SEARCH

SEARCH BY CITATION

Keywords:

  • cervical cancer;
  • healthcare access;
  • Hispanics;
  • incidence

Abstract

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

BACKGROUND.

Cervical cancer mortality rates have declined in the United States, primarily because of Papanicolaou testing. However, limited information is available about the incidence of the disease in the US-Mexico border region, where some of the poorest counties in the United States are located. This study was undertaken to help compare the patterns of cervical cancer incidence among women in the US-Mexico border region and other parts of the United States.

METHODS.

Age-adjusted cervical cancer incidence rates for border counties in the states bordering Mexico (California, Arizona, New Mexico, Texas) for the years 1998 to 2003 were compared with the rates for nonborder counties of the border states and with those of nonborder states. Differences were examined by age, race, ethnicity, rural residence, educational attainment, poverty, migration, stage of disease, and histology.

RESULTS.

Overall, Hispanic women had almost twice the cervical cancer incidence of non-Hispanic women in border counties, and Hispanic women in the border states had higher rates than did non-Hispanic women in nonborder states. In contrast, cervical cancer incidence rates among black women in the border counties were lower than those among black women in the nonborder states. Among white women, however, incidence rates were higher among those in nonborder states. Differences in cervical cancer incidence rates by geographic locality were also evident by age, urban/rural residence, migration from outside the United States, and stage of disease.

CONCLUSIONS.

Disparities in cervical cancer incidence in the US-Mexico border counties, when the incidence is compared with that of other counties and geographic regions, are evident. Of particular concern are the higher rates of late-stage cervical cancer diagnosed among women in the border states, especially because such cervical cancer is preventable. Cancer 2008;113(10 suppl):2964–73. Published 2008 by the American Cancer Society.

The US-Mexico border region extends about 2000 miles, from the Pacific Ocean to the Gulf of Mexico. It includes some of the poorest counties in the United States.1 More than a third of US border families live at or below the poverty line, and the unemployment rate is 2.5 to 3.0 times higher than in the rest of the United States. Lack of access to healthcare is an important problem in the border region.1 Other problems include an uneven distribution of physicians and other healthcare providers, a shortage of bilingual health information, and a lack of culturally sensitive systems of care. The public health concerns of the US-Mexico border region include relatively low cervical cancer screening rates, especially among older Hispanic women, and relatively high death rates from cervical cancer.2–5

Cervical cancer mortality rates have declined in the United States, but little information has been available about the incidence of the disease in the US-Mexico border region, where poverty, lower education, and inadequate access to healthcare services are prevalent social problems. A recent report about cervical cancer in Texas from 1999 to 2003 suggested that some Texas women may be more likely than others to develop cervical cancer, depending on where they live.6 Hispanic women living in 32 counties along the Texas-Mexico border had a 19% higher cervical cancer mortality rate than Hispanic women living in nonborder areas. Among women in Texas who were diagnosed with cervical cancer, only 40.2% of Hispanic women in the border counties had early-stage disease, compared with 49.3% of nonborder Hispanic women. White women in rural counties of Texas had a 27% higher cervical cancer incidence rate than white women residing in urban counties of the state. An analysis of cervical cancer incidence data from California over the period 1988 to 2002 found some variation in incidence rates by race/ethnicity and geographic locality.7 However, the incidence of cervical cancer in California counties adjacent to the US-Mexico border was not uniformly higher than in other counties in the state.

To help clarify these relationships, we examined cervical cancer incidence among women in the US-Mexico border counties during the years 1998 to 2003 by age, race, Hispanic ethnicity, education, poverty, urban/suburban/rural residence, stage, histology, and the percentage of the county population who had migrated to the United States. Cervical cancer incidence among women in US-Mexico border counties was compared with the incidence in other counties in the 4 border states (California, Arizona, New Mexico, and Texas), and also with the incidence in nonborder states. The current analysis differs from that of other articles included in this Supplement in that the focus is on cervical cancer incidence and stage at diagnosis in the US-Mexico border region, the residents of which are disproportionately impacted by cervical cancer.

MATERIALS AND METHODS

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

The databases used for all analyses were 1998 to 2003 data from the National Program of Cancer Registries (NPCR) and National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) program.8 The use of NPCR and SEER data allowed for a more comprehensive examination of cervical cancer incidence in border and nonborder counties of border states. Analyses were limited to 37 registries that met the US Cancer Statistics Publication Standard for data quality (USCS Publication Standard). Four of these registries (hereafter referred to as US-Mexico border states) were the state central registries for California, Arizona, New Mexico, and Texas. California is part of SEER and the NPCR, New Mexico is a SEER state, and Arizona and Texas are included in the NPCR. To provide the maximum number of cases with high-quality data, we used data for the years 1998 to 2003 for California, New Mexico, and Texas, whereas data for Arizona were limited to 1998 to 2002 (the years that met the USCS Publication Standard).

The remaining 33 registries (hereafter referred to as those for nonborder states) included the central registries for the District of Columbia and 32 states (Alabama, Alaska, Arkansas, Colorado, Connecticut, Delaware, Florida, Idaho, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maine, Massachusetts, Michigan, Missouri, Montana, Nebraska, New Jersey, New York, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode Island, South Carolina, Utah, Vermont, Washington, West Virginia, and Wisconsin). Each of the nonborder states had high-quality data available from 1998 to 2003.

