SEARCH

SEARCH BY CITATION

Keywords:

  • gastric adenocarcinoma;
  • race;
  • National Cancer Data Base;
  • lymph node evaluation;
  • multimodality therapy

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

BACKGROUND:

Race is associated with patterns of presentation and survival outcomes of gastric cancer in the United States. However, the impact of race on the receipt of guideline-recommended care is not well characterized. By using current recommendations, the authors examined the association between race and guideline-recommended treatments and identified factors that are predictive of variations in gastric cancer care.

METHODS:

By using the National Cancer Database for 1998 through 2005, 106,002 patients with gastric adenocarcinoma were identified. Multivariate analysis techniques were used to examine the association between race, the receipt of guideline-recommended care, and survival after adjusting for covariates.

RESULTS:

Although African-American and Hispanic patients were more likely to undergo adequate lymphadenectomy (≥15 lymph nodes) and to receive care at comprehensive cancer centers and high-volume facilities (for all, P ≤ .001), they were less likely to receive adjuvant multimodality therapy for American Joint Committee on Cancer stage IB through IV, lymph node-negative (M0) disease. Up to 60% of all patients who underwent gastrectomy failed to receive adequate lymphadenectomy and adjuvant multimodality therapy. The delivery of multimodality therapy varied significantly by stage and lymph node evaluation (P ≤ .001). These findings persisted on our multivariate analyses, indicating that African-American and Hispanic patients received adequate lymph node evaluation (P ≤ .001), whereas they were associated with receiving no adjuvant multimodality therapy (P ≤ .025).

CONCLUSIONS:

There were significant variations in treatment for gastric cancer among ethnic groups in the United States. It was noteworthy that, although nonwhite race was associated with improved surgical care, gastric cancer care remained suboptimal overall. Cancer programs need to identify procedures to maximize the delivery of adequate gastric cancer care to all patients. Cancer 2010. © 2010 American Cancer Society.

Gastric cancer remains the second most common cause of cancer-related deaths worldwide.1 Although its incidence continues to decline in the United States, the overall prognosis of gastric cancer remains poor.2 In this regard, several lines of investigations were conducted to improve the prognostication and survival outcomes of patients with gastric adenocarcinoma.3 Since 1997, researchers from the American Joint Committee on Cancer (AJCC) and others have revised the staging system to highlight the importance of adequate lymph node evaluation (≥15 lymph nodes).3, 4 In addition to its role in providing proper staging, adequate lymph node evaluation will also guide adjuvant therapy decisions, thus minimizing the risk of under treatment. Since 2001, prospective randomized trials, including the Intergroup 0116 trial and others, have supported the notion that multimodality therapy is effective when integrated in patients who undergo surgery for localized gastric cancer.5-7 Thus, over time, these important contributions have translated into treatment recommendations and guidelines to improve the quality of care and survival outcomes of gastric cancer in the United States.

To better understand the patterns of care and the prognosis of patients with gastric cancer in the United States, large, multihospital studies from the National Cancer Database (NCDB) demonstrated the presence of significant treatment variations of gastric cancer among US hospitals.8-12 We also demonstrated that race has an impact on patterns of disease presentation and survival rates in the United States.8 For example, African Americans and Hispanics were more likely than other races to present with AJCC stage III or IV disease and to receive no adjuvant therapy. Although advanced stage at presentation is not an unusual feature of gastric cancer in the West, these findings also suggest that ethnicity may influence whether a patient receives adequate treatment for gastric adenocarcinoma. However, the associations between ethnic-related, socioeconomic-related, or facility-related factors and guideline-recommended care for gastric cancer in the United States are not well characterized.

