SEARCH

SEARCH BY CITATION

Keywords:

  • breast neoplasms;
  • outcome assessment;
  • health insurance;
  • health services accessibility;
  • medically uninsured;
  • Medicaid;
  • neoplasm staging;
  • data bases;
  • retrospective studies

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

BACKGROUND

Individuals without medical insurance or with limited insurance are less likely than those with broader insurance coverage to receive preventive services and to seek timely medical care. The authors examined the associations of insurance status with stage at diagnosis among women with breast cancer.

METHODS

This study included women age ≥40 years who were diagnosed with invasive breast cancer from 1998 to 2003 and who were reported to the National Cancer Data Base. Multivariable logistic regression analyses were used to examine the associations of insurance status with more advanced-stage breast cancer at diagnosis while controlling for other patient characteristics.

RESULTS

Among the 533,715 women with breast cancer who were included in the current analysis, the proportions with advanced-stage (III/IV) cancer at diagnosis ranged from 8% among privately insured patients to 18% among uninsured patients and 19% among Medicaid patients; differences in the proportions of women with advanced-stage cancer were statistically significant (P < .0001). Regression analyses indicated that, compared with privately insured patients, uninsured patients and Medicaid patients had a greater likelihood of diagnosis at stage II (odds ratio [OR], ∼≈1.5) or at stages III/IV (OR, 2.4) versus stage I (P < .001). Black and Hispanic patients also were significantly more likely than white patients to be diagnosed at a more advanced stage (P < .001).

CONCLUSIONS

The results from this study provided strong evidence that patients without health insurance or with Medicaid coverage, as well as black and Hispanic patients, were more likely to present with advanced-stage breast cancer. These results are consistent with other reports that have documented less use of preventive services, including mammography, among uninsured women and delays in diagnosis and treatment for black and Hispanic women. Cancer 2007. © 2007 American Cancer Society.

More than 45 million Americans aged <65 years (or 20% of the nonelderly) lack health insurance, and many more who are underinsured (eg, intermittent lack of insurance coverage or enrollment in plans with limited benefits) do not have adequate access to health care.1 Lack of health insurance may have serious adverse consequences.2 Studies have shown that uninsured adults were less likely than the insured to receive preventive care, seek care in a timely manner, or receive recommended treatment.3–7

Breast cancer is the most common nonskin cancer and the secondleading cause of cancer-related mortality among women. Early detection and appropriate, timely treatments improve patient outcomes.8–10 Multiple studies have suggested that being uninsured or inadequately insured, which potentially is the case with certain types of public health insurance, may be barriers to early breast cancer detection and receipt of high-quality care.11–17 However, most of those studies involved data from a single state and/or years prior to 1998. Therefore, more current and generalizable information is needed about the impact of health insurance on the early detection of breast cancer.

In this study, we examined the associations between insurance status and stage of cancer at diagnosis among women who were diagnosed with breast cancer from 1998 to 2003 and who were reported to the National Cancer Data Base (NCDB). We compared the disease stage at diagnosis for uninsured women or women who were enrolled in public insurance plans (Medicaid, Medicare) with the disease stage at diagnosis among women who were enrolled in private health insurance plans.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

NCDB: 1998 to 2003

The NCDB, jointly sponsored by the American College of Surgeons and the American Cancer Society, is a national hospital-based cancer registry. The NCDB includes approximately 75% of cancer cases in the United States and collects data from >1400 hospitalsthat have Commission on Cancer (CoC)-approved cancer-treatment programs.18 The NCDB contains standardized data elements on patient demographics, patient insurance status, tumor site, stage and morphology, and first course of treatment in common with population-based registries. In addition, the NCDB contains information on patient zip code and county of residence, which is useful for incorporating area-based sociodemographic characteristics. Evaluations in progress suggest there are substantial similarities between CoC hospitals and non-CoC hospitals that have oncology programs.

Patient Sample

A data file for patients who were diagnosed with breast cancer (International Classification of Diseases for Oncology codes C50.0-C50.9) from 1998 to 2003 was prepared from the NCDB. The analysis included women aged ≥40 years who were diagnosed with unilateral stage I through IV breast cancer between 1998 and 2003. The sample was restricted further to include only those who were categorized as NCDB class 1 or 2 patients. Class 1 patients were those who were diagnosed at the reporting institution and received all or part of their first course of treatment at that facility. Class 2 patients were diagnosed elsewhere but received part or all of their treatment at the reporting facility. There were 712,171 patients who met these initial criteria.

