SEARCH

SEARCH BY CITATION

Keywords:

  • health disparities;
  • poverty;
  • socioeconomic status;
  • women's health

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Conflict of Interest Disclosures
  7. References

BACKGROUND:

In this study, the authors sought to understand the effects of patient race, ethnicity, and socioeconomic status (SES) on outcomes for cervical cancer.

METHOD:

The Florida Cancer Data System and the Agency for Health Care Administration data sets (1998-2003) were merged and queried. Survival outcomes for patients with invasive cervical cancer were compared between different races, ethnicities, and community poverty levels.

RESULTS:

In total, 5367 patients with cervical cancers were identified. The overall median survival was 43 months. Significantly longer survival was observed for Caucasians (47.1 months vs 28.8 months for African Americans [AA]; P < .001), Hispanics (52.8 months vs 41.6 months for non-Hispanics; P < .001), the insured (63 months vs 41.2 months for uninsured; P < .001), and patients from more affluent communities (53.3 months where <5% lived in poverty vs 36.9 months where >15% lived in poverty; P < .001). Surgery was associated with dramatically improved survival. AA women who were diagnosed with cervical cancer were significantly less likely to undergo surgical treatment with curative intent compared with Caucasian women (P < .001). However, on multivariate analysis, independent predictors of poorer outcomes were insurance status, tumor stage, tumor grade, and treatment. Neither race, nor ethnicity, nor SES was an independent predictor of poorer outcome.

CONCLUSIONS:

Race, ethnic, and SES disparities in cervical cancer survival were explained by late-stage presentation and under-treatment. Earlier diagnosis and greater access to surgery and other treatments would significantly improve the survival of women with cervical cancer. Cancer 2009. © 2008 American Cancer Society.

Cervical cancer represents the third most common gynecologic malignancy diagnosed in the United States. Approximately 11,070 new cases of invasive cervical cancer will be diagnosed in 2008, with an estimated 3870 deaths.1 Disparities in diagnosis, treatment, and outcome for cancers in African Americans (AA) and Caucasians have been documented over the past 30 years.2 The Annual Report to the Nation on the Status of Cancer (1975-2002) indicated that the incidence of cervical cancer deaths was highest for AA women (6.3 per 100,000), followed by Hispanic women (3.9 per 100,000), and Caucasian women (2.7 per 100,000).3 In the current study, we examined the differences in survival observed among patients with invasive cervical cancer based on race, ethnicity, and socioeconomic status (SES). We used a large state cancer registry with a diverse population and with comorbidity information that allowed for better correction for covariates.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Conflict of Interest Disclosures
  7. References

The study dataset consisted of data linked between the Florida Cancer Data System (FCDS), which is Florida's population-based incidence tumor registry, and the Agency for Healthcare Administration (AHCA), which is Florida's healthcare facility discharge database. FCDS data were used to identify all incident cases of invasive cervical cancer diagnosed in the state of Florida between 1998 and 2003. The FCDS dataset was enhanced with data linked from the AHCA dataset for the same years for both follow-up and comorbidities. Primary cases that were diagnosed antemortem and treated surgically with curative intent were identified. In addition, non-Florida residents were not included in the analysis, because follow-up for such patients, particularly survival information, may be inaccurate in up to 10% of such patients.4

AHCA maintains 2 databases (Hospital Patient Discharge Data and Ambulatory Outpatient Data) on all patient encounters within hospitals and freestanding ambulatory surgical and radiation therapy centers in Florida. The AHCA datasets that were used in this study contain diagnoses and procedures performed during every hospitalization or outpatient encounter in the state of Florida for the period from 1998 to 2003. Cases in the FCDS and AHCA datasets were linked on the basis of unique identifiers, birth, and sex.

Postal codes listed in the FCDS-AHCA database were used to determine community poverty levels according to the 2007 US Census Bureau report.5 Insurance and payor status was collected as an FCDS variable.

The staging criteria used by the FCDS are consistent with the National Cancer Institute's Surveillance, Epidemiology, and End Result summary staging and differ from International Federation of Gynecology and Obstetrics staging guidelines. In this study, ‘local’ staging represented disease that did not extend beyond the primary organ, whereas patients who had positive lymph nodes at the time of resection were classified as having ‘regional’ disease. Documentation of distant metastases during the perioperative period led to the classification of affected patients as having ‘distant’ disease.