In light of the above, we hypothesized that treatment variations of gastric cancer among different races exist in the United States. By using current treatment recommendations (>15-lymph nodes evaluated and the administration of adjuvant multimodality therapy, as recommended by the Intergroup 0116 trial),13 we examined the extent to which race independently impacts the delivery of guideline-recommended care in the United States. To provide insights into uncovered factors associated independently with deviation from guideline-recommended care, we also examined other patient-related, tumor-related, and facility-related factors in the NCDB.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

Data Source

The NCDB is a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society. The NCDB was established in 1988 to serve as a comprehensive clinical surveillance resource for cancer care in the United States. Currently, 1460 hospitals participate in the approvals program and respond to annual calls for the delivery of data to the NCDB. Participating hospitals use cancer registrars to abstract data from charts for each patient according to a standardized set of data elements and definitions.

Case Selection Criteria

In total, 106,002 cases of gastric adenocarcinoma (International Classification of Disease codes C16.0-C16.9) were reported to the NCDB between 1998 and 2005. Tumor histologies included adenocarcinoma, not otherwise specified (NOS); intestinal-type adenocarcinoma; carcinoma tumor; mucinous adenocarcinoma; signet ring cell carcinoma; and other. Patients with other primary gastric tumors, such as neuroendocrine tumors, lymphoma, or sarcoma, or with secondary neoplasm were excluded from our analysis.

Study Design

For our bivariate analysis, we compared patient demographics and tumor-related, treatment-related, and facility-related features by race. For patient demographics, we included age, sex, year of diagnosis, health insurance status, median household income, and geographic location according the US Census region. For tumor-related features, we evaluated the primary tumor's location in the stomach, histologic grade, and AJCC stage. For patterns of treatment, we assessed types of gastric resections, adequacy of lymph node evaluation (≥15 lymph nodes, 1-14 lymph nodes, or no lymph nodes evaluated), and types of adjuvant therapy (surgery alone, surgery followed by single-modality therapy, and surgery followed by multimodality therapy). For the treatment facility variable, we evaluated types of hospital systems and facility case volumes. Types of hospital systems were categorized as teaching hospital cancer programs, comprehensive community hospital cancer programs, and community hospital cancer programs. The facility case volume was categorized by quartiles into the 0 to 25th, >25th to <50th, 50th to <75th, and 75th to 100th percentiles.

The American College of Surgeons has executed a Business Associate Agreement that includes a data-use agreement with each of its Commission on Cancer (CoC)-approved hospitals. Data reported to the NCDB are retrospective and include no patient or physician names. Analyses are reported only at the aggregate level and are used to assist hospital cancer programs with quality assurance, not to make decisions about individuals and their care. The results reported here are in compliance with the privacy requirements of the Health Insurance Portability and Accountability Act of 1996 as reported in the Standards for Privacy of Individually Identifiable Health Information; Final Rule (45 Code of Federal Regulation, parts 160 and 164).

Statistical Analysis

We used the chi-square test and t test for comparisons of proportions across levels of categorical and continuous variables, respectively. Results were based on 2-sided tests with a significance level of .05 and were adjusted for all pairwise comparisons using the Bonferroni correction.

To meet our study goals and as a representative of previously published investigations of evolving gastric cancer treatment, we used the current National Comprehensive Cancer Network guidelines to assess data available from the NCDB.3-6, 13 These guidelines included 1) adequate lymph node evaluation of ≥15 lymph nodes; and 2) the administration of adjuvant multimodality therapy (chemoradiotherapy) for AJCC stage IB through IV lymph node-negative (M0) disease (according to the Intergroup 0116 trial).6 We then used logistic regression analysis to predict the effect of race on: 1) inadequate lymphadenectomy (<15 lymph nodes evaluated), 2) zero lymph node evaluation, and 3) lack of adjuvant chemoradiotherapy (multimodality) therapy (for AJCC stage IB-IV M0 disease), after adjusting for covariates. In our final logistic regression models, we included clinically relevant variables as well as those deemed significant in our bivariate analysis. To assess predictors of overall survival, we used a Cox proportional hazards model for patients who were diagnosed from 1998 to 2000 (survival times were available only for those years). Given the correlation between African-American and Hispanic race with uninsured status and lower annual income, we repeated our multivariate analysis with and without each variable. Similarly, and given the correlation between hospital type and facility, we repeated all multivariate analyses separately using hospital and volume. All analyses were performed using SPSS statistical software (SPSS for Windows, version 15.0; SPSS Inc, Chicago, Ill).