In total, 178,356 patients were excluded from the analysis, including 24,585 patients who were excluded because of missing information on health insurance or because they had types of government health insurance that were not included in the analysis (Indian Health Service, Public Health Service, other state-sponsored programs, and Veterans Affairs), 27,111 patients who were excluded because they had stage 0 disease (ie, in situ) or bilateral disease, 50,037 patients who were excluded because of missing information on race, and 58,881 patients who were excluded because of missing zip code-level education and/or income data. Finally, 17,922 patients were excluded because they were treated at facilities other than the 3 main types of CoC hospitals (described below). The final sample included 533,715 patients.

Variables

The dependent variable of interest was cancer stage at diagnosis. Two separate comparisons of cancer stage at diagnosis were performed: Patients who had stage I disease were compared with patients who had stage II disease, and patients who had stage I disease were compared with patients who had stage III or IV disease. Cancer stage was determined initially based on clinical values for tumor size, lymph node status, and metastases (TNM). If clinical information for T, N, or M were missing, then pathologic stage information was used instead.

The key explanatory variable was patient insurance status. Primary payer/insurance type at diagnosis was determined by using codes defined by the Facility Oncology Registry Data Standards (FORDS), which were grouped into the following categories: Medicaid, Medicare (which included both Medicare alone and Medicare with supplement), uninsured (which included FORDS codes for not insured-not otherwise specified [NOS], not insured-charity write-off, and not insured-self-pay), and private insurance plans (health maintenance organization, preferred provider organization, managed care-NOS, private insurance, the 3-option Civilian Health and Medical Program of the Uniformed Services [CHAMPUS/TRICARE], military, and insured-NOS). Individuals with any of these private insurance plans were grouped into a single category, because these plans represent either privately purchased insurance (purchased by the individual, a family member, and/or an employer) or insurance provided by the military that functions in a similar manner to private insurance (CHAMPUS/TRICARE). Because Medicare eligibility for individuals aged <65 years differs from those aged ≥65 years, the Medicare category was dichotomized by age into patients ages 40 years to 64 years versus patients aged ≥65 years.

Other sociodemographic variables that were included in the model were race, age, and zip code-based income and education. Race was categorized as white, black, Hispanic, and others (which included Asian American, American Indian, and other). Based on an evaluation of the unadjusted rate of advanced-stage disease by age among this population, age was classified into 3 overlapping splines (ie, linear functions) for individuals ages 40 years to 60 years, ages 60 years to 80 years, or aged ≥80 years. The median household income and the percentage of the population with a high school diploma for each patient's zip code of residence were included as proxies for the patient's socioeconomic status, because individual-level information on income and education was not available. Each of these variables was categorized based on national quartiles from the 2000 United States Census. The categories for the proportion of individuals with a high school diploma were <14%, from 14% to 19.9%, from 20% to 28.9%, and ≥29%; median household income was categorized as <$30,000, from $30,000 to $34,999, from $35,000 to $45,999, and ≥$46,000.

Other factors that were included in the model were state of residence, year of diagnosis, and treatment facility type. State of residence was included in the model to control for potential biases from state-level confounders, such as variation in cancer stage, percentage of breast cancer patients treated at CoC-approved facilities, and Medicaid eligibility policies and reimbursement rates.19, 20 Three types of treatment facilities that were included in the analysis were identified by using the NCDB CoC hospital category variable: community hospitals, communitycancer centers, and teaching/research hospitals. Community hospitals treat ≥300 cancer cases a year and have a full range of services for cancer care, but patients must be referred elsewhere for certain aspects of their treatment. Community cancer centers are facilities that offer the same range of services as the community hospitals but have ≥750 annual cancer cases and conduct weekly cancer conferences. Teaching/research facilities differ from community cancer facilities in that the teaching/research facilities have residency programs and ongoing cancer research. Patients who were treated at other types of facilities were excluded from the analysis.

Multivariate Statistical Analysis

Chi-square tests were used to examine differences between categorical variables by insurance status, with α = .05 for each comparison. Analyses with P values between .05 and .01 were considered marginally statistically significant, whereas analyses with P < .01 were considered significant. In addition, we used multivariate logistic regression models to assess the association of insurance coverage with the likelihood of later stage of breast cancer at diagnosis, controlling for patients' demographic, socioeconomic, and hospital characteristics. Using logistic regression analyses, patients who were diagnosed with stage I disease were compared with 2 groups: patients with stage II disease and patients with either stage III or stage IV disease. Patients with stage II disease were evaluated separately from those with stage I disease, because mortality rates differ significantly between these 2 groups. Patients with stage III and IV disease were combined into 1 group because of the smaller numbers of patients within each of these stages.