Statistical analysis was performed with SPSS Statistical Package version 15.0 (SPSS Inc., Chicago, Ill). Comparisons between categorical variables were made by using the chi-square test. Median survival rates were calculated by using the Kaplan-Meier method. Because the FCDS collects only primary cause of death, we analyzed only overall survival and not disease-specific survival. Survival was calculated from the time of the initial diagnosis to the date of last contact (or date of death). The univariate effects of demographic, clinical, and treatment variables on survival were tested by using the log-rank test for categorical values. To estimate the impact of race, ethnicity, and SES on survival outcomes, we used a Cox proportional hazards model and added demographic, clinical, and treatment variables in a stepwise fashion.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Conflict of Interest Disclosures
  7. References

Patient Demographics and Characteristics

Over the 5-year period studied, 5367 Florida residents with invasive tumors of the cervix were identified. Demographics, social characteristics, and tumor characteristics are summarized in Table 1. The majority of the patients were Caucasian (n = 4421 women; 82.4%) and non-Hispanic (n = 4454 women; 83.7%). The majority of women were ages 40 to 64 years (n = 2733 women; 50.9%) and were insured (n = 2567 women; 47.8%). Greater than 66% of tumors were of squamous cell histology (n = 3600 tumors; 67.1%). Localized disease was most common (n = 2442 tumors; 53.1%), followed by regional disease (n = 1679 tumors; 36.5%). Approximately 33.1% of the study population lived in a community where >15% of the area population was living below the poverty line.

Table 1. Demographic, Social, and Clinical Characteristics of the Study Group According to Race, Ethnicity, and Socioeconomic Status
 Entire CohortRaceEthnicityCommunity Poverty Level*
CharacteristicNo.Percentage of TotalCaucasianAAPNon-HispanicHispanicPABCDP
  • AA indicates African American; NOS, not otherwise specified.

  • *

    The community poverty level was defined as communities <5% (A), from 5.1% to 10% (B), from 10.1% to 15% (C), or >15.1% (D) below the poverty line.

Median age at diagnosis, y51.93 51.8552.29.44552.0451.35.25252.2951.9951.8251.73.84
   Percentage of Total
Age group, y             
 <40135725.325.325.1.30525.325.2.16624.725.22426.4.862
 40-64273350.951.349.3 50.553.4 51.450.65250.3 
 ≥65127723.823.425.7 24.221.4 23.824.32423.2 
Race             
 Caucasian442182.4   79.497.3<.0019389.285.668.6<.001
 AA94617.6   20.62.7 710.814.431.4 
Ethnicity             
 Non-Hispanic445483.780.897.6<.001   88.487.283.478.3<.001
 Hispanic86516.319.22.4    11.612.816.621.7 
Community poverty level             
 A101919.021.47.5<.0012013.5<.001     
 B152528.430.817.4 29.522.3      
 C104519.520.216 19.419.8      
 D177833.127.659.1 31.144.4      
Insurance             
 Insured256747.853.536.2<.00152.539.6<.00158.556.147.742.6<.001
 No insurance76714.313.820.8 1326 10.71216.819.2 
 Medicare5159.610.77.6 10.48.9 1010.410.89.5 
 Medicaid63611.910.621.4 11.418.1 7.810.412.716.9 
 Medicare/Medicaid NOS5219.79.812.4 10.96.7 10.89.610.710.2 
 Government831.51.61.6 1.80.8 2.31.51.21.6 
 Unknown2785.2           
Tobacco use             
 Yes205045.447.734.9<.0014927.7<.00147.546.148.641.9.005
 No246354.652.365.1 5172.3 52.553.951.458.1 
 Other854            
Histology             
 Squamous360067.165.375.6<.00167.266.7.93764.864.467.970.2<.001
 Adenocarcinoma88616.518.28.7 16.516.6 21.418.214.413.5 
 Epithelial4207.87.97.4 7.78.3 6.57.58.78.3 
 Other cancer4618.68.68.4 8.68.3 7.49.898 
Tumor stage             
 Localized244253.154.944.3<.00152.754.7.00854.656.252.549.9.034
 Regional167936.535.640.8 36.238.1 35.734.537.238.2 
 Distant48010.49.515 117.2 9.69.410.311.9 
 Unknown766            
Tumor grade             
 Well differentiated42013.214.28.1.00113.212.9.95616.613.91111.7.122
 Moderately differentiated136042.642.642.7 42.742.5 39.942.744.842.9 
 Poorly differentiated133741.940.946.4 41.841.8 41.540.64243.2 
 Undifferentiated762.42.32.8 2.32.7 2.12.92.32.2 
 Unknown2174            
Surgery             
 Yes321561.464.447.2<.0016660.5.00266.163.560.857.2<.001
 No202438.635.652.8 3439.5 33.936.539.242.8 
 Unknown128            
Chemotherapy             
 Yes261448.752.438<.00148.955.3<.0015451.54946.7.001
 No262548.947.662 51.144.7 4648.55153.3 
 Unknown128            
Radiation             
 Yes231044.643.151.5<.00144.744.2.79942.242.345.347.5.011
 No287155.456.948.5 55.355.8 57.857.754.752.5 
 Unknown186            