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

During the period studied, we identified 106,002 patients with gastric adenocarcinoma in the NCDB. Of those, 75,431 patients (71.6%) were white, 15,287 patients (14.4%) were African-American, 8669 patients (8.2%) were Hispanic, and 6615 patients (6.2%) were Asian.

When patient-related, tumor-related, treatment-related, and facility-related characteristics were compared by race, several significant differences were identified (Table 1). African Americans and Hispanics were more likely to be uninsured, to have lower annual household incomes, and to present with AJCC stage III and IV disease. However, they also were more likely to receive care at a comprehensive cancer center or a teaching center and at high-volume facilities.

Table 1. Patient Demographics, Tumor-Related, Treatment-Related, and Facility-Related Characteristics With Gastric Adenocarcinoma by Race (Total Cohort = 106,002)
FactorPercentage of PatientsPa
Whites, n = 75,431 (71.6%)African-Americans, n = 15,287 (14.4%)Hispanics, n = 8669 (8.2%)Asians/PI, n = 6615 (6.2%)
  • PI indicates Pacific Islander; NOS, not otherwise specified; AJCC, American Joint Committee on Cancer; Mtn, Mountain.

  • a

    Results were based on 2-sided tests with a significance level of P = .05. Tests were adjusted for all pairwise comparisons using the Bonferroni correction.

Age, y    <.0001
 <50813.921.315.7 
 50-6935.338.340.339 
 ≥7056.747.838.445.3 
Sex    <.0001
 Men64.658.35957.8 
 Women35.441.74142.2 
Year of diagnosis    <.0001
 199813.31311.211.5 
 199913.513.112.411.9 
 200013.512.81313.2 
 200111.611.811.411.7 
 200212.312.412.512.8 
 200312.312.713.112.6 
 200412.212.113.913.3 
 200511.412.112.512.9 
Insurance status    <.0001
 Insured94.289.882.688.5 
 Uninsured1.95.910.96.5 
 Unknown44.36.55 
Median income quartile, 2000    <.0001
 <$30,00011.641.927.89.9 
 $30,000-$35,99918.820.920.910.6 
 $36,000-$45,99929.12126.924.5 
 ≥$46,00040.516.324.355 
Location of tumor in stomach    <.0001
 Proximal42.614.818.613.6 
 Body6.98.910.49 
 Distal16.12924.232.1 
 Lesser and greater curvature10.41515.118.2 
 Overlapping lesion6.47.59.39.7 
 Stomach, NOS17.724.722.417.3 
Histologic grade    <.0001
 Low grade (1/2)29.230.222.325.9 
 High grade (3/4)54.553.361.962 
 Unknown16.316.515.812.2 
AJCC stage at diagnosis    <.0001
 I17.717.715.622.9 
 II11.410.51112 
 III17.516.31818 
 IV34.236.239.532.7 
 Unknown19.219.315.914.4 
Treating facility    <.0001
 Community cancer center17.315.818.713.7 
 Comprehensive community cancer center47.136.440.543.8 
 Teaching/research hospital35.547.840.842.5 
Facility case volume, percentile    <.0001
 75th-100th56.759.265.675.4 
 50th-<75th24.923.322.515.3 
 >25th-<50th14.114.297.5 
 0-25th4.43.32.91.9 
US Census region    <.0001
 Northeast and Atlantic28.719.918.119.2 
 Southeast and South23.642.613.410.4 
 Great Lakes and Midwest26.419.37.99.1 
 West, Mtn, and Pacific Coast21.418.260.661.2 
Adequacy of lymphadenectomy in those who underwent surgery for gastric cancer: No. of lymph nodes evaluated<.0001
  ≥1527.630.433.345.2 
  1-1461.661.658.349.6 
  None888.45.1 
Type of treatments received in entire cohort    <.0001
 No therapy26.131.827.221.3 
 Chemotherapy or radiotherapy only13.512.414.710.8 
 Chemoradiotherapy only8.24.24.43.1 
 Surgery alone31.932.230.738.9 
 Surgery and single modality6.87.69.39.5 
 Surgery and multimodality therapy12.911.312.815.9 
 Other0.60.60.80.5 

