SEARCH

SEARCH BY CITATION

Keywords:

  • Asians and Pacific Islanders;
  • blacks;
  • colorectal cancer;
  • Hispanics;
  • incidence;
  • poverty

Abstract

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

BACKGROUND.

Few studies of colorectal cancer incidence by rural, suburban, and metropolitan residence have been published.

METHODS.

The authors examined colorectal cancer incidence among men and women in U.S. counties classified as rural, suburban, and metropolitan for the period 1998–2001. They examined rural/suburban/metropolitan differences in incidence by age, race, Hispanic ethnicity, stage at diagnosis, histology, and percentage of the total county population below the poverty level, using data from the CDC's National Program of Cancer Registries, the NCI's Surveillance, Epidemiology, and End Results Program, and the 2000 U.S. Census.

RESULTS.

A total of 495,770 newly diagnosed or incident cases of colorectal cancer were included in this analysis (249,919 among men and 245,851 among women). Over the period 1998–2001, the colorectal cancer incidence rates among men tended to be lower among those who resided in rural areas, for each of the subgroups examined, with the exception of Asians and Pacific Islanders and those living in more affluent counties. Among women aged 75 years and older, the colorectal cancer incidence rates tended to be lower among rural than metropolitan or suburban residents, though the differences were slight. In multivariate analysis, the incidence of colorectal cancer was higher in metropolitan, suburban, and rural areas for blacks than that for whites (incidence rate ratios [RR] = 1.12, 1.07, and 1.06, respectively, all P < 0.015).

CONCLUSIONS.

This study suggests that black men who reside in metropolitan areas have a higher risk of colorectal cancer than black men who reside in rural areas. This finding suggests the need for diverse approaches for reducing colorectal cancer when targeting rural compared with metropolitan areas. Cancer 2006. © 2006 American Cancer Society.

Previous studies have suggested that there may be important rural/nonrural differences in colorectal cancer incidence and mortality among men and women in the United States.1–4 Such geographic variation may be partly due to differences in colorectal cancer screening because routine screening can reduce both mortality from the disease and morbidity over time.5 Residents of poor or medically underserved areas, such as some rural areas of the United States, may face important barriers to screening.6–8 Likewise, residents of some inner city areas may experience barriers. Atlases of cancer mortality and incidence indicate that colorectal cancer rates are highest in the Northeast and Midwest regions of the United States. These regions cover large geographic areas and include major metropolitan, suburban, and rural areas,1, 2, 4 thus making it difficult to identify rural/nonrural differences using these data. An analysis of colorectal cancer mortality in the Appalachian region of the United States from 1969 to 1999 showed that death rates had declined in the more recent years, but in 1999 the rates for white males and white females were still significantly higher in Appalachia than in the rest of the country.9 The higher death rates in Appalachia, which is disproportionately rural, may be due to the cancers being diagnosed at a later stage when survival is relatively poor, to poorer treatment, or to other factors. Lower screening rates may partly account for the higher than expected rates of late-stage colorectal cancer in some areas. Death certificate data do not include information about the decedents' stage of colorectal cancer at diagnosis or the histology, but incidence studies can provide such information and attempt to clarify rural/nonrural differences in the disease. However, a detailed examination of colorectal cancer incidence in metropolitan, suburban, and rural areas of the United States, with broad coverage by geographic location, race, and ethnicity, has not been published.

To help clarify these relationships, we examined colorectal cancer incidence among men and women in U.S. counties classified as rural, suburban, and metropolitan. We assessed incidence among rural, suburban, and metropolitan residents by age, race, Hispanic ethnicity, stage at diagnosis, histology, and poverty indication.

MATERIALS AND METHODS

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

For this analysis, we used data from the National Program of Cancer Registries (NPCR) reported to the Centers for Disease Control and Prevention (CDC) as of January 31, 2004, for registries that met the U.S. Cancer Statistics Publication Standard for data quality for all cancer sites combined.10 We also used data from the November 2003 submission to the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) Program.11 Data from Alaska, Hawaii, Illinois, and Minnesota were excluded because county-level data on rural/nonrural residence are not available for these states. SEER data from the Atlanta Metropolitan area were excluded because they included no rural residents. In California, stage data were not submitted to NPCR. Therefore, we excluded data from California from analyses involving stage of cancer diagnosis because stage data were available only from SEER registries in the state, and those SEER areas included no rural residents. The final analytic dataset included incidence data for 35 states and the District of Columbia, representing about 80% of the U.S. population. We combined data for invasive cancer diagnoses between 1998 and 2001, inclusive, among adults aged 20 years or older at the time of diagnosis.