Survival

Median survival rates for the entire study population are summarized in Table 2. The median survival time (MST) for the entire cohort was 43.03 months. Significantly longer survival was observed in Caucasians (47.1 months vs 28.8 months for AA; P < .001), Hispanics (52.8 months vs 41.6 months for non-Hispanics; P < .001), the insured (63 months vs 41.2 months for women without insurance; P < .001), and patients from more affluent communities (53.3 months vs 36.9 months in communities where >15% lived in poverty; P < .001).

Table 2. Median Survival According to Race, Ethnicity, and Socioeconomic Status
  Median Survival, mo
 Entire CohortRaceEthnicityCommunity Poverty Level*
VariableMedian Survival, moPCaucasianAAPNon- HispanicHispanicPABCDP
  • AA indicates African American; ND, not determined; NOS, not otherwise specified.

  • *

    The community poverty level was defined as communities <5% (A), from 5.1% to 10% (B), from 10.1% to 15% (C), or >15.1% (D) below the poverty line.

Age group, y             
 <4058.0<.00163.428.7<.00159.747.1.89280.1ND43.441.1<.001
 40-6451.6 55.032.1<.00148.871.2<.00154.553.858.045.0.039
 ≥6526.6 27.718.6.01624.434.4.01730.624.227.325.2.370
Race             
 Caucasian47.1<.001   46.153.2.00154.551.546.442.1.085
 AA28.8    28.820.8.26843.129.929.025.7.370
Ethnicity             
 Non-Hispanic41.6<.00146.128.8<.001   53.344.342.532.1<.001
 Hispanic52.8 53.220.8.001   59.071.238.250.1.266
Community poverty level             
 A53.3<.00154.543.1.47553.359.0.514     
 B45.8 51.529.9.00244.371.2.008     
 C42.4 46.429.0.01742.538.2.495     
 D36.9 42.125.7<.00132.150.1<.001     
Insurance             
 Insured63.0<.00167.834.0<.00160.8ND.00680.164.658.446.5.001
 No insurance41.2 46.129.8.14137.642.4.04353.246.134.639.5.645
 Medicare30.4 30.618.2.26030.434.4.10427.031.125.032.1.850
 Medicaid33.8 36.431.0.44132.347.1.03334.331.233.443.0.970
 Medicare/Medicaid  NOS26.4 29.618.9.00925.926.7.83335.922.146.822.9.036
 Government48.4 43.9ND.81743.9ND.89843.961.5ND38.6.474
Tobacco use             
 Yes45.6.95047.729.1.00144.754.5.134ND44.746.435.7.001
 No43.1 50.027.8<.00141.552.8<.00151.158.535.240.5.006
Histology             
 Squamous46.7<.00152.731.0<.00145.963.5.00160.256.746.737.3<..001
 Adenocarcinoma55.0 59.329.1<.00155.067.5.09558.859.347.143.0.120
 Epithelial25.4 26.817.5.11723.538.7.10920.521.826.127.1.616
 Other cancer27.5 30.615.6.02125.541.2.25320.026.127.928.8.874
Tumor stage             
 Localized90.5<.001ND90.5.00190.5ND.231NDND74.090.5.370
 Regional31.7 35.421.9.00130.450.0.00546.032.527.729.6.004
 Distant11.2 11.510.5.15410.912.0.22410.211.511.110.9.916
Tumor grade             
 Well differentiatedND<.001ND68.3.779NDND.046NDNDNDND.981
 Moderately  differentiated54.4 63.131.5<.00153.867.5.075ND60.855.241.7.011
 Poorly differentiated33.3 35.225.5.02531.342.2.02042.242.725.427.8.019
 Undifferentiated14.6 13.414.6.48113.4ND.08614.714.69.519.6.819
Surgery             
 Yes71.0<.00171.251.6.00169.971.0.979ND71.263.163.7.066
 No26.3 28.019.9<.00124.438.2<.00133.626.627.222.6.029
Chemotherapy             
 Yes33.0<.00135.425.7<.00130.546.9.00340.235.023.733.0.247
 No52.1 60.732.3.05551.263.2.01262.258.555.241.9<.001
Radiation             
 Yes31.6<.00134.224.9<.00129.847.1<.00139.631.827.729.6.014
 No63.1 66.736.7.00363.163.5.27571.066.465.454.4.016