In our cohort, 52,731 of 106,002 patients underwent gastrectomy. Of the patients who underwent gastrectomy, <30% received adequate lymphadenectomy (≥15 lymph nodes evaluated) and adjuvant multimodality therapy for AJCC stage IB through IV M0 disease. The delivery of multimodality therapy differed significantly by AJCC stage and by type of lymph node evaluation (Table 2). Although African-American and Hispanic patients were more likely to receive adequate lymphadenectomy (≥15 lymph nodes) and care at comprehensive cancer centers and high-volume facilities (P ≤ .001 for all), they were less likely to receive adjuvant multimodality therapy for AJCC stage IB through IV M0 disease.

Table 2. Types of Adjuvant Therapy by Lymph Node Evaluation and American Joint Committee on Cancer Stage After Gastrectomy for Gastric Cancer (n = 52,731)
VariableType of Treatment, %
Surgery AloneSurgery and Single-Modality Therapy: Chemotherapy or RadiotherapySurgery Followed by Multimodality Therapy: Combined Chemoradiotherapy
  • AJCC indicates American Joint Committee on Cancer.

  • a

    Compared with surgery alone. Results were based on 2-sided tests with a significance level of P = .05. Tests were adjusted for all pairwise comparisons using the Bonferroni correction.

Type of lymph node evaluation after gastrectomy: No. of lymph nodes evaluated   
  None10.286.9
  1-1463.15758
  ≥1526.735a35.1a
AJCC stage eligible to receive adjuvant multimodality therapy   
  I/IB74.912.4a12.7a
  II44.927a28.1a
  III5.646.9a47.5a
  IV M074.613.7a11.7a

Multivariate Analyses of Race and Lymph Node Evaluation After Gastrectomy

In our multivariate model, nonwhite patients had a lower risk of receiving inadequate lymph node evaluation (<15 lymph nodes). For example, Asian race predicted a nearly 50% reduction in the risk of inadequate lymphadenectomy compared with the risk among whites (for Asians, the odds ratio [OR] was 0.54; 95% confidence interval [CI], 0.50-0.58; P < .0001). Factors that were significantly predictive of inadequate lymphadenectomy also included advanced age, being a man, diagnosis before 2000, US Census region, primary tumor location, and treatment at a low-volume facility (Table 3). Furthermore, nonwhites were less likely to be at risk of having no lymph node evaluation after gastrectomy (Table 3). Because of the correlation between African-American and Hispanic race with uninsured status and lower annual income, we repeated our multivariate analysis after each variable, and the results continued to indicate that our race estimates remained unchanged (data not shown).

Table 3. Predictors of Lymph Node Evaluation After Gastrectomy (n = 52,731)
VariablePredictors of Inadequate Lymphadenectomy (<15 Lymph Nodes)Predictors of No Lymph Node Evaluation
OR95% CIPOR95% CIP
  1. OR indicates odds ratio; CI, confidence interval; AA, African American; PI, Pacific Islander; Mtn, Mountain; NOS, not otherwise specified.