We computed incidence rates and 95% confidence intervals (CIs) for the period 1998–2001. We compared the incidence of colorectal cancer in metropolitan, suburban, and rural counties of the United States, using the U.S. Department of Agriculture's urban/rural continuum codes that are based on information from the 2000 U.S. Census.12 Codes 0–3 correspond to 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. According to their county of residence, persons were assigned a county-level rural–urban continuum code and categorized as residents of a) metropolitan areas (codes 0–3), b) suburban areas (codes 4–5), or c) rural areas (codes 6–9).

We examined colorectal cancer incidence rates in metropolitan, suburban, and rural areas by age categories, gender, race, ethnicity, stage at cancer diagnosis, histology, and an area-based indicator of poverty. We categorized patient age at diagnosis in 5-year intervals for ages 20–84 years, with a final category of >85 years. Among rural, suburban, and metropolitan residents, incidence rates by residence were examined for black, white, and Asian/Pacific Islander men and women. We also examined differences in incidence by residence and Hispanic ethnicity. Data were not available for Asian/Pacific Islander or Hispanic subgroups. Incidence rates were not calculated separately for American Indians and Alaska Natives, but these persons were included in overall analyses.

Stage at diagnosis was categorized according to SEER summary stage (localized, regional, distant, or unstaged). Analyses by histologic type included the following ICD-O-3 groupings13: adenocarcinoma (histologic codes 8140–8147, 8160–8162, 8180–8221, 8250–8506, 8520–8550, 8560, 8570–8573, 8940–8941) including papillary carcinoma not otherwise specified (8050); nonadenocarcinoma (all other types except ‘unspecified’); and unspecified (not otherwise specified).

Poverty was defined as the percentage of the total county population below the federal poverty level as reported in the 2000 U.S. Census.14 We categorized counties as <10%, 10% to <20%, or ≥20% of the total county population below poverty level. Area-based measures of socioeconomic position have been widely used by other investigators to characterize important aspects of the social environment.15–17

All rates, except age-specific rates, were adjusted by the direct method to the 2000 U.S. standard population by 5-year age groups. Rates were compared only if there were at least 16 cases in each cell. Ninety-five percent CIs were estimated following a gamma distribution.18 The rate ratio test was used to compare incidence rates in rural areas to the other 2 groups. P < .05 indicated statistical significance. No adjustment was made for multiple comparisons.

We used negative binomial modeling techniques to examine the effects of age, gender, race (or Hispanic ethnicity), year of diagnosis, and percentage below the poverty level, and to adjust for these factors while examining metropolitan, suburban, and rural differences in the colorectal cancer incidence rates.19 Negative binomial models19 were used rather than Poisson models because of large amounts of overdispersion in the Poisson models. Two-way interactions between area (metropolitan, suburban, or rural) and each of the other covariates were examined to determine if the area effect was similar across levels of these variables. A statistically significant effect modification was observed between area (metropolitan, suburban, or rural) and both poverty level and gender. However, this effect was primarily due to the large sample size and power to detect small differences; there was not meaningful variation in the incidence rate ratios among groups. Therefore, these interactions were omitted from the model. Race and ethnicity could not be included in the same model because population data are not available for some race/ethnicity combinations, as race and Hispanic ethnicity were recorded separately. As a result, separate models were fit for these variables. Statistical testing in all models was performed using the likelihood ratio test.

The CDC Institutional Review Board approved this study.