A decrease in MST was observed as age increased in the study population. Survival was significantly longer with local disease stage than with distant disease (MST: 90.47 months vs 11.17 months; P < .001). Patients with well differentiated tumors fared better than those with poorly differentiated tumors (MST not reached vs 33.33 months; P < .001) (Fig. 1B). Women who underwent surgical extirpation had significantly longer survival (MST: 71 months vs 26.3 months; P < .001) (Fig. 1C). Kaplan-Meier survival curves illustrating these findings are shown in Figure 1.

thumbnail image

Figure 1. Kaplan-Meier survival curves for cervical cancer according to (A) stage of disease, (B) tumor grade, and (C) surgical resection with curative intent.

Download figure to PowerPoint

African Americans Have Worse Survival Outcomes

Univariate subset analysis demonstrated that AA race conferred a significantly poorer prognosis for cervical cancer. AA women did not differ significantly from Caucasian women with respect to age at diagnosis. However, AA women lived in communities with significantly higher levels of poverty; a significantly larger percentage was diagnosed with cervical cancer of squamous histology; and, at the time of diagnosis, AA women had significantly more regional and distant disease. In addition, AA women had more poorly differentiated tumors compared with their Caucasian counterparts.

The MST of AA women with cervical cancer was significantly shorter than for Caucasian women among all age strata. For all tumor stages and grades, with the exception of undifferentiated tumors, the MST for AA patients was significantly lower than for Caucasian patients. Almost 60% of AA patients who were diagnosed with cervical cancer lived in a community in which at least 15% of the local population was below the poverty line compared with only 27.6% of Caucasian patients (P < .001). At all poverty level strata, AA women had a shorter MST than Caucasian women.

Differences in treatment modality and outcomes between AA and Caucasian patients also were observed. AA women who were diagnosed with cervical cancer were significantly less likely to undergo surgical treatment with curative intent. Although more Caucasian patients underwent surgical resection (64.4% vs 47.2%; P < .001), AA patients received more radiation therapy (51.5% vs 43.1%; P < .001).

Ethnic Differences in Cervical Cancer

Differences in outcomes for patients with different ethnic backgrounds also were observed in univariate subset analyses. Hispanic Caucasian patients had a longer MST than non-Hispanic Caucasian patients (53.17 months vs 46.1 months; P < .001), whereas AA Hispanics and AA non-Hispanics did not differ significantly from one another with respect to survival. Across all tumor stages and grades, non-Hispanic women fared worse than Hispanic women; however, most results were not statistically significant.

More Hispanic patients lived in communities that had a >15% incidence of poverty compared with non-Hispanic patients (44.4% vs 31.1%; P < .001). Despite this, Hispanic patients had a significantly longer MST than non-Hispanic patients. Hispanic patients were significantly less likely than non-Hispanic patients to undergo surgery for cervical cancer (60.5% vs 66%; P = .002), but there was no statistical difference in the proportion of Hispanic patients and non-Hispanic patients who received chemotherapy or radiation. The MST in Hispanic patients was similar to that in non-Hispanic patients who underwent surgical treatment (71 months vs 69.9 months, respectively; P = .98).