Several different definitions of US-Mexico “border counties” have been proposed.9 For example, in Texas, variations have included 14 counties with boundaries that touch the US-Mexico border at some point; 32 counties in a Health Services Resources Administration (HRSA) border region, where “border counties” are defined as those counties where any part is within the 100-km (62.14 miles) region north of the US border with Mexico; and 43 counties in a South Texas border region.10

The HRSA definition of the border region presents some problems with certain counties; in California, for example, only a small portion of Riverside County is included in the 100-km region used for the HRSA definition. To reduce misclassification of geographic residence, we used the state definition of border counties for this analysis, rather than the HRSA definition, although in New Mexico and Texas the state definition and HRSA definition are identical. Table 1 provides a list of the 44 counties that the 4 border states (California, Arizona, New Mexico, Texas) defined as “border counties” and that we used for this study. The locations of these counties, and their population density, are shown in Figure 1. In this figure, the population density is derived from the LandScan world population database (LandScan 2002; Oak Ridge National Laboratory, Oak Ridge, Tenn),11 which has a grid cell size of 30 arc seconds (about 1 km2). High-density areas have between 2000 and 7000 people per square kilometer; low-density areas have between 0 and 10 people per square kilometer.

thumbnail image

Figure 1. US-Mexico border counties and population density are shown for the United States, 2002.

Download figure to PowerPoint

Table 1. Counties Identified as “Border Counties” by Border States (Arizona, California, New Mexico, and Texas)
Border StateBorder Counties
Arizona4 counties: Cochise, Pima, Santa Cruz, Yuma
California2 counties: San Diego, Imperial
New Mexico6 counties: Dona Ana, Grant, Hidalgo, Luna, Otero, Sierra
Texas32 counties: Brewster, Brooks, Cameron, Crockett, Culberson, Dimmit, Duval, Edwards, El Paso, Frio, Hidalgo, Hudspeth, Jeff Davis, Jim Hogg, Kenedy, Kinney, La Salle, Maverick, McMullen, Pecos, Presidio, Real, Reeves, Starr, Sutton, Terrell, Uvalde, Val Verde, Webb, Willacy, Zapata, Zavala
Total44 counties

Results were aggregated for the purpose of examining the 44 counties in Table 1 as a single entity, as compared with the aggregated results for the remaining 316 counties in the border states. In addition, comparisons were made with aggregated results for the nonborder states as a single entity.

Counties were categorized as “urban,” “suburban,” or “rural,” using the US Department of Agriculture 2003 Rural-Urban Continuum Codes.12 Codes 1-3 correspond to “urban” (metropolitan) counties (including metropolitan areas with populations of about 250,000 to greater than 1 million); codes 4-5 correspond to predominately suburban populations of 20,000 or greater, but less than 250,000; and codes 6-9 correspond to rural populations and small towns of up to 19,999.

The analysis was limited to microscopically confirmed cases of invasive cervical carcinomas, as defined earlier (International Classification of Diseases for Oncology-3 8010-8671 and 8940-8941).13 The stage at diagnosis was categorized by using SEER Summary Stage 1977 for cancers from 1998 to 2000, and SEER Summary Stage 2000 for cancers from 2001 to 2003. Data for the 2 staging systems were combined, because the differences observed in comparative analyses were minimal.14 To minimize problems with small numbers, we combined the regional stage with the distant stage, resulting in 3 groups: early stage (localized),1 late stage,2 and unknown stage/unstaged.3 Analysis of histology was limited to identification of percentages, rates (with 95% confidence intervals for those rates), and rate ratios by region for squamous cell carcinomas, adenocarcinomas, adenosquamous/glassy cell, small cell/neuroendocrine, and other/unspecified carcinomas.

We examined invasive cervical cancer incidence rates in the US-Mexico border counties by age categories, race (white, black, Asian/Pacific Islander), Hispanic ethnicity, urban/suburban/rural residence, educational attainment (percentage of county population aged 25 years or more who did not graduate from high school), poverty (percentage of persons in the county below the poverty level), the percentage of persons in a county living in the same house for over 5 years (as a potential marker for individuals more likely to have an established physician to make screening recommendations and also to have an established social network to encourage screening), and the percentage of persons who moved from outside the United States into the county. In this analysis, Hispanic ethnicity was not mutually exclusive of race.13 The categories for education, poverty, no migration, and moved from outside the United States were calculated by dividing the percentage for each variable into 2 groups. The dividing point for creating the 2 groups was the median percentage for the border counties for that variable.

Age-specific rates, direct age-adjusted incidence rates, and rate ratios were calculated in SEER Stat (National Cancer Institute, Bethesday, Md) using the 2000 US standard population with 19 age groups, and expressed per 100,000 females. The 19 age groups were: <1 year, 1-4 years, 5-9 years, 10‒14 years, 15-19 years, 20‒24 years, 25-29 years, 30-34 years, 35-39 years, 40-44 years, 45-49 years, 50-54 years, 55-59 years, 60-64 years, 65-69 years, 70-74 years, 75-79 years, 80-84 years, and ≥85 years.15 Ninety-five percent confidence intervals were estimated by use of an approach proposed by Tiwari et al.16 As a result of misclassification of data on race for American Indians and Alaska Natives, incidence rates were not calculated separately for them, but these persons were included in the overall analyses.