Race      
 WhiteReferent  Referent  
 AA0.880.83-0.93<.00010.750.68-0.83<.0001
 Hispanic0.870.81-0.93<.00010.880.78-0.99.033
 Asian/PI0.540.50-0.58<.00010.550.48-0.64<.0001
Age, y      
 <50Referent  Referent  
 50-691.121.05-1.20<.00011.050.94-1.17.435
 ≥701.501.41-1.60<.00011.090.99-1.22.090
Sex      
 MenReferent  Referent  
 Women0.950.91-0.98.0061.050.99-1.12.128
Year of diagnosis      
 19981.201.11-1.30<.00011.151.02-1.29.019
 19991.201.10-1.29<.00011.030.92-1.17.582
 20001.111.02-1.20.011.000.89-1.13.992
 20011.060.98-1.14.1800.990.87-1.12.842
 2002Referent  Referent  
 20030.890.83-0.97.0040.940.83-1.06.307
 20040.860.80-0.93<.00011.010.89-1.14.918
 20050.770.71-0.83<.00010.910.80-1.03.142
Region      
 Northeast and AtlanticReferent  Referent  
 South and Southeast1.311.24-1.38<.00011.201.11-1.31<.0001
 Midwest and Great Lakes1.111.05-1.18<.00010.970.88-1.06.435
 West, Mtn, and Pacific Coast1.050.99-1.11.0851.030.94-1.13.518
Location of primary tumor      
 ProximalReferent  Referent  
 Body0.850.78-0.92<.00011.130.99-1.26.060
 Distal1.111.05-1.17<.00010.940.86-1.02.143
 Curvatures0.840.79-0.90<.00010.880.79-0.97.013
 Overlapping lesions0.670.62-0.73<.00010.780.68-0.90.001
 Stomach, NOS1.121.05-1.19.0011.601.46-1.74<.0001
Facility case volume, percentile      
 75th-100thReferent  Referent  
 50th-75th1.361.29-1.43<.00011.281.19-1.38<.0001
 26th-49th1.431.34-1.53<.00011.441.32-1.58<.0001
 0-25th1.511.35-1.69<.00011.651.43-1.91<.0001

Multivariate Analyses of Race and Lack of Adjuvant Multimodality Therapy for American Joint Committee on Cancer Stage IB Through IV M0 Disease

African-American race and Hispanic race both predicted the lack of adjuvant multimodality therapy for AJCC stage IB through IV M0 disease (African-American race: OR, 1.12; 95%CI, 1.04-1.12; P = .003; Hispanic race: OR, 1.12; 95%CI, 1.01-1.23; P = .025). Other factors that were significantly predictive of a lack of adjuvant multimodality therapy were older age, being a woman, diagnosis before 2002, US Census region, nonproximal gastric tumors, and AJCC stage (Table 4). Because African-American and Hispanic race were correlated with uninsured status and lower annual income, as discussed above, we repeated our multivariate analysis after adding each variable. When insurance status was added to our model, the estimates for both race and insurance became nonsignificant (African-American race: OR, 0.93; 95%CI, 0.83-1.04; P = .222; uninsured status: OR, 0.86; 95%CI, 0.70-1.04; P = .119), suggesting that the impact of race on multimodality therapy was confounded by insurance status. Therefore, we reported our results (in Table 4) with both of these variables removed.

Table 4. Predictors of Lack of Multimodality Therapy After Gastrectomy (n=52,731)
VariablePredictors of Lack of Multimodality Therapy
OR95% CIP
  1. OR indicates odds ratio; CI, confidence interval; AA, African American; PI, Pacific Islander; Mtn, Mountain; NOS, not otherwise specified; AJCC, American Joint Committee on Cancer; M0, no metastasis.