Effects of Area Poverty on Prognosis of Cervical Cancer

Univariate analysis demonstrated that the community poverty level also affects MST in cervical cancer. Patients who lived in communities in which >15% of the population lived in poverty had significantly worse survival outcomes compared with communities in which there was less poverty. There was no significant difference in MST for patients who lived in communities of all poverty levels of all age groups. Patients from communities in which >15% of the population lived in poverty more frequently presented with advanced-stage disease compared with patients who lived in communities with a smaller proportion of poverty. Patients who lived in communities with the greatest amount of poverty were less likely to receive surgical treatment for their cervical cancer. Furthermore, among all patients who underwent surgical treatment, the patients who lived in the poorest communities experienced worse survival outcomes compared with patients who lived in communities with the least amount of poverty (MST: 63.7 months vs not yet reached; P = .06)

Tobacco

Caucasian women who were diagnosed with cervical cancer were significantly more likely to use tobacco compared with AA women (47.7% vs 34.9%; P < .001). The survival of Caucasian smokers was significantly longer than that of their AA counterparts (47.7 months vs 29.1 months; P = .001). A survival advantage also was observed for Caucasians who did not smoke compared with AA nonsmokers (50.03 months vs 27.8 months; P < .001). A greater proportion of non-Hispanic women than Hispanic women were smokers (49% vs 27.7%; P < .001). In the Hispanic cohort who smoked, survival was longer than survival in non-Hispanic smokers (54.47 months vs 44.73 months); however, this result was not statistically significant. Abstinence from tobacco use conferred a significantly greater survival advantage for Hispanic women compared with non-Hispanic women (MST: 52.83 months vs 41.53 months; P < .001).

Comorbidities in Cervical Cancer

The frequency of comorbidities (Elixhauser) in the study population is shown in Table 3. On univariate analysis, AA women were more likely to suffer from renal failure, human immunodeficiency virus/acquired immunodeficiency syndrome, and rheumatoid arthritis/collagen vascular disease compared with Caucasian women. No differences in comorbidities were noted between Hispanic and non-Hispanic women or by differing community poverty levels. On multivariate analysis, comorbid conditions did not predict survival significantly in patients who were diagnosed with cervical cancer (data not shown).

Table 3. Elixhauser Comorbidities According to Race, Ethnicity, and Socioeconomic Status
Frequency
 Percentage by RacePercentage by EthnicityPercentage by Community Poverty Level*
Overall PercentageCaucasianAAPNon-HispanicHispanicPABCDP
  • AA indicates African American.

  • *

    The community poverty level was defined as communities <5% (A), from 5.1% to 10% (B), from 10.1% to 15% (C), or >15.1% (D) below the poverty line.

0.520.570.32.0690.60.3.0410.70.70.40.4.141
0.971.000.85.0521.00.9.0211.41.01.10.6.115
0.580.610.42.0690.60.5.0380.80.80.40.4.104
0.190.200.11.0690.20.2.0290.40.20.20.1.860
0.320.360.11.0560.40.0.0210.60.40.20.2.087
2.933.231.48.0143.21.6.0323.83.83.21.7.012
0.220.270.00.0380.20.5.0030.70.20.00.1.003
0.350.380.21.0690.40.0.0190.30.50.60.1.033
1.731.950.74.0251.90.8.0282.62.01.71.0.055
0.710.720.63.0540.70.6.0360.60.70.90.7.060
0.070.090.00.0570.10.0.0360.10.10.10.1.165
0.390.450.11.0470.50.0.0170.30.40.90.2.006
0.240.230.32.0380.30.0.0250.40.30.20.2.151
0.540.610.21.0530.60.2.0381.10.60.50.2.036
0.130.160.00.0490.20.0.0320.20.10.10.1.161
0.070.020.32<.0010.10.0.0360.10.10.00.1.113
0.110.090.21.0260.10.1.0380.00.10.20.1.067
1.161.131.27.0131.30.6.0371.61.31.10.8.158
0.300.290.32.0520.30.2.0400.50.20.50.2.064
0.320.340.21.0700.30.3.0280.60.30.20.2.098
0.610.630.53.0600.70.5.0390.60.80.40.6.062
1.681.701.59.0201.90.8.0311.92.01.41.4.080
0.390.430.21.0670.40.1.0330.70.50.20.2.069
0.280.290.21.0690.30.2.0390.50.10.50.2.038
0.390.430.21.0670.40.2.0420.60.50.50.2.111
0.090.110.00.0540.10.0.0340.20.10.20.0.062
0.170.180.11.0690.20.0.0290.40.00.30.1.019
0.560.630.21.0510.60.5.0351.00.70.30.4.062