To evaluate further the potential combined effects of insurance status with age or race on the likelihood of having advanced disease at diagnosis, we also examined interaction terms of age and insurance or race and insurance in separate multivariate logistic regressions. All statistical tests were 2-sided. Standard errors were adjusted for Huber standard errors correction.21 We used Stata software (version 8.0) for all regression analyses.22

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Descriptive Statistics of the Sample

The sample included a total of 533,715 patients from 1457 hospitals in all 50 states and the District of Columbia. The descriptive statistics are shown in Table 1. Among these patients, ∼≈55% were privately insured, 37% were aged ≥65 years and had Medicare as their primary payer, 2% were aged <65 years and had with Medicare as their primary payer, 2% were uninsured, and 3% were enrolled in Medicaid. Overall, 55% of patients had stage I disease at diagnosis, 36% had stage II disease, and 9% had stage III or IV disease. The proportions with advanced stage cancer at diagnosis (stage III or IV) ranged from 8% among privately insured patients to 18% among uninsured patients and 19% among patients with Medicaid. Differences in the proportions of advanced-stage cancer by insurance type were statistically significant (P < .0001).

Table 1. Summary Statistics of Women Age ≥40 Years With Breast Cancer From 1998 to 2003 in the National Cancer Data Base
Payer statusInsurance status, Proportion of patients in each categoryOverall
Private*UninsuredMedicaidMedicare
Age <65 yAge ≥65 y
  • *

    Includes managed care, other insurance carriers-not otherwise specified, TRICARE/CHAMPUS, military (Facility Oncology Registry Data Standards; codes 10, 20, 53, and 54). Excludes patients with Veterans Affairs, Indian Health Service, or Public Health Service listed as the primary payer.

Sample size, no. of patients293,72411,81018,15712,997197,027533,715
Individual-level characteristics
 Disease stage at diagnosis
  Stage I0.540.380.380.490.590.55
  Stage II0.370.450.430.390.320.36
  Stage III0.070.140.150.090.070.08
  Stage IV0.010.040.040.020.020.02
 Age at diagnosis, y55.6954.4055.2556.0975.6663.03
  40–640.690.710.680.610.000.43
  65–740.290.270.290.390.720.45
  ≥750.020.010.030.000.280.12
 Race
  White0.840.600.560.750.890.85
  Black0.090.210.240.190.070.09
  Hispanic0.030.130.130.040.020.03
  Other race/ethnicity0.040.060.070.020.010.03
Resident zip code-level characteristics
 Proportion of population with a high school diploma, %
  <140.110.270.300.230.130.13
  14–19.90.190.260.270.260.220.20
  20–28.90.220.180.200.210.240.22
  ≥290.480.290.230.300.410.44
 Median household income, 2000 US dollars
  ≤$30,0000.070.200.210.160.100.09
  $30,000–34,9990.140.210.220.220.180.16
  $35,000–45,9990.250.270.280.280.280.26
  ≥$46,0000.540.320.290.340.440.49
Hospital-level characteristics
 Type of cancer center
  Community cancer center0.170.170.200.210.220.19
  Comprehensive community cancer center0.490.330.370.460.520.49
  Teaching/research cancer center0.340.500.430.330.260.32
 Year
  19980.170.190.150.170.180.18
  19990.170.180.140.160.170.17
  20000.170.170.150.160.170.17
  20010.180.170.170.180.180.18
  20020.150.140.180.160.150.15
  20030.160.150.210.160.150.16

Substantial differences by insurance type also were observed for a number of sociodemographic characteristics. There were significantly (P < .0001) greater proportions of nonwhite patients among the uninsured and Medicaid populations (40% and 44%, respectively) than among the private insurance and Medicare populations aged ≥65 year (16% and 11%, respectively). Zip code-level variables also indicated that uninsured patients, Medicaid patients, and patients aged <65 years who had Medicare were more likely to reside in areas with lower median incomes and less education than were patients who had private insurance (each P < .0001). Uninsured and Medicaid patients were less likely to be treated at comprehensive community cancer centers and were more likely to be treated at teaching/research hospitals than patients who had private insurance or Medicare.