RESULTS

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

A total of 495,770 newly diagnosed cases of colorectal cancer were included in this analysis (249,919 among men and 245,851 among women). Age-specific colorectal cancer incidence rates by metropolitan, suburban, and rural residence among men and among women, for the period 1998–2001, are shown in Tables 1 and 2, respectively. Among rural, suburban, and metropolitan residents, the rates increased with advancing age and were highest among men in the oldest age categories. Among men aged 75–84 years, the colorectal cancer incidence rates were lower among those who resided in rural areas than among those who resided in other areas. Among women aged 75 and older, the colorectal cancer incidence rates tended to be lower in rural than metropolitan areas; the differences were slight, but statistically significant.

Table 1. Age-Specific Colorectal Cancer Incidence Rates* by Metropolitan, Suburban, Rural Resident Status among Men, United States, 1998–2001
Age (yr)TotalMetropolitanSuburbanRural
CasesRate95% CICasesRate95% CICasesRate95% CICasesRate95% CI
  • *

    Rates are per 100,000 persons.

  • Data are from selected population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance Epidemiology and End Results Program and meet high-quality data criteria: Alabama, Arizona, California, Colorado, Connecticut, District of Columbia, Florida, Idaho, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maine, Massachusetts, Michigan, Missouri, Montana, Nebraska, New Jersey, New Mexico, New York, North Carolina, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode Island, South Carolina, Texas, Utah, Vermont, Washington, West Virginia, Wisconsin, Wyoming. These registries cover ˜80% of the U.S. population. Alaska, Atlanta (Georgia), Hawaii, Illinois, and Minnesota were excluded (see Methods).

20–242380.8(0.7–0.9)1920.7(0.6–0.9)190.8(0.5–1.3)261.0(0.7–1.5)
25–296262.0(1.8–2.2)5181.9(1.8–2.1)532.6(2.0–3.5)552.3(1.7–3.0)
30–3413144.0(3.7–4.2)11344.0(3.8–4.2)874.1(3.3–5.1)933.7(3.0–4.5)
35–3928728.0(7.7–8.3)24077.9(7.6–8.2)2008.4(7.3–9.6)2628.8(7.7–9.9)
40–44539215.3(14.8–15.7)450015.2(14.7–15.6)39516.2(14.6–17.9)49215.7(14.3–17.1)
45–49963130.7(30.1–31.4)794830.5(29.9–31.2)73732.7(30.4–35.2)93831.9(29.9–34.0)
50–5415,86258.7(57.8–59.7)13,05458.5(57.5–59.5)125563.5(60.1–67.2)154259.3(56.4–62.4)
55–5920,758101.0(99.6–102.3)16,948100.9(99.4–102.4)1600102.7(97.7–107.9)2194102.6(98.3–107.0)
60–6426,883164.1(162.2–166.1)21,638164.6(162.4–166.8)2220168.4(161.4–175.5)3005161.1(155.4–167.0)
65–6935,119246.4(243.8–249.0)28,285248.0(245.1–250.9)2873245.7(236.8–254.9)3944240.0(232.6–247.7)
70–7441,428326.8(323.7–330.0)33,328328.1(324.6–331.7)3451329.1(318.2–340.3)4633321.2(312.1–330.6)
75–7939,972404.1(400.2–408.1)32,499407.0(402.6–411.4)3311413.0(399.1–427.3)4142382.2(370.6–394.0)
80–8428,401471.5(466.0–477.0)23,020474.5(468.4–480.7)2349482.0(462.7–501.8)3024449.1(433.2–465.4)
≥8521,423531.6(524.5–538.8)17,348538.5(530.5–546.6)1639515.0(490.3–540.5)2430504.0(484.1–524.4)
Table 2. Age-Specific Colorectal Cancer Incidence Rates* by Metropolitan, Suburban, Rural Resident Status among Women, United States, 1998–2001
Age (yr)TotalMetropolitanSuburbanRural
CasesRate95% CICasesRate95% CICasesRate95% CICasesRate95% CI
  • *

    Rates are per 100,000 persons.

  • Data are from selected population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance Epidemiology and End Results Program and meet high-quality data criteria (see Table 1 footnote for list of registries). These registries cover ∼80% of the U.S. population.