Race   
 WhiteReferent  
 AA1.121.04-1.12.003
 Hispanic1.121.01-1.23.025
 Asian/PI0.920.83-1.01.098
Age, y   
 <50Referent  
 50-691.501.39-1.63<.0001
 ≥704.634.27-5.02<.0001
Sex   
 MenReferent  
 Women1.201.13-1.26<.0001
Year of diagnosis   
 19982.832.56-3.14<.0001
 19992.612.35-2.88<.0001
 20001.241.13-1.36<.0001
 20010.990.90-1.09.803
 2002Referent  
 20030.910.83-0.99.037
 20040.910.82-0.99.046
 20050.920.84-1.02.107
Region   
 Northeast and AtlanticReferent  
 South and Southeast1.161.08-1.24<.0001
 Midwest and Great Lakes0.970.90-1.04.375
 West, Mtn, and Pacific Coast1.251.16-1.34<.0001
Location of primary tumor   
 ProximalReferent  
 Body1.401.25-1.56<.0001
 Distal1.251.16-1.34<.0001
 Curvatures1.151.06-1.24<.0001
 Overlapping lesions1.311.18-1.46<.0001
 Stomach, NOS1.361.25-1.47<.0001
Histologic grade   
 Low gradeReferent  
 High grade0.850.80-0.91<.0001
AJCC stage   
 IBReferent  
 II0.450.41-0.49<.0001
 III0.370.35-0.40<.0001
 IV M00.600.54-0.66<.0001
 Unstaged0.350.29-0.42<.0001
Facility case volume, percentile   
 75th-100thReferent  
 50th-75th0.920.86-0.97.005
 26th-49th1.060.97-1.14.188
 0-25th1.130.98-1.31.096

Multivariate Analyses of Race and Overall Survival

White race predicted poorer overall survival outcomes compared with other racial groups. Additional factors that were significantly predictive of worse overall survival rates in the entire cohort were advanced age, being a man, proximal tumors, high tumor grade, advanced AJCC stage, low facility volume, and receiving no adjuvant multimodality therapy. In patients who underwent gastric cancer surgery, inadequate lymph node evaluation and lack of adjuvant multimodality therapy were additional independent predictors of worse survival outcomes (Table 5).

Table 5. Predictors of Survival in Patients With gastric Adenocarcinoma: 1998 to 2000
VariableEntire Cohort, n = 41,924After Gastric Cancer Surgery, n = 21,112
HR95% CIPHR95% CIP
  1. HR indicates hazard ratio; CI, confidence interval; AA, African American; PI, Pacific Islander; NOS, not otherwise specified; Mtn, Mountain; LNs, lymph nodes; AJCC, American Joint Committee on Cancer; M0, no metastasis.

Race      
 WhiteReferent  Referent  
 AA1.000.97-1.04.6961.000.95-1.00.91
 Hispanic0.880.84-0.93<.00010.870.81-0.93<.001
 Asian/PI0.760.72-0.81<.00010.780.72-0.84<.0001
Age, y      
 <50Referent  Referent  
 50-701.111.06-1.56<.00011.111.04-1.18.002
 >711.481.42-1.54<.00011.541.45-1.64<.0001
Sex      
 MenReferent  Referent  
 Women0.900.88-0.92<.00010.930.89-0.96<.0001
Year of diagnosis      
 1998Referent  Referent  
 19990.990.96-1.02.5151.040.99-1.08.099
 20000.980.95-1.01.1291.010.97-10.5.524
Region      
 Northeast and AtlanticReferent  Referent  
 South and Southeast1.161.08-1.24<.00011.030.98-1.08.323
 Midwest and Great Lakes0.970.90-1.04.1420.990.94-1.04.706
 West, Mtn, and Pacific Coast1.251.16-1.34.0011.051.00-1.10.048
Tumor location      
 ProximalReferent  Referent  
 Body0.910.87-0.96<.00010.860.74-0.86<.0001
 Distal0.980.95-1.02.1010.870.79-0.87<.0001
 Curvatures0.850.82-0.89<.0010.720.68-0.76<.0001
 Overlapping lesions1.151.10-1.20<.00010.970.90-1.04.390
 Stomach, NOS1.030.99-1.06.4990.900.85-0.95<.0001
Histologic grade      
 Low gradeReferent  Referent  
 High grade1.311.28-1.35<.00011.311.26-1.36<.0001
AJCC stage      
 IReferent  Referent  
 II1.611.50-1.72<.00011.991.87-2.12<.0001
 III2.582.42-2.74<.00013.503.30-3.70<.0001
 IV M04.063.82-4.32<.00016.365.99-6.75<.0001
 Unstaged2.372.32-2.53<.00012.081.92-2.25<.0001
No. of LNs evaluated      
 ≥15Referent  Referent  
 1-141.210.76-0.82<.00011.291.24-1.35<.0001
 None1.030.99-1.08.1541.481.37-1.59<.0001
Hospital volume, percentile      
 75th-100thReferent  Referent  
 50th-75th1.031.01-1.06.0871.040.99-1.09.057
 26th-49th1.091.06-1.14<.00011.131.07-1.19<.0001
 0-25th1.111.07-1.19<.00011.181.07-1.29<.0001
Types of treatments received      
 No therapy2.702.60-2.81<.0001   
 Chemotherapy or radiotherapy2.222.13-2.32<.0001   
 Chemoradiotherapy1.581.50-1.67<.0001   
 Surgery aloneReferent  Referent  
 Surgery plus single modality0.880.84-0.92<.00010.810.77-0.85<.0001
 Surgery plus multimodality therapy0.740.71-0.77<.00010.690.66-0.72<.0001