Multivariate Analysis

The results of multivariate analysis using the Cox regression model are summarized in Table 4. Independent predictors of survival for patients with cervical cancer in the final model included insurance payor status, tumor stage, tumor grade, surgical extirpation, chemotherapy, and radiation therapy. Differences in survival after a diagnosis of cervical cancer between Hispanic women and non-Hispanic, AA, and Caucasian women and between variable levels of community poverty failed to reach statistical significance in the final multivariate model.

Table 4. Multivariate Regression Analyses*
VariableHRLowerUpperP
  • HR indicates hazard ratio; NOS, not otherwise specified.

  • *

    The regression model included variables for Elixhauser comorbid conditions.

  • The community poverty level was defined as communities <5% (A), from 5.1% to 10% (B), from 10.1% to 15% (C), or >15.1% (D) below the poverty line.

Age, y    
 <40Reference group   
 40-640.910.7451.109.346
 ≥651.080.8211.413.593
Race    
 CaucasianReference group   
 African American1.100.9121.315.329
Ethnicity    
 Non-HispanicReference group   
 Hispanic0.840.6661.055.133
Community poverty level    
 A (wealthiest)Reference group   
 B1.070.8581.343.535
 C1.220.9591.540.106
 D (poorest)1.210.9731.507.086
Insurance    
 InsuredReference group   
 No insurance1.220.9731.520.085
 Medicare1.351.0201.792.036
 Medicaid1.220.9741.530.083
 Medicare/Medicaid NOS1.581.2102.052<.001
 Government1.180.6562.124.581
Histology    
 SquamousReference group   
 Adenocarcinoma1.210.9951.479.056
 Epithelial1.350.9391.946.105
 Other cancer1.060.7901.431.685
Tumor stage    
 LocalizedReference group   
 Regional2.692.1943.306<.001
 Distant6.795.3398.635<.001
Tumor grade    
 Well differentiatedReference group   
 Moderately differentiated1.441.0581.961.020
 Poorly differentiated1.931.4222.618<.001
 Undifferentiated3.702.2776.016<.001
Surgery    
 NoReference group   
 Yes0.380.3150.449<.001
Chemotherapy    
 NoReference group   
 Yes0.740.6210.887.001
Radiation    
 NoReference group   
 Yes0.720.5990.877<.001

Women who had Medicare coverage were 35% more likely to die during the study period (P = .04). Women who had Medicare/Medicaid coverage, not otherwise specified, were 58% more likely to die during the study period (P < .001) compared with women who had private insurance coverage.

Clinical and treatment characteristics proved to have the largest effect on survival outcomes. Compared with women who had localized disease at the time of diagnosis, women who had regional disease were 2.69 times more likely to die during the study period, and women with distant disease were 6.79 times more likely to die during the study period (P < .001). A dose-response effect was observed between tumor grade and the risk of death during the study period. Women who had undifferentiated tumors were 3.70 times more likely to die during the study period compared with women who had well differentiated tumors (P < .001). Women who underwent surgical extirpation had a 62% decreased risk of death compared with women who did not undergo surgery, whereas women who received radiation therapy had a 28% decreased risk of death compared with women who did not receive radiation therapy (P < .001). The receipt of chemotherapy also decreased the risk of death significantly (by 26%) during the study period (P = .001) compared with women who received no chemotherapy.

A stepwise, multivariate analysis using the Cox regression model is summarized in Table 5. After correcting for clinical characteristics and treatment, the hazard ratio (HR) for death during the study period for AA women with cervical cancer decreased substantially from 1.23 to 1.10 and was no longer statistically significant. Likewise, the HR for death during the study period for Hispanic women with cervical cancer increased from 0.78 to 0.84 and was no longer statistically significant after the addition of treatment into the model. For poverty, the addition of treatment—surgical extirpation, radiation therapy, and chemotherapy—decreased the HR from 1.32 to 1.21, and the model was no longer statistically significant.