To smooth variation observed in the age-specific rates based upon 5-year age categories, smoothed age-specific rates were calculated using 15-year age groups that advance by 5 years. The starting point for calculating the 15-year moving averages was the number of cases and the corresponding population for each of 19 5-year age groups (Census P25-1130). The moving average rate was calculated as the sum of the total cases for a 15-year period divided by the sum of population denominators for the same period. For example, the 15-year age group 10-24 years was calculated by adding data for the 5-year age groups 10-14 years, 15-19 years, and 20-24 years. Next, the 15-year age group 15-29 years was calculated by adding data for the 5-year age groups 15-19 years, 20-24 years, and 25-29 years. Next, the 15-year age group 20-34 years was calculated by adding data for the 5-year age groups 20-24 years, 25-29 years, and 30-34 years. This process of advancing the 15-year age groups by 5 years was continued through the remaining age groups.15

RESULTS

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

The median age was in the 45-49 year age category for all 3 locations (46.9 years for other counties in border states, and 48.4 years for both border counties and for nonborder states). Figure 2 shows the smoothed age-specific incidence rates. The cervical cancer incidence rates were higher in the border counties and nonborder counties of border states than in nonborder states for older women. In border counties, 5-year age-specific incidence rates were higher than in other counties in border states to age ≥85 years. The 5-year age-specific rates in other counties in border states were higher than rates in nonborder states to age 75 to 79 years.

thumbnail image

Figure 2. Invasive, age-specific cervical cancer carcinomas incidence rates (per 100,000 women) are shown by 15-year moving average age groups and location in relation to the US-Mexico border, United States, 1998-2003. Data are from 37 population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance, Epidemiology, and End Results program, meet high-quality data criteria (Watson et al13), and provide county-level incidence data. See Materials and Methods for a list of the 37 registries. Arizona data are from 1998 to 2002. These registries cover approximately 76.5% of the US population. The starting point for calculating the 15-year moving averages was the number of cases and the corresponding population for each of 19 5-year age groups (Census P25-1130). The moving average rate was calculated as the sum of the total cases for a 15-year period divided by the sum of population denominators for the same period. For example, the 15-year age group 10-24 years was calculated by adding data for the 5-year age groups 10-14 years, 15-19 years, and 20-24 years. Next, the 15-year age group 15-29 years was calculated by adding data for the 5-year age groups 15-19 years, 20-24 years, and 25-29 years. Then the 15-year age group 20-34 was calculated by adding data for the 5-year age groups 20-24 years, 25-29 years, and 30-34 years. This process of advancing the 15-year age groups by 5 years was continued through the remaining age groups.

Download figure to PowerPoint

Table 2 shows the overall age-adjusted incidence rates of cervical cancer and rates by race, Hispanic ethnicity, and socioeconomic factors in the 3 regions (border counties, nonborder counties in border states, and nonborder states). Overall, both women in border counties and other counties of border states had higher cervical cancer incidence rates than women in nonborder states, but rates in women in border counties were not significantly different from rates in women in other counties in border states. Cervical cancer incidence rates among white women in border counties were higher than in the other counties of border states, and rates among white women in other counties in border states were higher than among white women in nonborder states. Hispanic women had higher cervical cancer incidence rates than non-Hispanic women in all 3 regions. In addition, Hispanic women in other counties of border states had significantly higher incidence than Hispanics in nonborder states.

Table 2. Age-Adjusted Cervical Carcinoma Rates* and Rate Ratios for Selected Characteristics, by Location in Relation to the US-Mexico Border, Using Nonborder States as the Reference Group, United States, 1998 to 2003
CharacteristicBorder CountiesOther Counties in Border StatesNonborder States (Reference Group)
CasesRateRate Ratio (95% CI)CasesRateRate Ratio (95% CI)CasesRateRate Ratio
  • CI indicates confidence interval (calculated with Tiwari et al16 modification).

  • Percentages indicate percent of total cases for that subcategory.

  • Percentage <high school education indicates percentage of county population aged 25 years or more who did not graduate from high school (Census 2000, SF3 table P37). The percentage is calculated with the formula: ([P037003+…+P037010+P037020+…+P037027]/P037001)·100.

  • Percentage persons <poverty level indicates percentage of persons in the county whose incomes were below the poverty level in 1999 (Census 2000, SF3 table P87). The percentage is calculated with the formula: (P087002/P087001)·100.

  • In same house >5 y indicates percent of persons in the county living in the same house in 2000 as in 1995 (Census 2000, SF3 table P24). The percentage is calculated with the formula: (P024002/P024001)·100.

  • Moved from outside the US indicates percentage of persons in the county who moved from outside the United States into the county (Census 2000, SF3 table P24). The percentage is calculated with the formula: ([P024013+P024016]/P024001)·100.