20–242290.8(0.7–0.9)1830.7(0.6–0.8)170.8(0.5–1.3)291.3(0.9–1.9)
25–295291.7(1.6–1.9)4381.7(1.5–1.8)382.0(1.4–2.8)512.3(1.7–3.0)
30–341,2043.7(3.5–3.9)10223.6(3.4–3.8)783.9(3.1–4.9)1044.3(3.5–5.2)
35–3925907.2(6.9–7.4)22137.2(6.9–7.5)1516.4(5.5–7.6)2257.7(6.7–8.7)
40–44505814.1(13.7–14.5)423914.0(13.6–14.4)37315.3(13.8–17.0)44514.4(13.1–15.8)
45–49838525.9(25.4–26.5)702826.0(25.4–26.6)59226.3(24.2–28.5)76326.4(24.6–28.3)
50–5412,75045.2(44.4–46.0)10,62045.1(44.3–46.0)91745.9(43.0–49.0)120746.5(43.9–49.2)
55–5915,28269.3(68.2–70.4)12,56169.3(68.1–70.5)116070.8(66.8–75.0)155470.1(66.7–73.7)
60–6419,881109.6(108.0–111.1)16,161109.8(108.1–111.5)1548109.3(103.9–114.9)2162110.3(105.7–115.0)
65–6927,319164.1(162.2–166.1)22,051163.7(161.5–165.9)2181165.0(158.1–172.1)3069169.3(163.4–175.4)
70–7435,693221.5(219.2–223.8)29,030221.6(219.0–224.1)2898227.0(218.8–235.4)3753220.7(213.7–227.9)
75–7940,678288.0(285.2–290.8)33,467290.5(287.4–293.6)3147283.6(273.8–293.7)4042275.2(266.8–283.9)
80–8436,267358.2(354.5–361.9)29,613361.4(357.3–365.6)2797348.4(335.6–361.5)3845346.9(336.0–358.0)
≥8539,986410.0(406.0–414.1)32,444414.5(410.0–419.0)3089400.5(386.5–414.8)4430390.8(379.3–402.4)

Table 3 shows age-adjusted colorectal cancer incidence rates among men, for the period 1998–2001, by residence, race, ethnicity, stage, and percentage below poverty level. The colorectal cancer incidence rates among all men, white men, and black men were lower among those who resided in rural areas than those in metropolitan areas (Table 3). The opposite was true for Asian and Pacific Islander men; rates were higher in rural than in metropolitan areas, although the number of cases was relatively small (Table 3).

Table 3. Age-Adjusted Colorectal Cancer Incidence Rates* by Race, Ethnicity, Stage, Poverty Status, and Metropolitan, Suburban, Rural Resident Status among Men Aged 20 Years and Older, United States, 1998–2001
CharacteristicTotalMetropolitanSuburbanRural
CasesRate95% CICasesRate95% CICasesRate95% CICasesRate95% CI
  • No adjustments were made for multiple comparisons to control the Type I error-rate. Statistically significant differences should be interpreted with caution.

  • *

    Rates are per 100,000 persons and are age-adjusted to the 2000 U.S. standard population.

  • Data are from selected population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance Epidemiology and End Results Program and meet high-quality data criteria (see Table 1 footnote for list of registries). These registries cover ∼80% of the U.S. population. California was excluded from analysis by stage because of missing stage data. Hispanic origin is not mutually exclusive from race categories.

  • P < 0.05 for testing for differences between metropolitan vs. rural areas.

  • §

    P < 0.05 for testing for differences between suburban vs. rural areas.