Estimates From Multivariate Logistic Models

Table 2 presents the odd ratios (ORs) from the multivariable logistic regression models. Results from 2 separate models are presented in the table: a comparison of patients with stage I versus stage II disease and a comparison of patients with stage I disease versus stage III or IV disease. The direction of the results is the same for both models, although the magnitude of the ORs tends to be greater for the comparison with more advanced disease (stage III or IV). The multivariable analyses indicated that uninsured and Medicaid patients were substantially more likely to be diagnosed at a later stage of cancer than privately insured patients (both P < .001). Medicaid and uninsured patients were ∼≈50% more likely to be diagnosed at stage II rather than stage I compared with privately insured patients (OR, ∼≈1.5), and they were from 2.4 to 2.5 times more likely to be diagnosed at stage III or IV rather than at stage I compared with privately insured patients. Medicare patients aged <65 years were marginally more likely than privately insured patients to be diagnosed at stage II and were significantly more likely to be diagnosed at stage III or IV compared with privately insured patients. Medicare patients aged ≥65 years were significantly less likely to be diagnosed at stage II than at stage I compared with privately insured patients, whereas there was no statistically significant difference between these older Medicare patients and privately insured patients with regard to being diagnosed with stage III or IV disease.

Table 2. Estimated Odds Ratios From Multivariate Logistic Regression Among Women With Breast Cancer From 1998 to 2003 in the National Cancer Data Base*
CharacteristicStage II vs Stage I, n = 484,099Stage III/IV vs Stage I, n = 342,824
ORLower confidence limitUpper confidence limitORLower confidence limitUpper confidence limit
  • OR indicates odds ratio.

  • *

    ORs and 95% confidence intervals were calculated from multivariate logistic regression analyses, as described in the text (see Materials and Methods). Estimates of the 50 state dummies are not shown, and standard errors were adjusted with Huber standard errors correction.

  • Age was included in the models as 3 separate splines, as described in the text (see Materials and Methods).

Insurance coverage
Privately insured (reference)1.000  1.000  
 Uninsured1.5161.4551.5802.4172.2862.556
 Medicaid1.4711.4221.5222.4822.3732.597
 Medicare
  Age <65 y1.0811.0411.1231.3111.2361.390
  Age ≥65 y0.9080.8920.9240.9850.9561.014
Age, y
 40–591.0021.0001.0031.0041.0021.006
 60–790.9820.9810.9840.9830.9810.985
 ≥801.0151.0121.0181.0461.0411.051
Race
 White (reference)1.000  1.000  
 Black1.5161.4841.5491.8531.7951.913
 Hispanic1.3281.2841.3741.3601.2901.433
 Other race/ethnicity1.2281.1831.2731.1531.0831.227
Proportion of population with a high school diploma, %
 <14 (Reference)1.000  1.000  
 14–19.90.9360.9150.9580.9050.8740.931
 20–28.90.9100.8870.9340.8180.7870.847
 ≥290.8480.8260.8710.7310.7010.759
Median household income, 2000 US dollars
 <$30,000 (Reference)1.000  1.000  
 $30,000–34,9991.0010.9751.0270.9710.9341.023
 $35,000–45,9991.0200.9931.0480.9930.9541.037
 ≥46,0001.0331.0031.0640.9930.9491.052
Type of cancer center
 Teaching/research cancer center (reference)1.000  1.000  
 Community cancer center1.0481.0301.0671.1431.1121.198
 Comprehensive community cancer center0.9870.9731.0010.9910.9691.028
Year
 1998 (Reference)1.000  1.000  
 19991.0110.9911.0310.9530.9240.996
 20001.0190.9991.0390.9360.9060.977
 20011.0120.9931.0320.9260.8970.970
 20020.9930.9731.0130.9670.9350.991
 20030.8920.8740.9111.2851.2461.334

The regression analyses also evaluated the impact of age and race/ethnicity on the likelihood of being diagnosed with later stage disease. Age at diagnosis, as discussed above (see Materials and Methods), was included in the regression models using 3 separate linear functions: patient age at diagnosis from 40 years to 59 years, from 60 years to 79 years, or ≥80 years. The ORs represent changes in the likelihood of being diagnosed with stage II breast cancer or with stage III or IV breast cancer (compared with stage I breast cancer) for each additional year of age. Among individuals ages 40 years to 59 years, although the results are statistically significant (P = .006), the magnitude of the OR indicates that increasing age had little impact on the likelihood of being diagnosed at a later stage. For individuals ages 60 years to 79 years, the risk of later-stage diagnosis decreases slightly with increasing age; whereas, from age 80 years onward, the risk increases slightly with increasing age.