Given the correlation between facility type and facility volume, we repeated all of our multivariate models by separately using facility type or facility volume, which continued to indicate that our race estimates, along with other significant factors (in Tables 3-5), remained unchanged (data not shown).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

In this large, multihospital study from the NCDB, our results demonstrated the presence of significant treatment differences among racial groups with gastric cancer in the United States. Although African American and Hispanics were likely to receive adequate lymph node evaluation, they were less likely to receive adjuvant multimodality therapy. It is noteworthy that, according to our current results, gastric cancer treatment overall deviates from current recommended guidelines in the United States. Our findings suggest patterns of under treatment that extend beyond racial disparities to a wider national problem. In our study, nearly 66% of patients failed to receive adequate lymph node examination or adjuvant multimodality therapy for AJCC stage IB through IV M0 disease. To our knowledge, this is the largest multihospital study to present a global overview of gastric cancer care in the United States.

In striking contrast to studies that demonstrated an association between nonwhite race and inferior cancer care and, ultimately, unfavorable long-term survival,14-21 we observed that nonwhite race predicted improved surgical care, as defined by adequate lymph node evaluation. Because African Americans and Hispanics tend to be uninsured and of lower socioeconomic status, it is possible that university centers are the likely providers of their surgical care: These findings are contrary to popular opinion. Our observations and those of others also suggest that academic and comprehensive cancer centers are associated with higher case volumes.22-30 Whether this particular association between nonwhite race and better surgical care exists in other malignancies deserves further investigation.

Consistent with numerous studies of racial disparities in cancer care in the United States, we also observed that African-American and Hispanic races were associated with a failure to receive adjuvant multimodality therapy. The strong association between African-American and Hispanic race and lack of insurance and lower socioeconomic status is a likely barrier to accessing cancer programs and multidisciplinary cancer care. However, in the current study, we observed that the effect of race on receipt of multimodality therapy was confounded by uninsured status.31, 32 To allow for future interventions, mechanisms to better understand the relation between hospital volume and insurance status with the receipt of multimodality therapy are needed.