  • The cases in nonborder state subcategories do not total to the 44,032 cases listed for the US age-adjusted rate because: a) the race category does not include cases for American Indian/Alaska Native, Other Unspecified, and Unknown; and b) 75 cases in the nonborder state category had missing information on county characteristics for rural-urban status, education, poverty, in same house, and moved from outside the US.

  • *

    Data are from 37 population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance, Epidemiology, and End Results program, meet high-quality data criteria (Watson et al13), and provide county-level incidence data. See Materials and Methods for a list of the 37 registries. Arizona data is from 1998 to 2002. These registries cover approximately 76.5% of the US population.

  • Rates are per 100,000 women and age-adjusted to the 2000 US standard population (19 age groups, Census P25-1130).

  • The rate ratio indicates that the rate is significantly different (P<.05) from the reference group (nonborder states).

  • §

    Hispanic origin is not mutually exclusive from race categories (white, black, and Asian/Pacific Islander).

  • Information is suppressed because the number of cases is <16.

Rural-Urban         
 Urban1,5839.51.1 (1.1-1.2)13,7539.21.1 (1.1-1.1)35,1578.51.0
 Suburban10110.51.2 (0.9-1.4)56110.01.1 (1.0-1.2)3,5009.01.0
 Rural9211.51.2 (0.9-1.5)6879.51.0 (0.9-1.1)5,3009.71.0
Race         
 White1,6219.81.2 (1.2-1.3)12,1779.31.1 (1.1-1.2)34,8428.11.0
 Black285.10.4 (0.3-0.6)1,42110.60.8 (0.8-0.9)7,08312.81.0
 Asian/Pacific Islander919.11.1 (0.8-1.3)1,0978.11.0 (0.9-1.0)9668.51.0
Ethnicity§         
 Hispanic1,04813.91.1 (1.0-1.1)5,11814.81.1 (1.1-1.2)3,82613.21.0
 Non-Hispanic7287.00.8 (0.8-0.9)9,8837.90.9 (0.9-1.0)40,2068.41.0
% <high school education         
 ≤40.151,2508.81.0 (1.0-1.1)14,9949.31.1 (1.1-1.1)43,6338.61.0
 >40.1652612.91.0 (0.9-1.2)32412.41.0
% persons <poverty level         
 ≤25.211,2508.81.0 (1.0-1.1)14,9049.31.1 (1.1-1.1)42,5838.61.0
 >25.2252612.51.0 (0.9-1.1)978.90.7 (0.6-0.9)1,37412.61.0
In same house >5 y         
 ≤59.16%1,5049.31.1 (1.0-1.1)14,5809.31.1 (1.1-1.1)26,6318.51.0
 >59.17%27212.81.4 (1.3-1.6)4219.31.0 (0.9-1.1)17,3268.91.0
Moved from outside the US         
 ≤2.81%7112.01.4 (1.1-1.8)3,6059.21.1 (1.0-1.1)30,2218.51.0
 >2.82%1,7059.61.0 (1.0-1.1)11,3969.31.0 (1.0-1.0)13,7369.11.0
Age-adjusted rate1,7769.71.1 (1.1-1.2)15,0019.31.1 (1.0-1.1)44,0328.71.0

Of the 44 counties in the border region, 9 (20.5%) were urban, 6 (13.6%) were suburban, and 29 (65.9%) were rural. Cervical cancer incidence rates in urban areas were significantly higher in border counties and other counties in border states than in nonborder states, but with no statistical difference in urban rates comparing border counties and other counties in border states. Suburban rates in other counties in border states also were significantly higher than in nonborder states.

Women in border counties and border states with little migration/movement (eg, counties in which a relatively large proportion of residents in 2000 had lived in the same house as in 1995) had relatively high rates of cervical cancer incidence. The counties with limited migration/movement are predominantly rural; 19 of 22 of the counties with >59.2% of residents in the same house are rural, and all 22 of the counties with ≤2.8% of residents who migrated from outside the United States are rural. Lower education and poverty were both associated with cervical cancer incidence rates in exploratory analyses.

Table 3 shows age-adjusted incidence rates of cervical cancer in the 3 study areas by stage at diagnosis. The percentage of unknown/unstaged cervical cancer in border counties was 10.5%, compared with 7.6% in nonborder states. The rates of cervical cancer diagnosed at a late stage were higher in border counties and in other counties in border states, as compared with the rates for nonborder states, but there was no significant difference in late-stage rates, comparing border counties with other counties in border states. The crossover between early- and late-stage cervical incidence by age (Fig. 3) occurs about 5 years earlier for Hispanic women in border counties (age 45-59 years) than for women in nonborder states (age 50-64 years), and the cervical cancer incidence rate is higher in older Hispanic women in border counties than in non-Hispanic women.