Overall§249,91992.4(92.1–92.8)202,81992.8(92.4–93.2)20,18993.9(92.6–95.3)26,78090.1(89.1–91.2)
Race
 White§221,58191.9(91.5–92.3)177,34192.3(91.9–92.8)18,96093.7(92.4–95.1)25,18389.8(88.7–90.9)
 Black20,33897.4(96.0–98.8)18,22198.3(96.8–99.8)94395.3(89.2–101.9)115987.8(82.7–93.1)
 Asian/Pacific Islander467363.7(61.7–65.7)457363.3(61.3–65.3)6193.1(67.7–127.6)3799.2(67.9–142.3)
Ethnicity
 Hispanic13,81176.1(74.7–77.5)12,65676.3(74.9–77.8)53279.2(72.1–87.0)61974.8(68.7–81.4)
 Non-Hispanic§236,09493.6(93.2–94.0)190,15294.2(93.7–94.6)19,65494.4(93.1–95.8)26,16190.6(89.5–91.7)
Stage
 Localized§77,65233.0(32.8–33.3)61,38233.3(33.1–33.6)698133.6(32.8–34.4)924931.5(30.9–32.2)
 Regional84,32335.8(35.5–36.0)66,94736.3(36.0–36.5)731435.1(34.3–35.9)10,02534.2(33.6–34.9)
 Distant36,38315.3(15.1–15.4)28,81915.4(15.2–15.6)322715.3(14.8–15.9)432114.7(14.2–15.1)
 Unstaged22,29110.0(9.8–10.1)17,3979.9(9.8–10.1)202210.2(9.7–10.6)283410.1(9.7–10.5)
Percentage below poverty level
 <1078,59294.7(94.0–95.4)68,91294.7(94.0–95.4)452295.5(92.7–98.3)515894.4(91.8- 97.0)
 10 to <20§149,00591.7(91.2–92.2)118,97891.7(91.2–92.3)13,53094.0(92.4–95.6)16,49789.9(88.5–91.3)
 ≥2022,19191.7(90.5–92.9)14,92993.8(92.3–95.3)213790.3(86.4–94.3)512586.8(84.4–89.2)

Among Asian and Pacific Islander women, rates were higher in rural than in metropolitan areas (Table 4). Among non-Hispanic women, rates were lower in rural than in metropolitan areas (Table 4).

Table 4. Age-Adjusted Colorectal Cancer Incidence Rates* by Race, Ethnicity, Poverty Status, and Metropolitan, Suburban, and Rural Resident Status among Women Aged 20 Years and Older, United States, 1998–2001
CharacteristicTotalMetropolitanSuburbanRural
CasesRate95% CICasesRate95% CICasesRate95% CICasesRate95% CI
  • No adjustments were made for multiple comparisons to control the Type I error-rate. Statistically significant differences should be interpreted with caution.

  • *

    Rates are per 100,000 persons and are age-adjusted to the 2000 U.S. standard population.

  • Data are from selected population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance Epidemiology and End Results Program and meet high quality data criteria (see Table 1 footnote for list of registries). These registries cover ∼80% of the U.S. population. California was excluded from analysis by stage because of missing stage data. Hispanic origin is not mutually exclusive from race categories.

  • P < 0.05 for testing for differences between metropolitan vs. rural areas.

  • §

    P < 0.05 for testing for differences between suburban vs. rural areas.

Overall245,85166.8(66.6–67.1)201,07067.1(66.8–67.4)18,98666.8(65.8–67.8)25,67966.4(65.5–67.2)
Race
 White215,33665.9(65.6–66.2)173,63166.0(65.7–66.4)17,76766.6(65.6–67.6)23,84665.6(64.7–66.4)
 Black22,95374.8(73.8–75.8)20,57375.1(74.1–76.2)97171.3(66.9–76.0)139974.1(70.2–78.1)
 Asian/Pacific Islander§451648.2(46.7–49.7)439647.9(46.4–49.4)5753.7(37.7–76.3)61103.8(76.8–139.0)
Ethnicity
 Hispanic12,07651.4(50.4–52.3)11,23451.7(50.7–52.7)36846.8(42.0–52.0)47451.5(46.9–56.4)
 Non-Hispanic233,76067.9(67.6–68.1)189,82468.2(67.9–68.5)18,61667.3(66.4–68.3)25,20466.7(65.9–67.6)
Stage
 Localized§73,64123.0(22.8–23.2)58,54323.1(22.9–23.3)646923.6(23.0–24.1)860022.6(22.1–23.1)
 Regional85,54826.8(26.6–26.9)68,40927.0(26.8–27.2)711725.9(25.3–26.5)999126.4(25.9–26.9)
 Distant34,70411.0(10.9–11.1)27,86211.1(11.0–11.3)291310.8(10.5–11.3)391110.6(10.3–10.9)
 Unstaged23,8527.1(7.0–7.2)18,9997.2(7.1–7.3)19386.6(6.4–7.0)28777.0(6.8–7.3)
Percentage below poverty level
 <1077,23168.5(68.0–69.0)67,71068.2(67.7–68.7)437969.5(67.4–71.6)514271.0(69.1–73.1)
 10 to <20§145,93366.3(65.9–66.6)117,58366.2(65.9–66.6)12,77467.4(66.2–68.6)15,57665.6(64.5–66.6)
 ≥20§22,57166.5(65.7–67.4)15,77768.6(67.5–69.6)183357.9(55.3–60.7)496164.3(62.5–66.1)