Our study should be interpreted with the following limitations. First, the NCDB only receives reports from CoC-approved programs, which capture 75% of cancer cases in the United States. We could not account for non-CoC programs that, if included, may have altered our observations. Second, it has been demonstrated that lack of health insurance is a barrier to receiving medical care in a timely manner. Because >90% of patients in the NCDB were insured and <4% were either uninsured or had unknown insurance status, it is possible that our study may have underestimated the effect of being uninsured (as a factor) on the delivery of optimum gastric cancer care in the NCDB. Therefore, our multivariate analyses were repeated with and without insurance status. Given the finding that the NCDB does not collect this information, the current study does not assess variations in the receipt of neoadjuvant therapy for gastric cancer or referral patterns to a medical/radiation oncologist by race. Finally, comorbidities are natural barriers to the receipt of cancer care. For example, African-American race is associated with high rates of comorbidities.33 It is possible that the delivery of adjuvant multimodality therapy to African Americans is confounded by comorbidities.

Nevertheless, our study presents important implications for assessing the current status of US cancer care and opening avenues for future investigations to improve the quality of multidisciplinary cancer care. First, the paradoxical disparities observed in our study between nonwhite race and improved surgical care are worthy of note and future evaluation. A plausible explanation is that university centers are the likely providers for patients with inadequate or no health insurance. To our knowledge, studies evaluating the association between types of health insurance and referrals to university and comprehensive cancer centers are sparse. Second, it has been demonstrated that higher facility volumes are predictive of improved surgical care and favorable long-term survival.34 However, this was not the case for patients who were enrolled in the Intergroup 0116 trial of adjuvant chemoradiotherapy for resectable esophagogastric adenocarcinoma.35 In that study, investigators observed no association between hospital procedure volume and tumor recurrence or survival after gastric cancer surgery.6, 35 We observed that high facility volume (or facility type) had a marginal effect on whether multimodality therapy was delivered to patients with gastric cancer. Because of the demanding financial burden of adjuvant multimodality therapy, uninsured or underinsured patients may receive their surgical care at high-volume centers (or university centers in the current study) but may not necessarily receive their adjuvant multimodality treatments afterward. In the era of multidisciplinary cancer care, future studies should investigate any underlying variation between facility type or volume and delivery of the entire process of cancer care. Third, although racial disparities in cancer care have been evaluated extensively, our current study provides insights into potential variations between US regions and the delivery of optimum care. Given the limitations of the literature in these areas, exploring this line of health services research may offer opportunities to improve regional performance. Fourth, our study demonstrated that nonwhites were more likely to receive adequate lymph node evaluation. Our multivariate analyses also demonstrated that the receipt of adjuvant therapy predicted favorable survival outcomes after adjusting for race and adequate lymph node evaluation. In these models, combined multimodality therapy after gastrectomy had a more protective effect on survival (hazard ratio, 0.81) than single multimodality therapy (hazard ratio, 0.69). Whether or not the observed benefit in survival for patients with adequately staged gastric cancer after receiving adjuvant therapy varies by race deserves the attention of a future study. Finally, our observations, and those of others, suggest a concerning overview of gastric cancer care in the United States36 and may represent a marker of suboptimal overall cancer care in this country. Our findings also highlight the importance of measuring compliance with recommended care. The frequency of evaluating 12 lymph nodes was investigated recently in the NCDB as a colon cancer quality measure, and it was reported that 60% of CoC programs failed to meet this standard from 2004 to 2005.37 Similar quality measures should be considered to allow treating facilities to assess their compliance with recommended evidence-based guidelines and to maximize the delivery of optimum cancer care in United States.

In conclusion, in this large, multihospital study from the NCDB, we observed that there were significant gastric cancer treatment variations among racial groups in the United States. Our findings also suggested that adequate lymphadenectomy and multimodality therapy for gastric cancer are underused in the United States. Whereas nonwhite race was associated with improved surgical care, overall, gastric cancer care deviated from currently recommended guidelines. Cancer programs need to identify procedures to maximize the delivery of multimodality therapy and surgical care to all patients. Future studies are needed to investigate underlying associations between hospital attributes and the administration of adjuvant therapy after gastric cancer surgery and to establish quality measures for gastric cancer care in the United States.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

We thank Ms. Virginia Mohlere for her expert editorial assistance.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES