SEARCH

SEARCH BY CITATION

Keywords:

  • colorectal cancer;
  • stage at diagnosis;
  • screening;
  • rural;
  • socioeconomic status

Abstract

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

BACKGROUND.

Rural residence has been associated with increased risk of late stage cancer diagnosis, but it is unknown if this is related to lower socioeconomic status (SES) of rural residents or to other factors. This study examined the impacts of SES and urban/rural status on colorectal cancer (CRC) stage at diagnosis in California.

METHODS.

Cases of CRC among persons ≥50 years of age diagnosed from 1988–2000 were obtained from the California Cancer Registry. A composite census based SES measure was used in the analysis, and the RUCA (Rural Urban Commuting Areas) classification scheme was used to categorize the residence at diagnosis as urban, large town, or small town. Multivariate logistic regression was used to examine the association between SES, urban/rural status, and late stage at diagnosis.

RESULTS.

In multivariate models, SES had the strongest association with stage at diagnosis among individuals living in urban areas. As SES level increased, odds of late stage at diagnosis decreased. Individuals in the highest SES category had lower odds of being diagnosed at late stage when compared with those in the lowest SES category (O.R. = 0.91, 95% C.I. = 0.87, 0.94). For individuals who lived in large towns and small rural towns, SES was not significantly associated with stage at diagnosis. We found no significant differences in stage at diagnosis by urban/rural status within SES categories.

CONCLUSIONS.

These data suggest that the relationship between SES and the risk of late stage colorectal cancer varies among rural and urban populations. Further research into the factors that influence access to and utilization of colorectal cancer screening in rural areas is needed. Cancer 2006. © 2006 American Cancer Society.

Colorectal cancer is the third most common cancer among men and women in California, and ranks third among the leading causes of cancer-related deaths for both genders.1 Stage at diagnosis is the most significant prognostic factor for colorectal cancer survival,2 and is associated with a variety of factors including age, race, urban/rural status, and socioeconomic status (SES).3–7 Although earlier research focused primarily on racial differences in stage at diagnosis, in recent years, there has been growing recognition of the importance of SES as an equally important determinant.8 SES is a construct consisting of two primary factors: access to resources suchas income and education, as well as status or social position in society.9 Living in low SES areas is associated with decreased utilization of screening services and later stage at diagnosis for several cancer sites, including breast, prostate and colorectal cancers.4, 6, 10, 11 SES may be a marker for various factors such as access to health care and low literacy levels,12–15 or it may represent differences in lifestyle factors that affect cancer risk, such as diet and physical activity.16

Living in rural areas (rurality) has also been associated with later stage at diagnosis,3, 5, 7 possibly due to reduced access to or utilization of early screening tests. Several surveys of rural residents have identified numerous barriers to obtaining preventative health care services including: lack of health insurance, long travel distances, and lack of knowledge about screening guidelines.17–20 The definition of urban/rural status shows considerable variation across studies. Some researchers use population density as a measure,7 while others substitute distance to treatment facilities as a measure of rurality.3, 21 Less is known about the effects of rural/urban status on stage at diagnosis than other factors, possibly due to difficulties in quantifying the variable.

The objectives of this study were twofold: 1) to examine stage distributions of colorectal cancer cases in California with respect to SES and urban/rural status and 2) to assess the impacts of SES and urban/rural status on colorectal cancer stage at diagnosis after controlling for associated demographic factors. To our knowledge, this is the first population-based study of the effects of both SES and urban/rural status on colorectal cancer stage at diagnosis in California.

MATERIALS AND METHODS

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

Study Population

Colorectal cancer cases were drawn from the California Cancer Registry (CCR) database. The CCR is the largest population-based cancer registry, for a geographically contiguous area in the world, and collects incidence reports on over 130,000 new cases of cancer diagnosed annually in California. Cancer reporting in California is legally mandated and was fully implemented in 1988 with standardized data collection and quality control procedures.22–25 Our study population included all individuals diagnosed with colorectal cancer (ICD-O-3 sites C18.0–C18.9, C19.9, C20.9, and C26.0),26 for whom this was the first diagnosis of cancer, aged 50 or older, diagnosed from 1988–2000 in California.