Further analyses were carried out to examine age-adjusted colorectal cancer incidence rates by rural/suburban/metropolitan residence, race, ethnicity, stage, and percentage below poverty level by histologic type (adenocarcinoma, nonadenocarcinoma, unspecified) and gender (results not shown). The majority of cases are adenocarcinomas, and the results mirrored those observed for all histologies combined. The small number of nonadenocarcinoma cases precluded the identification of meaningful differences among subgroups.

In multivariate analysis, the incidence of colorectal cancer was higher in all areas for blacks than for whites (Table 5). In contrast, the incidence of colorectal cancer was lower among Asians and Pacific Islanders than among whites in metropolitan and suburban areas, although the number of cases was relatively small. Hispanic ethnicity was associated with lower incidence (RR = 0.81, 95% CI = 0.79–0.82, P < .001). There was no interaction between Hispanic ethnicity and area (metropolitan, suburban, or rural).

Table 5. Negative Binomial Model Predicting Colorectal Cancer Incidence among Men and Women Aged 20 years and Older, United States, 1998–2001*
CharacteristicLikelihood ratio χ2DFPIncidence rate ratio95% CI
  • *

    Data are from selected population-based cancer registries that participate in the National Program of Cancer Registries and/or the Surveillance Epidemiology and End Results Program and meet high-quality data criteria (see Table 1 footnote for list of registries). These registries cover ∼80% of the U.S. population.

  • Race by area interaction included in model (P < 0.0001).

Age
 35–44 vs. <354946.856<0.00015.134.89–5.40
 45–54 vs. <3518.4717.63–19.36
 55–64 vs. <3548.8246.64–51.12
 65–74 vs. <35102.8298.25–107.60
 75–84 vs. <35159.69152.57–167.14
 >85 vs. <35199.17190.04–208.70
Gender
 Female vs. Male625.211<0.00010.750.74–0.76
Race
 Metropolitan609.446<0.0001  
  Black vs. White1.121.09–1.14
  API vs. White0.640.62–0.67
 Suburban  
  Black vs. White1.071.02–1.13
  API vs. White0.810.67–0.97
 Rural  
  Black vs. White1.061.01–1.11
  API vs. White1.200.98–1.46
Area
 White42.256<0.0001  
  Metropolitan vs. Rural1.010.99–1.03
  Suburban vs. Rural1.010.99–1.04
 Black  
  Metropolitan vs. Rural1.061.01–1.11
  Suburban vs. Rural1.020.96–1.09
 API  
  Metropolitan vs. Rural0.540.44–0.67
  Suburban vs. Rural0.680.52–0.90
Year
 1999 vs. 199836.163<0.00010.970.95–0.99
 2000 vs. 19980.960.94–0.98
 2001 vs. 19980.930.91–0.95
Percent below poverty level
 10-<20 vs. <104.7120.09470.980.97–1.00
 >20 vs. <100.980.96–1.00

DISCUSSION

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

The results of this study suggest that, after adjustment for age, black men who reside in metropolitan areas of the United States are at greater risk of colorectal cancer than black men who reside in rural areas. In contrast, Asians and Pacific Islanders who live in rural areas of the United States have a higher risk than Asians and Pacific Islanders in metropolitan areas; the number of cases, however, was relatively small. The reasons for the different patterns in rural versus metropolitan incidence rates for blacks and for Asians and Pacific Islanders are not known. These disparities in colorectal cancer incidence by race and residence may be partly due to an interaction of race and geographic variation in preventive practices such as physical activity, diet, and colorectal cancer screening, although large geographic differences in stage at diagnosis were not observed. Nevertheless, prior studies have shown that colorectal cancer screening rates are lower among rural men and women than those residing in large metropolitan areas, and that colorectal cancer screening rates are low relative to those of other screening tests such as mammography.7, 8 Colorectal cancer screening may increase or decrease colorectal cancer incidence. Rural residence has been inversely associated with both preventive behaviors and socioeconomic position.7, 8 Many metropolitan areas of the United States, however, are socioeconomically diverse and include both affluent areas and inner city areas where many households have incomes below the poverty level.

Rural/nonrural differences in colorectal cancer incidence and mortality may reflect geographic differences in access to health care.17 Access to professional advice about risk factor modification (e.g., counseling by a primary care physician to exercise more) and access to screening and appropriate follow-up care for colorectal cancer are complex issues. For example, one factor might be the distance a patient has to travel from a sparsely populated rural area to a facility providing preventive health care services or follow-up care. Access may also involve other considerations that are important in both rural and nonrural areas, e.g., health insurance, hours that clinics are open, language barriers, and availability of culturally appropriate and sensitive health care.

The current analysis is limited by the approach that was taken to define rural, suburban, or metropolitan residence. Although the urban/rural continuum codes are commonly used in epidemiologic analysis, rural/nonrural residence based on county-level information may be less satisfactory than sub-county units of analysis. The geographic size of counties varies widely. Rural populations may exist within the boundaries of metropolitan areas, and metropolitan areas may overlap geopolitical boundaries and extend into areas classified as rural or nonmetropolitan. Consequently, some misclassifications likely occurred in the current analysis. The observed differences in rural versus metropolitan colorectal cancer incidence would likely have been greater if the analyses had been based on units of analysis smaller than counties. We were unable to define metropolitan, suburban, and rural status at a smaller geographic level because the county was the smallest geographic unit of analysis in our dataset. Other limitations include small numbers within some subgroups of interest and wide confidence intervals. Nevertheless, the current study had several strengths including the fact that the data covered 80% of the U.S. population.

The proportion of colorectal cancer cases that are histologically confirmed might introduce bias into this analysis. Misclassification may occur through the inclusion of cases that are not histologically confirmed. However, more than 96% of the cases included in this analysis were histologically confirmed, and so, any biases should be small. We included all colorectal cancer cases in the descriptive analyses and then stratified some analyses by histological type (adenocarcinoma versus nonadenocarcinoma).

Analyses by Hispanic ethnicity may have been biased because of variation across states (including states that are predominately rural or predominately nonrural) in the accuracy and completeness of coding for Hispanic ethnicity.20 Many Asian and Pacific Islander men and women in the continental U.S. reside in urban areas of California and New York, and there may be regional heterogeneity in ethnicity-specific incidence that the current study did not address.

In the current study, we did not observe large geographic differences in stage at diagnosis. Spatial analyses of geographic differences in late-stage colorectal cancer, such as the study by Rushton et al.21 in Iowa, are also likely to be useful for identifying geographic patterns in colorectal cancer, which may be due to differential access to screening, nutritional factors, physician practice patterns, or other factors.

In conclusion, the results from this study add to the literature on disparities in colorectal cancer incidence by rural/nonrural residence including effect modification by race. An interesting and potentially important finding was that the incidence rate of colorectal cancer was higher among black men who reside in metropolitan areas of the United States than among black men who reside in rural areas. There may also be a higher rate of colorectal cancer incidence among Asians and Pacific Islanders living in rural areas of the United States than among those living in metropolitan areas.