thumbnail image

Figure 3. Invasive, age-specific cervical cancer carcinoma incidence rates (per 100,000 women) are shown by 15-year moving average age groups, stage, Hispanic ethnicity, and location in relation to the US-Mexico border, United States, 1998 to 2003. Data are from 37 population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance, Epidemiology, and End Results (SEER) program, meet high-quality data criteria (Watson et al13), and provide county-level incidence data. See Materials and Methods for a list of the 37 registries. Arizona data are from 1998 to 2002. These registries cover approximately 76.5% of the US population. The starting point for calculating the 15-year moving averages was the number of cases and the corresponding population for each of 19 5-year age groups (Census P25-1130). The moving average rate was calculated as the sum of the total cases for a 15-year period divided by the sum of population denominators for the same period. For example, the 15-year age group 10-24 years was calculated by adding data for the 5-year age groups 10-14 years, 15-19 years, and 20-24 years. Next, the 15-year age group 15-29 years was calculated by adding data for the 5-year age groups 15-19 years, 20-24 years, and 25-29 years. Then the 15-year age groups 20-34 was calculated by adding data for the 5-year age groups 20-24 years, 25-29 years, and 30-34 years. This process of advancing the 15-year age groups by 5 years was continued through the remaining age groups. For 1998-2000, SEER Summary Stage 1977 was used; for 2001-2003, SEER Summary Stage 2000 was used. Data for the 2 staging systems are combined because the differences observed in comparative analyses were minimal (Watson et al13). Early stage is the SEER Summary Stage localized category, and late stage is the sum of SEER Summary Stage regional and distant categories. Hispanic origin is not mutually exclusive from race categories (white, black, and Asian/Pacific Islander). Information was suppressed when the number of cases was <16 or when the majority of cases were in the 5-year age group 15-19 years.

Download figure to PowerPoint

Table 3. Cervical Carcinoma Rates* and Rate Ratios by Stage and Location in Relation to the US-Mexico Border, Using Nonborder States as the Reference Group, United States, 1998 to 2003
 Border CountiesOther Counties in Border StatesNonborder States (Reference Group)
StageCases, n (%)RateRate Ratio (95% CI)Cases, n (%)RateRate Ratio (95% CI)Cases, n (% )RateRate Ratio
  • n indicates number of cases; %, percentage of total cases for that subcategory; CI, confidence interval (calculated with Tiwari et al16 modification).

  • *

    Data are from 37 population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance, Epidemiology, and End Results (SEER) program, meet high-quality data criteria (Watson et al13), and provide county-level incidence data. See Materials and Methods for a list of the 37 registries. Arizona data is from 1998 to 2002. These registries cover approximately 76.5% of the US population.

  • Rates are per 100,000 women and age-adjusted to the 2000 US standard population (19 age groups, Census P25-1130).

  • For 1998 to 2000, SEER Summary Stage 1977 was used; for 2001 to 2003, SEER Summary Stage 2000 was used. Data for the two staging systems are combined, because the differences observed in comparative analyses were minimal (Watson et al13). Early stage is the SEER Summary Stage localized category, and late stage is the sum of SEER Summary Stage regional and distant categories.

  • §

    The rate ratio indicates that the rate is significantly different (P < .05) from the reference group (nonborder states).

Early stage833 (46.9)4.51.0 (0.9-1.0)7712 (51.4)4.71.0 (1.0-1.0)23,080 (52.4)4.61.0
Late stage756 (42.6)4.1§1.2 (1.1-1.3)6295 (42.0)3.9§1.2 (1.1-1.2)17,609 (40.0)3.41.0
Unknown/unstaged187 (10.5)1.0§1.6 (1.3-1.8)994 (6.6)0.61.0 (0.9-1.0)3343 (7.6)0.61.0

Table 4 shows incidence rates of cervical cancer in the 3 areas by histology. Squamous cell carcinomas were the most common, accounting for >70% of all cases in the 3 study areas. For the total population, the rates for squamous carcinoma in border and other counties in border states were significantly higher than in nonborder states, and the rate in border counties was significantly higher than in the other counties in border states. Squamous carcinoma rates were higher in Hispanics in border counties and in other counties in border states, compared with nonborder states. The pattern for Hispanic women was that squamous cancer rates were higher in border counties than in nonborder states. The pattern for non-Hispanic women was that squamous cancer rates were lower in border counties than in nonborder states.

Table 4. Cervical Carcinoma rates* by Histology, Ethnicity, and Location in Relation to the US-Mexico Border, Using Nonborder States as the Reference Group, United States, 1998 to 2003
 Border CountiesOther Counties in Border StatesNonborder States
 Cases, n (%)Rate (95% CI)Cases, n (%)Rate (95% CI)Cases, n (%)Rate (95% CI)
  • n indicates number of cases; %, percentage of total cases for that subcategory; CI, confidence interval (calculated with Tiwari et al16 modification).

  • *

    Data are from 37 population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance, Epidemiology, and End Results program, meet high-quality data criteria (Watson et al13), and provide county-level incidence data. See Materials and Methods for a list of the 37 registries. Arizona data is from 1998 to 2002. These registries cover approximately 76.5% of the US population.

  • Rates are per 100,000 women and age-adjusted to the 2000 US standard population (19 age groups, Census P25-1130).

  • Hispanic origin is not mutually exclusive from race categories (white, black, and Asian/Pacific Islander).

  • §

    The rate ratio indicates that the rate is significantly different (P < .05) from the reference group (nonborder states).

  • Information is suppressed because the number of cases is <16.

Total      
 Squamous carcinoma1310 (73.8)7.1 (6.7-7.5)§10,844 (72.3)6.7 (6.6-6.8)§31,656 (71.9)6.2 (6.2-6.3)
 Adenocarcinoma312 (17.6)1.7 (1.5-1.9)2876 (19.2)1.8 (1.7-1.8)§8394 (19.1)1.6 (1.6-1.7)
 Adenosquamous/glassy cell65 (3.7)0.4 (0.3-0.4)656 (4.4)0.4 (0.4-0.4)§1583 (3.6)0.3 (0.3-0.3)
 Small cell/neuroendocrine21 (1.2)0.1 (0.1-0.2)167 (1.1)0.1 (0.1-0.1)515 (1.2)0.1 (0.1-0.1)
 Other/unspecified68 (3.8)0.4 (0.3-0.5)458 (3.1)0.3 (0.3-0.3)§1884 (4.3)0.4 (0.3-0.4)
Hispanic      
 Squamous carcinoma829 (79.1)11.0 (10.2-11.8)§3883 (75.9)11.2 (10.9-11.6)§2816 (73.6)9.7 (9.3-10.1)
 Adenocarcinoma137 (13.1)1.8 (1.5-2.1)§820 (16.0)2.3 (2.2-2.5)637 (16.6)2.2 (2.0-2.4)
 Adenosquamous/glassy cell29 (2.8)0.4 (0.3-0.6)239 (4.7)0.7 (0.6-0.8)§135 (3.5)0.5 (0.4-0.6)
 Small cell/neuroendocrine40 (0.8)0.1 (0.1-0.2)40 (1.0)0.1 (0.1-0.2)
 Other/unspecified46 (4.4)0.6 (0.5-0.8)136 (2.7)0.4 (0.3-0.5)§198 (5.2)0.7 (0.6-0.8)
Non-Hispanic      
 Squamous carcinoma481 (66.1)4.6 (4.2-5.0)§6961 (70.4)5.6 (5.5-5.7)§28,840 (71.7)6.1 (6.0-6.1)
 Adenocarcinoma175 (24.0)1.7 (1.4-2.0)2056 (20.8)1.6 (1.6-1.7)7757 (19.3)1.6 (1.6-1.7)
 Adenosquamous/glassy cell36 (4.9)0.4 (0.2-0.5)417 (4.2)0.3 (0.3-0.4)1448 (3.6)0.3 (0.3-0.3)
 Small cell/neuroendocrine127 (1.3)0.1 (0.1-0.1)475 (1.2)0.1 (0.1-0.1)
 Other/unspecified22 (3.0)0.2 (0.1-0.3)§322 (3.3)0.3 (0.2-0.3)§1686 (4.2)0.3 (0.3-0.4)

Adenocarcinoma was the second most common cancer, accounting for 17% to 19% of all carcinomas. Rates for adenocarcinomas and adenosquamous/glassy cell carcinomas in other counties in border states were significantly higher than in nonborder states. Rates for other/unspecified carcinomas in other counties in border states were lower than in nonborder states. Adenocarcinoma rates were lower among Hispanic women in border counties, compared with Hispanic women in other counties in border states and nonborder states.

DISCUSSION

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

The results of this analysis confirm and expand upon prior reports that were based upon data from individual states on the incidence rates of cervical cancer in the US-Mexico border region. The use of NPCR and SEER data combined allowed for a more comprehensive examination of cervical cancer incidence in border and nonborder counties of border states than in earlier studies. Our finding of higher incidence rates among Hispanic women than among non-Hispanic women is consistent with prior studies from Texas,6 California,7 and the United States as a whole.17 In the current study, Hispanic women in the border states had a higher incidence of cervical cancer than non-Hispanic women in nonborder states. The incidence of cervical cancer among Hispanic women was not significantly different, comparing border counties with nonborder counties of border states. This pattern of risk is similar to that reported by Nasseri et al,7 and illustrates the need for incorporating geographic residence in research on disparities in cancer incidence.

One of the interesting findings in the present study was the pattern of cervical cancer incidence among black women. Whereas the cervical cancer incidence rates were lower among black women in border counties than in nonborder states, the opposite was true among white women. This pattern may be attributable to relatively high incidence rates among black women in other geographic localities, such as rural areas in the southern United States and in metropolitan areas outside the border region. However, the numbers of cervical cancer cases among black women in the border counties was small. In addition, the NPCR data for 1998 to 2003 did not include data from some states in the Southeast that have a relatively high proportion of blacks.