Description of Variables

The outcome of interest in this analysis was stage at diagnosis. SEER summary stage was the classification scheme used to measure stage at diagnosis.27 In the univariate analysis, stage at diagnosis was presented in the following categories: in situ, localized, regional, distant, and unknown. For the purposes of the multivariate analysis, however, the in situ and localized categories were considered early stage, and regional and distant categories were considered late stage. Unstaged cases or those with unknown stage were excluded from the multivariate analysis.

The measure of SES used in this analysis was a composite measure previously created by Yost et al. using CCR and census data.28 Census files were linked to the CCR file based on the cases' block group of residence at the time of diagnosis. Cases that were not able to be geocoded to a street address (5.5% of cases) were randomly allocated to census block groups within their county of residence. Cases diagnosed during the period 1988–1995 were linked to 1990 census data, and cases diagnosed from 1996 forward were linked to 2000 census data. Principal components analysis was then used to create a composite SES score using several census variables, including median household income, education level, proportion below 200% poverty level, and median house value. Quintiles of this SES score were used in the analysis, with a value of one representing the lowest SES level and a value of five representing the highest SES level.

Degree of rurality was measured using the Rural Urban Commuting Areas (RUCAs) categories. RUCAs refer to a census tract-based classification scheme that utilizes the standard Bureau of Census urban area and place definitions in combination with commuting information to characterize all of the nation's census tracts regarding their rural and urban status and relationships.29 Cases diagnosed during the period 1988–1995 were linked to 1990 RUCA codes and cases diagnosed from 1996 forward were linked to the 2000 RUCA codes. On the basis of the classification strategies recommended by RUCA developers and our sample size, individual RUCA codes were grouped into three categories: urban, large town, and small rural town. This classification was based on two factors: population size and the proportion of residents commuting to an urban core. Areas were classified as urban if the population was 50,000 or more, or if the population was less than 50,000, but 30% or more residents commuted to an urban core. If fewer than 30% of the population commuted to an urban core, areas with populations ranging from 10,000 to 49,999 were categorized as large towns, and areas with less than 10,000 population were categorized as small rural towns.

Other demographic variables included in the analysis were age (50–64, 65–74, 75–84, 85+ years), race/ethnic group (non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic Asian/Pacific Islander), sex (male and female), and marital status (married and unmarried). The race/ethnic categories were mutually exclusive. Individuals identified as Hispanic were included only in the Hispanic category, and could be of any race. A total of 1512 individuals classified as either non-Hispanic American Indian/Alaskan Native, other race or unknown race were excluded from the analysis. The unmarried category of marital status included individuals who self-reported as single, widowed, or divorced.

Statistical Analysis

Descriptive statistics were used to summarize the demographic characteristics of the study population. χ2 tests were used to examine the associations between SES and urban/rural status, as well as the associations between stage at diagnosis and demographic variables. Because the results of the bivariate analysis of SES and urban/rural status revealed a significant association between the variables, the multivariate analysis consisted of unconditional logistic regression models stratified on urban/rural status. Three separate models were run to examine the independent effect of SES on stage at diagnosis within each category of urban/rural status. A dichotomous stage variable (early vs. late stage) was the outcome of interest. The models were adjusted for the following variables: age, race/ethnicity, sex, and marital status. Odds ratios and 95% confidence intervals were generated to evaluate the associations between SES and stage at diagnosis within each urban/rural group. Similar models stratified on SES level were run in order to examine the independent effect of urban/rural status on stage at diagnosis.

A sensitivity analysis was performed to address the issue of unknown stage (n = 12,503). Many of the unstaged/unknown cases were thought to be distant cancers because in approximately 73% of these cases no surgery was performed and the diagnosis of cancer was made based on either an autopsy or death certificate only. To address this issue, cases with unknown stage were classified as distant stage and models were run, in order to compare them with the models in which they were excluded. Since the effect measures were similar in magnitude, the decision was made to exclude the cases with unknown stage in multivariate analyses. All analyses were conducted using SAS (V. 9.1, SAS Institute, Cary, NC).

RESULTS

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

There were a total of 180,413 colorectal cancer cases, aged 50 and older diagnosed during the period 1988–2000. Demographic and clinical characteristics of the study population are summarized in Table 1. Approximately 78% of cases were non-Hispanic white and nearly two-thirds of the cases were between the ages of 65 and 84. One-third of the population was in the lowest two SES categories, while 45% were in the two highest SES categories. The vast majority of cases lived in urban areas. Only about 4% and 3% of cases lived in large towns and small rural towns, respectively. A little over half of cases were diagnosed at late stage (regional or distant), and 7% of cases had unknown/unstaged cancers.