Acknowledgements

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

We appreciate the in-kind support from all the contributors to this monograph and also are grateful to Jessica King for the preparation of the analytic files and to Faruque Ahmed for his leadership of the Colorectal cancer monograph project.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES
  • 1
    Pickle LW, Mason TJ, Howard N, et al. Atlas of U.S. Cancer Mortality Among Nonwhites: 1950–1980. Washington, DC: U.S. Government Printing Office; 1990. DHHS pub NIH 90–1582.
  • 2
    Devesa SS, Grauman DJ, Blott WJ, et al. Atlas of Cancer Mortality in the United States, 1950–94. Bethesda, MD: National Cancer Institute; 1999. NIH pub 99–4564.
  • 3
    Fulton JP, Correa CN, Hirschenberger W, et al. Urbanization and cancer incidence, United States, 1988–1992. In: HoweHL, ed. Cancer Incidence in North America, 1989–1993, Vol 1. Sacramento, CA: North American Association of Central Cancer Registries; 1997: VI-1-VI-9.
  • 4
    Pickle LW, Feuer EJ, Edwards BK. U.S. Predicted Cancer Incidence, 1999: Complete Maps by County and State from Spatial Projection Models. Bethesda, MD: National Cancer Institute; 2003. NCI Cancer Surveillance Monograph Series, No. 5, NIH pub 03-5435.
  • 5
    Ries LA, Wingo PA, Miller DS, et al. The annual report to the nation on the status of cancer, 1973–1997, with a special section on colorectal cancer. Cancer. 2000; 88: 23982424.
  • 6
    Coughlin SS, Thompson TD, Seeff L, et al. Breast, cervical, and colorectal cancer screening in a demographically defined region of the Southern United States. Cancer. 2002; 95: 22112227.
  • 7
    Coughlin SS, Thompson TD. Colorectal cancer screening practices among men and women in rural and nonrural areas of the United States, 1999. J Rural Health. 2004; 20: 118124.
  • 8
    Casey MM, Call KT, Klingner JM. Are rural residents less likely to obtain recommended preventive health care services? Am J Prev Med. 2001; 21: 182188.
  • 9
    Armstrong LR, Thompson T, Hall HI, et al. Colorectal cancer mortality among Appalachian men and women, 1969–1999. Cancer. 2004; 101: 28512858.
  • 10
    Jackson-Thompson J, German RR, Faruque A, Lai S-M, Friedman C. Descriptive epidemiology of colorectal cancer in the United States, 1998–2001. Cancer. (in press).
  • 11
    Surveillance, Epidemiology, and End Results (SEER) Program. SEER*Stat Database: Incidence—SEER 18 Regs. Submitted Nov 2003 for expanded races and Hispanics (1990–2001 varying); released June 2004 based on Nov 2003 submission. Cancer Statistics Branch, Surveillance Research Program, DCCPS, National Cancer Institute, 2004. Available at URL: www.seer.cancer.gov
  • 12
    U.S. Department of Agriculture Economic Research Service. 2003 Rural–urban continuum codes. Available at URL: www.ers.usda.gov/data/RuralUrbanContinuumCodes Accessed January 24, 2006.
  • 13
    YoungJL JR, RoffersSD, RiesLAG, et al., eds. SEER Summary Staging Manual—2000: Codes and Coding Instructions. Bethesda, MD: National Cancer Institute; 2001. NIH pub 01-4969.
  • 14
    Census 2000, Summary File 3, Technical Documentation. U.S. Census Bureau, 2002.
  • 15
    Massey DS, Eggers ML. The ecology of inequality: Minorities and the concentration of poverty, 1970–1980. Am J Sociol. 1990; 95: 11531188.
  • 16
    Krieger N. Overcoming the absence of socioeconomic data in medical records: Validation and application of a census-based methodology. Am J Public Health. 1992; 82: 703710.
  • 17
    Singh GK, Miller BA, Hankey BR, et al. Area Socioeconomic Variations in U.S. Cancer Incidence, Mortality, Stage, Treatment, and Survival, 1975–1999. Betheseda, MD: NCI; 2003. NCI Cancer Surveillance Monograph Series, No. 4, NIH pub 03-5417.
  • 18
    Fay MP, Feuer EJ. Confidence intervals for directly standardized rates: A method based on the gamma distribution. Stat Med. 1997; 16: 791801.
  • 19
    Byers AL, Allore H, Gill TM, et al. Application of negative binomial modeling for discrete outcomes. A case study in aging research. J Clin Epidemiol. 2003; 56: 559564.
  • 20
    Greenlee R. Evaluation of Methods to Estimate and Compare Race/Ethnicity Specific Cancer Incidence Rates Using Combined Central Cancer Registry Data. Sacramento, CA: North American Association of Central Cancer Registries; 2004: VI-1-VI-9.
  • 21
    Rushton G, Peleg I, Banerjee A, et al. Analyzing geographic patterns of late-stage colorectal cancer in Iowa. J Med Syst. 2004; 28: 223236.