Compared with white patients, black patients were substantially more likely to be diagnosed at later stages. Black patients had an ∼≈52% increase in the likelihood of being diagnosed at stage II and an 85% increase in the likelihood of being diagnosed at stage III or IV (both significant at P < .001). Hispanic patients also were more likely to be diagnosed at a more advanced stage, with ORs ranging from 1.33 to 1.36 compared with white patients (P < .001 for both). Patients who were categorized in the other race category (which included Asians and Native Americans) were more likely to be diagnosed at stage II or stages III/IV (P < .001 for both), although the magnitude of the increased likelihood was smaller than that for black or Hispanic patients.

Residence in an area with a lower percentage of high school graduates also was associated significantly with an increased likelihood of diagnosis at more advanced stages (all comparisons were significant at P < .001). In contrast, zip code-based median household income was not associated significantly with later stage of diagnosis.

There was no significant difference in stage at diagnosis between teaching/research centers and comprehensive community cancer centers when controlling for the other regression variables. In contrast, patients who were treated at community hospitals were slightly more likely than those who were treated at teaching/research hospitals to be diagnosed at a more advanced stage (P < .001).

We also examined whether the year of diagnosis was associated significantly with disease stage at diagnosis. The likelihood of being diagnosed at stage II compared with stage I did not differ significantly during the years from 1999 through 2002 compared with 1998 but was significantly lower in 2003. In contrast, the likelihood of being diagnosed at stages III/IV was significantly lower (although qualitatively similar) from 1999 through 2002 compared with 1998 but was significantly higher in 2003. The changes observed for diagnosis in 2003 may reflect temporal differences in coding of cancer stage. In 2003, a new coding system was implemented by the American Joint Committee on Cancer that affected the classification of patients into stage II versus stage III, resulting in certain patients who would have been coded as stage II being classified instead as stage III.23 Thus, the decreased odds of being diagnosed with stage II disease and the increased odds of being diagnosed with stage III/IV disease in 2003 likely reflects this secular change of stage migration.

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

The principle finding of this study on the disease stage at diagnosis among a large national population of breast cancer patients was that patients who are uninsured or have Medicaid insurance have substantially elevated risks of being diagnosed with advanced disease compared with privately insured patients. Furthermore, black women and Hispanic women have increased risk, even after controlling for insurance status. These associations remained statistically significant in analyses that controlled for age, zip code-based sociodemographic characteristics, facility type, and year of diagnosis. Our results suggest that uninsured and underinsured women are not receiving optimal screening and/or follow-up. It has been demonstrated that early detection reduces morbidity and mortality from breast cancer. Women who are diagnosed at more advanced stages of breast cancer experience lower survival, more debilitating treatment, and greater long-term treatment-related morbidity.

Our results are in agreement with and extend those from previous, smaller studies. An analysis that used data from the American College of Surgeons from 1983 to 1990 indicated that medically indigent patients were more likely to have advanced breast cancer.24 Ayanian et al. studied a cohort of 4675 women with invasive breast cancer from the 1980s and observed that the women who were uninsured or who had Medicaid insurance were more likely to have advanced disease.25 Roetzheim et al., using data from Florida, also reported that uninsured and Medicaid patients with breast cancer were more likely to be diagnosed with advanced disease compared with patients who had commercial insurance.16 Our results provide information from a large, national breast cancer population, increasing the generalizability of these results as well as controlling for potentially important confounders.

Numerous studies also have examined the relation between race/ethnicity and breast cancer stage at diagnosis; some of those studies simultaneously examined the role of insurance and socioeconomic status. Elmore et al. evaluated 100 black and 300 white breast cancer patients using data from the Yale-New Haven Hospital registry.26 Insurance and income explained some of the differences they reported between the black and white populations; however, even controlling for those factors, black patients were more likely to have larger tumors. Using data from the Surveillance, Epidemiology, and End Results registries in Detroit and Los Angeles, Lantz et al. also observed that black and Hispanic women were less likely to be diagnosed with early-stage breast cancer than white women after controlling for socioeconomic factors.27

Several factors probably contribute to the increased risk of disease among the uninsured and Medicaid populations. Uninsured individuals and those with Medicaid are less likely to have regular sources of medical care in general28, 29 and, specifically, are less likely to receive regular mammography.30–33 The age-adjusted prevalence of mammography in the past 2 years, based on the 2003 National Health Interview Survey (NHIS), was 73.1% for women with health insurance and 40.2% for women without health insurance.34 Recently published studies are inconsistent regarding whether the prevalence of recent mammography among black women differs compared with that among white women.35, 36 Data from the 2003 NHIS showed similar age-adjusted prevalence among black and white women, with slightly lower prevalence for Hispanic/Latina women.34 However, among the black population and the uninsured and Medicaid populations, individuals with abnormal screening mammograms are less likely to receive timely follow-up and are more likely to experience greater delays in diagnosis and treatment.26, 37 Delays in treatment also have been reported among Hispanic women, particularly those who prefer Spanish to English.38 One recent study reported that the lack of a usual health care provider was an even stronger predictor than the lack of health insurance for inadequate follow-up of abnormal or inconclusive mammograms and that black race/ethnicity was associated with inadequate follow-up even after controlling for other factors.39

The National Breast and Cervical Cancer Early Detection Program (NBCCEDP) was initiated in 1991 to provide early detection programs for low-income, uninsured women in all 50 states. Although the number of women served by the program grew from 5430 in 1991/1992 to over half a million in 2003, only 13.2% of the approximately 4 million women in the United States who had no health insurance and a family income <250% of the Federal Policy Level received mammograms through the NBCCEDP.34, 40

The observed association between Medicaid insurance and more advanced stage of breast cancer diagnosis does not distinguish women who became eligible for Medicaid coverage because of their diagnosis from women who were covered by Medicaid prior to diagnosis. Breast cancer patients who are uninsured at the time of initial hospitalization may receive retroactive Medicaid coverage. The passage of the Breast and Cervical Cancer Treatment Act, effective October 1, 2000, gave states the option to provide Medicaid coverage for medical assistance and follow-up treatment for women who were diagnosed with cancer through the NBCCEDP; as of 2005, all states and the District of Columbia had elected to provide this coverage. Thus, the ORS associated with Medicaid insurance may be modified by retroactive enrollment of patients who initially were uninsured. Using Medicaid data from California (Medi-Cal), Perkins et al. reported that 18% of women with breast cancer who, according to the population-based cancer registry, had Medi-Cal insurance at the time of their diagnosis were not covered by Medi-Cal during the year prior to diagnosis, and this group was substantially more likely to present with advanced disease.41 Bradley et al. reported that cancer patients who were enrolled in Medicaid at the time of or after their diagnosis were more likely to have advanced disease than patients who were enrolled prior to their diagnosis; however, this difference was not statistically significant, and both groups had significantly increased risks of advanced disease compared with non-Medicaid patients.11 Bradley et al. also reported that, although women with breast cancer who were enrolled in Medicaid either before or after diagnosis had significantly decreased survival compared with non-Medicaid patients, there was no significant difference in survival between the 2 Medicaid populations.12 No information was available from the NCDB to determine whether patients for whom Medicaid was the primary payer were covered prior to diagnosis or received coverage retroactively.

The current study also was limited regarding the sociodemographic characteristics of individual patients. No individual-level information on education or income was available in this data set; instead, we used categories of education and income based on the patients' zip codes of residence. Although these area-level variables are likely to be correlated with individual socioeconomic status and independently may influence access to care, they are less specific than individual-level data in capturing socioeconomic status.

A potential limitation of our study is that the NCDB includes only patients who are treated in CoC-approved treatment facilities. Thus, any systematic biases in referral patterns to CoC-approved facilities that relate to both insurance status and stage at diagnosis potentially may influence our findings. We hypothesized that these biases may be greatest for teaching facilities, which treat a higher proportion of uninsured and Medicaid-insured patients and also treat a high proportion of patients with more advanced-stage disease. However, no difference in the association between insurance status and more advanced stage of breast cancer was observed between teaching/research hospitals and comprehensive community cancer centers, facility types that, combined, treated ∼≈80% of breast cancer patients in the study. Moreover, prior studies have suggested that clinical characteristics of patients reported to the NCDB are similar to those in population-based registries, and a variety of facility types, including some National Cancer Institute (NCI)-designated cancer centers, choose not to participate in the CoC-approvals program. Thus, it is unlikely that the strong associations between insurance status and stage at diagnosis observed in the current study resulted from referral bias.

Despite some limitations, this study provides strong evidence that patients without health insurance and those with Medicaid insurance, as well as black and Hispanic patients (separate from any influence of insurance status), are significantly more likely to present with more advanced breast cancer. This evidence is consistent with other studies in which uninsured and underinsured women had less access to preventive services, including mammography. Programs that provide breast cancer screening to uninsured or underinsured women reportedly increase the proportion of patients who are diagnosed at earlier stages.42, 43 Future research should focus on evaluating the impact of activities to decrease these disparities, including broader insurance coverage with sufficient levels of preventive services, improved access to primary and preventive care, and intervention to facilitate rapid diagnosis, follow-up, and appropriate treatment of individuals with breast cancer.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Holahan J, Cook A. Changes in economics conditions and health insurance coverage, 2000–2004. Health Aff (Millwood). 2005; W5( suppl web exclusives): 498508.
  • 2
    Institute of Medicine. Insuring America's Health: Principles and Recommendations. Washington, DC: National Academy Press; 2004.
  • 3
    Ayanian JZ, Weissman JS, Schneider EC, Ginsburg JA, Zaslavsky AM. Unmet health needs of uninsured adults in the United States. JAMA. 2000; 284: 20612069.
  • 4
    Beckles GL, Engelgau MM, Narayan KM, Herman WH, Aubert RE, Williamson DF. Population-based assessment of the level of care among adults with diabetes in the U.S. Diabetes Care. 1998; 21: 14321438.
  • 5
    Breen N, Wagener DK, Brown ML, Davis WW, Ballard-Barbash R. Progress in cancer screening over a decade: results of cancer screening from the 1987, 1992, and 1998 National Health Interview Surveys. J Natl Cancer Inst. 2001; 93: 17041713.
  • 6
    DeVoe JE, Fryer GE, Phillips R, Green L. Receipt of preventive care among adults: insurance status and usual source of care. Am J Public Health. 2003, 93: 786791.
  • 7
    Powell-Griner E, Bolen J, Bland S. Health care coverage and use of preventive services among the near elderly in the United States. Am J Public Health. 1999; 89: 882886.
  • 8
    Berry DA, Cronin KA, Plevritis SK, et al. Cancer Intervention and Surveillance Modeling Network (CISNET) collaborators. Effect of screening and adjuvant therapy on mortality from breast cancer. N Engl J Med. 2005; 353: 17841792.
  • 9
    Christensen LH, Engholm G, Cortes R, et al. Reduced mortality for women with mammography-detected breast cancer in east Denmark and south Sweden. Eur J Cancer. 2006; 42: 27732780.
  • 10
    Cronin KA, Feuer EJ, Clarke LD, Plevritis SK. Impact of adjuvant therapy and mammography on U.S. mortality from 1975 to 2000: comparison of mortality results from the CISNET breast cancer base case analysis. J Natl Cancer Inst Monogr. 2006;( 36): 112121.
  • 11
    Bradely CJ, Given CW, Roberts C. Late stage cancers in a Medicaid-insured population. Med Care. 2003; 41: 722728.
  • 12
    Bradley CJ, Gardiner J, Given CW, Roberts C. Cancer, Medicaid enrollment, and survival disparities. Cancer. 2005; 103: 17121708.
  • 13
    Lee-Feldstein A, Feldstein PJ, Buchmueller T, Katterhagen C. The relationship of HMOs, health insurance, and delivery systems to breast cancer outcomes. Med Care. 2000; 38: 705718.
  • 14
    McDavid K, Tucker TC, Sloggett A, Coleman MP. Cancer survival in Kentucky and health insurance coverage Arch Int Med. 2003; 163: 21352144.
  • 15
    Richardson LC, Tian L, Voti L, et al. The roles of teaching hospitals, insurance status, and race/ethnicity in receipt of adjuvant therapy for regional-stage breast cancer in Florida. Am J Public Health. 2006; 96: 160166.
  • 16
    Roetzheim RG, Pal N, Tennant C, et al. Effects of health insurance and race on early detection of cancer. J Natl Cancer Inst. 1999; 91: 14091415.
  • 17
    Roetzheim RG, Gonzalez EC, Ferrante JM, Pal N, Van Durme DJ, Krischer JP. Effects of health insurance and race on breast carcinoma treatments and outcomes. Cancer. 2000; 89: 22022213.
  • 18
    American College of Surgeons. What is the NCDB? Available from URL: http://www.facs.org/cancer/ncdb/ncdbabout.html Accessed March 8, 2008.
  • 19
    Kaiser Family Foundation. State Medicaid Fact Sheet. Available from URL: http://www.kff.org/mfs/pdfs/pdf_106.pdf Accessed March 8, 2007.