Table 5. Stepwise Analysis of Survival in Cervical Cancer
VariableHRLowerUpperP
  • HR indicates hazard ratio.

  • *

    Race: African Americans compared with Caucasians.

  • Demographics: age, race, ethnicity, and poverty.

  • Comorbidities: Elixhauser comorbid conditions.

  • §

    Clinical characteristics: tumor grade, stage, and histology.

  • ||

    Treatment: surgery, chemotherapy, and/or radiation.

  • Ethnicity: Hispanics compared with non-Hispanics.

  • #

    Poverty: communities >15 % below the poverty level compared with communities <5% below the poverty level.

Race*1.461.3021.627<.001
 Race+demographics1.341.1911.513<.001
 Race+demographics+comorbidities1.311.1611.478<.001
 Race+demographics+comorbidities+insurance1.231.0821.400.0015
 Race+demographics+comorbidities+insurance+clinical characteristics§1.190.9931.418.0603
 Race+demographics+comorbidities+insurance+clinical characteristics+treatment||1.100.9121.315.3291
Ethnicity0.740.6460.858<.001
 Ethnicity+demographics0.770.6680.894<.001
 Ethnicity+demographics+comorbidities0.760.6550.878<.001
 Ethnicity+demographics+comorbidities+insurance0.740.6340.863<.001
 Ethnicity+demographics+comorbidities+insurance+clinical characteristics0.780.6240.976.0300
 Ethnicity+demographics+comorbidities+insurance+clinical characteristics+treatment0.840.6661.055.1331
Poverty#1.331.1581.528<.001
 Poverty+demographics1.281.1031.474.0010
 Poverty+demographics+comorbidities1.261.0911.461.0017
 Poverty+demographics+comorbidities+insurance1.251.0721.455.0044
 Poverty+demographics+comorbidities+insurance+clinical characteristics1.321.0631.634.0119
 Poverty+demographics+comorbidities+insurance+clinical characteristics+treatment1.210.9731.507.0858

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Conflict of Interest Disclosures
  7. References

Differences in cancer survival based on race, ethnicity, and SES remain major issues despite the recognition of these inequalities for more than 30 years. In an effort to understand the outcomes of patients with cervical cancer and potentially to improve survival, we examined a population-based registry to identify global prognostic factors that are important in the survival of patients who are diagnosed with cervical cancer. In the current analysis, emphasis was placed on the effects of race, ethnicity, and area-based poverty level on overall survival. Although our report is consistent with others who also reported racial and socioeconomic disparities in cancer,6-11 this study was unique, in that a diverse patient population with all age groups, as well as their comorbid conditions, was included in the analysis. To our knowledge, this study represents the largest, most comprehensive analysis of these variables on the outcomes of patients with invasive cervical cancer to date.

After controlling for sociodemographic, clinical, and treatment characteristics (as well as comorbidities), we observed that race, ethnicity, and poverty were no longer independent predictors of survival in patients with invasive cervical cancer. With respect to race, these results are supported by previous studies that lacked comorbidity data.7, 12, 13 In the cohort reported herein, AA patients had significantly more regional and distant disease and more poorly differentiated tumors compared with their Caucasian counterparts. AA women also were significantly less likely to undergo surgical treatment with curative intent. Our comorbidity data certainly did not indicate that AA women were more likely to be at increased operative risk secondary to poor health; however, the findings of more advanced disease at presentation may indicate that the extent of disease made these patients more likely to be inoperable at time of diagnosis, leading to more frequent use of palliative chemotherapy and radiation treatment.

The survival of AA patients who did undergo surgery was significantly shorter in the univariate analysis; however, after controlling for sociodemographic variables and clinical disease characteristics, there was no survival advantage for Caucasian patients who underwent surgery compared with AA patients. These results indicate that, for patients who undergo surgical treatment for invasive cervical cancer, race alone does not have a significant impact on survival. AA women reportedly are more likely to be screened for cervical cancer than their Caucasian counterparts.14-18 Nonetheless, our current findings suggest that a large fraction of these women may fail to be screened, resulting in the development of frank malignancy, or that poor survival outcomes for AA patients, regardless of poverty level, may be the result of poor follow-up of these screening results, ultimately resulting in delays in treatment or less than adequate treatment and presentation with advanced-stage disease.