The incidence of cervical cancer increases with age, from the early 20s until the 40 to 44-year age category, although there may be some differences across racial and ethnic groups. This is the active childbearing age for most women, when they receive prenatal care and regular Papanicolaou (Pap) testing. The continued higher cervical cancer rates in border counties after ages 40 to 44 years are consistent with studies suggesting that cervical cancer rates in border counties are a reflection of the human papillomavirus (HPV) epidemiology found in many Latin American countries, including Mexico. Data from HPV studies in the US-Mexico border region and other parts of Latin American have suggested that in addition to a peak prevalence of HPV infection in the younger age groups, Hispanic women may experience a secondary peak at older ages, such as age 50 to 60 years.18, 19 The dual peak, combined with a lack of screening among older women in border counties, may explain why older women continue to have such high rates of cervical cancer. The decline among older women in nonborder counties may be explained by better access to screening and follow-up. Older Hispanic women, especially women who live in border counties, may be less likely to undergo routine Pap testing3 and may be less likely to be compliant in receiving appropriate follow-up after an abnormal Pap test.20 Access to Pap testing is significant in prevention of cervical cancer. Our data show that up to age 30 to 34 years, the incidence rates are similar throughout the United States. However, in the “post reproductive” age group, women with lower socioeconomic status may have less access and thus get less screening and more cervical cancer.

A potentially important finding in the present study was the relatively high rates of cervical cancer among urban and suburban women in border states, as compared with the rates for urban and suburban women in nonborder states. The cervical cancer incidence rate among rural women in the border counties was also higher, although differences with nonborder states did not achieve statistical significance. These differences in cervical cancer incidence rates by geographic locality may reflect differences in access to and acceptance of the healthcare system, including access to culturally appropriate and sensitive cervical cancer screening. Previous authors have emphasized that the percentage of persons in a county living in the same house over 5 years is a potential marker for individuals more likely to have an established physician to make screening recommendations and also to have an established social network to encourage screening.21 Future studies should examine factors not addressed in the current study, such as structural barriers faced by some women, distance required to travel to screening facilities, lack of transportation, or assistance with child care.22, 23

Of particular concern are cases diagnosed at a late stage, because cervical cancer is much more likely to be successfully treated if detected early. The rates of cervical cancer diagnosed at a late stage were relatively high in the border counties and nonborder counties of border states, as compared with the rates in nonborder states.

Previous analyses have reported that Hispanic women have higher rates of adenocarcinoma than black or white women.24 This analysis had similar findings in the nonborder states where there were relatively large numbers, but some differences were observed in border counties.

With respect to limitations of the current study, the small number of women in rural areas of the border counties likely limited the statistical power to detect differences. A further issue is that some misclassification likely occurred as a result of the lack of individual-level data for factors such as education and income. In addition, some states have concluded that the US Department of Agriculture county-level rural-urban codes should not be used in their state-level analyses, because counties are relatively large geographic areas that can include both rural and urban census tracts. As an alternative, these states recommend that rural-urban codes be assigned at the census tract level. For purposes of the current study, the county is the smallest unit of geography available in the existing computerized dataset for all states, a fact that likely introduced some misclassification of rural/urban residence.

A further issue is that the histologic analysis was limited to cancer cases with microscopic confirmation. This restriction helps assure that case identification is accurate. However, limiting the analysis to microscopically confirmed cases resulted in the exclusion of 1.1% of all cases, suggesting that this exclusion was not an important study limitation.

Although data on HPV infection were unavailable for the current study, HPV is thought to be responsible for nearly all cervical cancers. Transmission of HPV is primarily through sexual contact.25, 26 Oncogenic HPV types 16 and 18 account for about 70% of cervical cancers,27, 28 and these types are included in the recently recommended quadrivalent HPV vaccine. Among females who were not exposed to HPV 16 and HPV 18, the vaccine demonstrated high efficacy in preventing future HPV 16 and 18 cervical precancer. Vaccination efforts might be expected to be particularly beneficial among the higher-risk subgroups identified in this analysis. As vaccination efforts increase in both Mexico and the United States, there is increasing concern that the vaccine at its current price may not reach those at highest risk of developing precancerous or cancer stages, especially immigrant populations.