  • 20
    Santerre RE. The inequity of Medicaid reimbursement in the United States. App Health Econ Health Policy. 2002; 1: 2532.
  • 21
    White H. A heteroskedasticity-consistent covariate matrix estimator and a direct test for heteroskedasticity. Econometrica. 1980; 48: 817838.
  • 22
    StataCorp. Stata Statistical Software, release 8.0. College Station, Tex: Stata Corporation; 2003.
  • 23
    Woodward WA, Strom EA, Tucker SL, McNeese MD, Perkins GH, Schecter NR, Singletary SE, Theriault RL, Hortobagyi GN, Hunt KK, Buchholz TA. Changes in the 2003 American Joint Committee on Cancer staging for breast cancer dramatically affect stage-specific survival. J Clin Oncol. 2003; 21: 32443248.
  • 24
    Osteen RT, Winchester DP, Hussey DH, et al. Insurance coverage of patients with breast cancer in the 1991 Commission on Cancer Patient Care Evaluation study. Ann Surg Oncol. 1994; 1: 462467.
  • 25
    Ayanian JZ, Kohler BA, Abe T, Epstein AM. The relation between health insurance coverage and clinical outcomes among women with breast cancer. New Engl J Med. 1993; 329: 326331.
  • 26
    Elmore JG, Nakano CY, Linden HM, Reisch LM, Ayanian JZ, Larson EB. Racial inequities in the timing of breast cancer detection, diagnosis, and initiation of treatment. Med Care. 2005; 43: 141148.
  • 27
    Lantz PM, Mujahid M, Schwartz K, et al. The influence of race, ethnicity, and individual socioeconomic factors on breast cancer stage at diagnosis. Am J Public Health. 2006; 96: 21732178.
  • 28
    Berk ML, Schur CL. Access to care: how much difference does Medicaid make? Health Aff (Millwood). 1998; 17: 169180.
  • 29
    Greenwald HP, O'Keefe S, Dicamillo M. The importance of public sector health care in an underserved population. JHealth Hum Serv Adm. 2004; 27: 142157.
  • 30
    Chen JY, Diamant A, Pourat N, Kagawa-Singer M. Racial/ethnic disparities in the use of preventive services among the elderly. Am J Prev Med. 2005; 29: 388395.
  • 31
    Coughlin SS, Uhler RJ, Bobo JK, Caplan L. Breast cancer screening practices among women in the United States, 2000. Cancer Causes Control. 2004; 15: 159170.
  • 32
    Parker J, Gebretsadik T, Sabogal F, Newman J, Lawson HW. Mammography screening among California Medicare beneficiaries: 1993–1994. Am J Prev Med. 1998; 15: 198205.
  • 33
    Sambamoorthi U, McAlpine DD. Racial, ethnic, socioeconomic, and access disparities in the use of preventive services among women. Prev Med. 2003; 37: 475484.
  • 34
    American Cancer Society. Cancer Prevention and Early Detection: Facts and Figures 2006. Atlanta, Ga: American Cancer Society; 2006.
  • 35
    Scharpf TP, Rimm AA. Mammography utilization rates among young white and black women in the USA. Public Health. 2006; 120: 937941.
  • 36
    Smith-Bindman R, Miglioretti DL, Lurie N, et al. Does utilization of screening mammography explain racial and ethnic differences in breast cancer? Ann Intern Med. 2006; 144: 541553.
  • 37
    Gwyn K, Bondy ML, Cohen DS, et al. Racial differences in diagnosis, treatment, and clinical delays in a population-based study of patients with newly diagnosed breast carcinoma. Cancer. 2004; 100: 15951604.
  • 38
    Katz SJ, Lantz PM, Paredes Y, et al. Breast cancer treatment experiences of Latinas in Los Angeles County. Am J Public Health. 2005; 95: 22252230.
  • 39
    Jones BA, Dailey A, Calvocoressi L, et al. Inadequate follow-up of abnormal screening mammograms: findings from the Race Differences in Screening Mammography Process Study (United States). Cancer Causes Control. 2005; 16: 809821.
  • 40
    Tangka FK, Dalaker J, Chattopadhyay SK, et al. Meeting the mammography screening needs of underserved women: the performance of the National Breast and Cervical Cancer Early Detection Program in 2002–2003 (United States). Cancer Causes Control. 2006; 17: 11451154.
  • 41
    Perkins CI, Wright WE, Allen M, Samuels SJ, Romano PS. Breast cancer stage at diagnosis in relation to duration of Medicaid enrollment. Med Care. 2001; 39: 12241233.
  • 42
    McCoy CB, Pereyra M, Metsch LR, Collado-Mesa F, Messiah SE, Sears S. A community-based breast cancer screening program for medically underserved women: its effect on disease stage at diagnosis and on hazard of death. Rev Panam Salud Publica. 2004; 15: 160167.
  • 43
    Oluwole SF, Ali AO, Adu A, et al. Impact of a cancer screening program on breast cancer stage at diagnosis in a medically underserved urban community. J Am Coll Surg. 2003; 196: 180188.