In addition to differences in survival by race, we also observed that Hispanic women with invasive cervical cancer tended to present with more advanced disease, but they experienced longer survival. This observation is consistent with other studies that identified Hispanic ethnicity as a protective factor in cervical cancer survival.7, 13, 19, 20 In an analysis of the Texas Cancer Registry, Eggleston et al7 reported that Hispanic women were 30% less likely to die from cervical cancer during follow-up after adjusting for confounders despite being diagnosed at a later stage. Patel et al13 reported that, after adjusting for disease characteristics, Hispanic women were at a 26% decreased risk of death compared with non-Hispanic women over the follow-up period. Those authors attributed the survival advantage to the diagnosis of cervical cancer at a younger age for Hispanic women; however, we observed that the survival advantage persisted, although we observed no significant difference in age at diagnosis between Hispanics and non-Hispanics. In contrast to a previous analysis using the FCDS registry from 1981 to 1989,19 which reported that Hispanics had disproportionately later stage cervical cancer compared with non-Hispanics, our analysis indicated that non-Hispanics presented with significantly later stage cancer, possibly explaining our findings. These results also may point to improved cervical cancer screening and access to medical care for Hispanic women.

Studies that have examined SES as an independent predictor of cervical cancer mortality have produced conflicting results.6-11 Movva et al12 determined that, after adjustment for confounding variables, race no longer had a significant impact on survival, but women who resided in a ‘working-poor’ census tract had an increased risk of death relative to women who lived in a ‘professional’ census tract. Eggleston et al7 observed that the association between lower SES and poorer survival was consistent across all racial/ethnic groups. We observed significantly decreased survival across all ages and races with increasing levels of community poverty in our univariate analysis. Women living in communities that had >15% of the population living in poverty had significantly decreased MST compared with women living in communities with less poverty, even when they underwent surgery for their cervical cancer. After adjustment for confounding variables in our multivariate analysis, community poverty level was no longer an independent predictor of survival. Another study of women aged ≥65 years with cervical cancer reported that neither racial/ethnic differences nor socioeconomic factors were independent predictors of a poor outcome.21 Similar to the findings in our study, Coker et al21 reported that women who underwent surgical intervention had significantly longer overall survival compared with women who did not undergo surgery. These results suggest that, when socioeconomically disadvantaged women with invasive cervical cancer undergo surgically, their survival outcomes are no different from those of women from more affluent communities. Our findings probably are more robust, because we included a more heterogeneous group of patients with multiple payor sources. Much like the association between AA race and survival, these results likely indicate that poorer women more frequently are subject to incomplete follow-up of screening results and treatment.

Because access to care has been highlighted as a major factor affecting cancer survival, many organizations have increased their efforts to level the playing field in the burden of cancer. The American Cancer Society has made this issue the centerpiece of their goal to be attained by the year 2015,22 and the US Department of Health and Human Services has made a commitment to reduce cancer disparities with the Healthy People 2010 initiative. These efforts are focused primarily on modifiable socioeconomic factors, such as poverty level, education, and healthcare. The results of our study suggest that programs that increase the likelihood that AA and poor populations will seek treatment and education programs aimed at better compliance with treatment programs may have the most substantial impact on cervical cancer survival.

The data collected from large cancer registries provide insight into tumor behavior and allow us to examine outcomes from current treatment strategies.4, 23-27 Although this represents an excellent database for comparative outcomes analysis, it is not without limitations. The use of area-based poverty as a proxy for SES may result in misclassification of some patients whose postal codes do not reflect accurately the true income level of the individual. In addition, the FCDS records only primary cause of death; consequently, we were unable to include disease-specific survival in our analysis. Furthermore, although data on radiotherapy and chemotherapy were examined, information on specific regimens and dosages were not available.

In conclusion, after controlling for treatment modality, we could not demonstrate racial or SES disparities in cervical cancer survival. The inequalities observed in the univariate analysis were explained by tumor characteristics and under-treatment; poor outcomes disappeared after correcting for these factors. Programs geared toward diminishing disparities in cervical cancer survival should be aimed at improved screening and earlier diagnosis, because the results from this study suggest that there is equality in the management of cervical cancer after diagnosis and proper treatment.

References

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Conflict of Interest Disclosures
  7. References