In summary, the results from this study add to the literature on disparities in cervical cancer incidence in the US-Mexico border counties and other geographic localities within border states, including differences in incidence by Hispanic ethnicity, race, rural/nonrural residence, and other factors. The association between the areas in which women live and the incidence of cervical cancer may be attributable to several factors, including screening utilization and access to healthcare.2, 3, 22, 23 Of particular concern are cases diagnosed at a late stage, because cervical cancer is highly preventable through routine screening and much more likely to be successfully treated if diagnosed at an early stage.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Health Resources Services Administration. Assuring a Healthy Future Along the U.S.-Mexico Border. A HRSA Priority. Rockville, MD: US Department of Health and Human Services; 2000.
  • 2
    Byrd TL,Cardenas VM,Pauli A. Self-reported use of cervical cancer screening among Hispanic women of 2 US-Mexico border counties. J Womens Cancer. 2001; 3: 4955.
  • 3
    Coughlin SS,Uhler RJ,Richards T,Wilson KM. Breast and cervical cancer screening practices among Hispanic and non-Hispanic women residing near the United States-Mexico border, 1999–2000. Fam Comm Health. 2003; 26: 130139.
  • 4
    Tsui J,Saraiya M,Thompson T,Dey A,Richardson L. Cervical cancer screening among foreign-born women in the United States by birthplace and duration. J Womens Health. 2007; 16: 14471457.
  • 5
    Seeff LC,McKenna MT. Cervical cancer mortality among foreign-born women living in the United States, 1985–1996. Cancer Detect Prev. 2003; 27: 203208.
  • 6
    WilliamsMA,MokryB,RisserD, et al, eds. Cervical Cancer in Texas 2006. Austin, TX: Texas Department of State Health Services; 2006. Available at: http://www.dshs.state.tx.us/tcr/report_cervical.shtm. Accessed on July 22, 2008.
  • 7
    NasseriK,CressRD,LeiserowitzG, eds. Cervical Cancer in California, 2006. Santa Barbara, CA: Public Health Institute, Tri-Counties Cancer Surveillance Program; 2006.
  • 8
    Bernard VB,Johnson CJ,Thompson TD, et al. Examining the association between socio-economic factors and human papillomavirus-related cancers. Cancer. 2008; 113(10 suppl ): 2910-2918.
  • 9
    U.S.-Mexico Border Health Commission. U.S.-Mexico Border Environmental Health Initiative. Available at: http://borderhealth.cr.usgs.gov/links.html. Accessed on July 22, 2008.
  • 10
    Combs S. The Border: On the Brink. Texas State Comptroller of Public Accounts. Available at: http://www.window. state.tx.us/specialrpt/brink/. Accessed on July 22, 2008.
  • 11
    LandScan™ Global Population Database, North America 2002. Oak Ridge, TN: Oak Ridge National Laboratory. Available at: http://www.ornl.gov/landscan/ Accessed on July 22, 2008.
  • 12
    US Department of Agriculture Economic Research Service. 2003 Rural-Urban Continuum Codes. Available at: http://www.ers.usda.gov/Data/RuralUrbanContinuumCodes/ Accessed on July 22, 2008.
  • 13
    Watson M,Saraiya M,Ahmed F, et al. Using population-based cancer registry data to assess the burden of HPV-associated cancers in the United States: overview of methods. Cancer. 2008; 113(10 suppl ): 28412854.
  • 14
    Howe HL,Jamison PM,Havener L, et al. Site-Specific Comparison of Summary Stage 1977 and Summary Stage 2000 Coding. North American Association of Central Cancer Registries 2005. Available at: http://www.naaccr.org/index.asp?Col_SectionKey=11&Col_ContentID=397 Accessed on July 22, 2008.
  • 15
    Rothman KJ,Greenland S. Modern Epidemiology. 2nd ed. Philadelphia, PA: Lippencott-Raven; 1998.
  • 16
    Tiwari RC,Clegg LX,Zou Z. Efficient interval estimation for age-adjusted cancer rates. Stat Methods Med Res. 2006; 15: 547569.
  • 17
    U.S. Cancer Statistics Working Group. United States Cancer Statistics: 2004 Incidence and Mortality. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute; 2007.
  • 18
    Giuliano AR,Papenfuss MR,Denman CA, et al. Human papillomavirus prevalence at the USA-Mexico border among women 40 years of age and older. Int J STD AIDS. 2005; 16: 247251.
  • 19
    Franceschi S,Herrero R,Clifford GM, et al. Variations in the age-specific curves of human papillomavirus prevalence in women worldwide. Int J Cancer. 2006; 1: 26772684.
  • 20
    Greenberg H,Duarte-Gardea M,Quezada OR. Implications regarding atypical squamous cells of undetermined significance among women residing in a US-Mexico border city. Int J Gynecol Cancer. 2006; 16: 10141016.
  • 21
    Roos LL,Walld R. Neighborhood, family and health care. Can J Public Health. 2007; 98.( suppl 1): S54S61.
  • 22
    Tortolero-Luna G,Glober GA,Villarreal R,Palos G,Linares A. Screening practices and knowledge, attitudes, and beliefs about cancer among Hispanic and non-Hispanic white women 35 years old or older in Nueces County, Texas. J Natl Cancer Inst Monogr 1995;( 18): 4956.
  • 23
    Suarez L,Roche RA,Nichols D,Simpson DM. Knowledge, behavior, and fears concerning breast and cervical cancer among older low-income Mexican-American women. Am J Prev Med. 1997; 13: 137142.
  • 24
    Saraiya M,Ahmed F,Krishnan S,Richards TB,Unger ER,Lawson HW. Cervical cancer incidence in a prevaccine era in the United States, 1998-2002. Obstet Gynecol. 2007; 109(2 pt 1): 360370.
  • 25
    Giuliano AR,Papenfuss M,Schneider A, et al. Risk factors for high-risk type human papillomavirus infection among Mexican-American women. Cancer Epidemiol Biomarkers Prev. 1999; 8: 615620.
  • 26
    Bosch FX,de Sanjose S. Chapter 1: Human papillomavirus and cervical cancer—burden and assessment of causality. J Natl Cancer Inst Monogr. 2003;( 31): 313.
  • 27
    Smith JS,Linsay L,Hoots B, et al. Human papillomavirus type distribution in invasive cervical cancer and high-grade cervical lesions: a meta-analysis update. Int J Cancer. 2007; 121: 621632.
  • 28
    Clifford GM,Smith JS,Plummer M, et al. Human papillomavirus types in invasive cervical cancer worldwide: a meta-analysis. Br J Cancer. 2003; 88: 6373.