over time for NSCLC.

The reasons for these disparities are multifactorial. One reason documented by Lathan et al29 is that black patients were less likely to be appropriately staged and to receive recommendations for surgery and more likely to refuse surgical treatment compared with whites.31 One common belief among blacks, besides lack of trust in the healthcare system,32, 33 is that of tumor spread and metastasis when the cancer is exposed to air.34 Blacks also tend to use spirituality, prayer, and healing techniques to deal with their illness,35 and the belief that “when your time is up, it is up” is widely accepted by this ethnic group. Patients with low median income are also less likely to receive surgical treatment for early stages of NSCLC because of limited access to healthcare, and black patients most often have this experience.10 The aforementioned reasons may cause blacks to less often seek medical attention for cancer symptoms. In addition, blacks are most often seen at county hospitals, which often provide lower quality medical therapy.36, 37

A strength of the current study is the large, retrospective, population based from 17 SEER areas that account for 26% of the US population; thus, a greater generalization of our study findings can be applied. Moreover, data collected by SEER is of high quality and is subject to rigorous review. Another unique strength of this study is that we were able to observe changes in disparity that took place during a 12-year time period. In addition, the receipt of surgery for stages I-II for blacks remained relatively stable and did not change drastically during the time period selected for the study. Alternatively, for chemotherapy for stages I-II, the disparity in receipt of treatment for blacks compared with whites was in the 1996-1999 period, and for stages III-IV, this occurred in 2000-2002. It is critical to know when these disparities took place, which ethnic groups were severely affected, and how the disparities changed over time, so that efforts can be concentrated on targeting problems and rectifying discrepancies in administration of treatments for the purposes of decreasing mortality and improving quality of life.

Furthermore, although SEER registries have well-validated information on surgery and radiation therapy, the completeness of data can still be enhanced by combining them with Medicare claims on these therapies. Moreover, chemotherapy and growth factors can be uniquely identified from Medicare data. Finally, comorbid conditions that may have an impact on the receipt of various therapies can be identified from Medicare data and can be controlled in analyses for racial disparities in receiving these cancer treatments.

There are several limitations to this study. First, there was no information on patient preferences for treatment and refusals despite physicians' recommendations or on other related issues that may have played a role in treatment decisions. Black Medicare patients compared with whites are less likely to be able to afford coinsurance, deductibles, copayments, other out-of-pocket expenses, and treatments not covered by Medicare, which are important to the selection of treatment options. In this study, both ethnic groups had Medicare insurance; however, blacks were less likely to receive the recommended treatment for NSCLC. Although blacks are less likely to have supplemental insurance coverage, these treatment disparities, in theory, should not exist.

Second, because our population was limited to Medicare beneficiaries ages 65 years or older, the findings may not be generalizable to younger patients. Due to their unavailability, certain variables found in the literature to be related to the receipt of cancer therapy were not included in our analyses. For example, treatment at a teaching hospital14, 29 and the use of a high-volume versus a low-volume hospital11 were reported to be risk factors, but they were not included in the adjusted model. Another limitation is that there are some missing values for several variables in the SEER data. For example, tumor stage had 11,510 missing or unknown entries that could lead to some misclassifications of tumor stages and subsequently to inappropriate treatment.

In conclusion, there were substantial disparities in receiving recommended treatments between blacks and whites, and these disparities have been relatively stable during the past 12 years. To reduce disparities in receipt of treatment for NSCLC, efforts should focus on providing appropriate quality treatment and educating blacks on the value of having these treatments.

Acknowledgements

We acknowledge the creation of this database from the National Cancer Institute, Centers for Medicare and Medicaid Services, and Surveillance, Epidemiology, and End Results tumor registries. Analyses, interpretation, and reporting are the sole responsibilities of the authors.

Conflict of Interest Disclosures

This study was supported by grant from the Agency for Healthcare Research and Quality (RO1-HS016743).

Ancillary