Table 1. Demographic and Clinical Characteristics of Colorectal Cancer Cases, and Proportion Diagnosed at Late Stage among Persons Ages 50 Years and Older, California, 1988–2000 (N = 180,413)*
VariableNo. (%)Late stage (%)
  • SES indicates socioeconomic status; RUCA, Rural Urban Commuting Areas.

  • *

    Data are from the California Cancer Registry.

  • Late stage cases are those that are classified as regional or distant.

  • P < 0.0001.

  • §

    A total of 1512 individuals classified as one of the following were excluded from the analysis: American Indian/Alaskan Native, other race, or unknown race.

  • The ‘unmarried’ category includes those individuals who are categorized as never married, widowed, or divorced.

  • Mantel trend test: Odds ratio = 0.91, P < 0.00001.

  • #

    P = 0.05.

  • **

    SEER Summary Stage.

  • ††

    Cases with unknown stage (n = 12,503) were excluded from the analysis.

Age group
 50–6443,638 (24.2)58.0
 65–7458,983 (32.7)55.7
 75–8455,875 (31.0)56.7
 85+21917 (12.1)59.7
Race/ethnicity§
 Non-Hispanic white139563 (78.0)56.8
 Non-Hispanic black11382 (6.4)58.3
 Hispanic15991 (8.9)58.8
 Non-Hispanic Asian/Pacific Islander11965 (6.7)58.6
Sex
 Male92695 (51.4)55.4
 Female87718 (48.6)58.7
Marital status
 Married100513 (57.6)55.9
 Unmarried74154 (42.4)59.8
SES level
 1 (Low)25324 (14.0)59.1
 235526 (19.7)58.6
 339272 (21.8)56.6
 440447 (22.4)55.9
 5 (High)39844 (22.1)55.8
RUCA#
 Urban168522 (93.4)57.0
 Large town6459 (3.6)56.2
 Small rural town5432 (3.0)58.5
Stage at diagnosis**††
 In situ11797 (6.5)
 Localized60397 (33.5)
 Regional65355 (36.2)
 Distant30361 (16.8)
 Unknown12503 (7.0)

The proportion of cases with late stage at diagnosis differed significantly by age, race/ethnicity, sex, SES and rurality (Table 1). Subgroups with the greatest proportion of cases detected at late stage included people 85 and older (59.7%), Hispanics (58.8%), females (58.7%), unmarried individuals (59.8%), and small rural town residents (58.5%). There was a significant inverse dose-response relationship between SES and stage at diagnosis. As SES level increased, the proportion of individuals diagnosed at late stage decreased (Mantel trend test: P< 0.0001).

The bivariate analysis of SES and urban/rural status revealed a significant association between the two variables (Table 2). A significantly larger proportion of individuals who lived in small rural towns (68.5%) were in the two lowest SES strata when compared with those living in both large towns (58.8%) and urban areas (31.3%). Conversely, approximately 47% of individuals in urban areas belonged to the highest two SES strata, compared with 13% and 8% of individuals in large and small rural towns, respectively.

Table 2. SES and Urban/Rural Status among Colorectal Cancer Cases Ages 50 Years and Older, California, 1988–2000 (n = 167,910)*
SES levelUrban, no.Large town, no.Small rural town, no.Total, no.
  • *

    SES indicates socioeconomic status. Data are from the California Cancer Registry. Cases with unknown stage (n = 12,503) were excluded. P < 0.0001.

  • Values in parentheses indicate Column %.

1 (Low)20502 (13.0)1628 (27.4)1065 (22.0)23195 (13.8)
228707 (18.3)1868 (31.4)2253 (46.5)32828 (19.5)
333705 (21.5)1686 (28.3)1123 (23.2)36514 (21.8)
436800 (23.4)662 (11.1)330 (6.8)37792 (22.5)
5 (High)37400 (23.8)109 (1.8)72 (1.5)37581 (22.4)
Total157114 (100.0)5953 (100.0)4843 (100.0)167910 (100.0)

Table 3 summarizes the results of the multivariate analysis stratified on urban/rural category. The association between SES and stage at diagnosis is presented in three different models, each representing a category of urban/rural status. SES had a strong association with stage at diagnosis among individuals living in urban areas. As SES level increased, odds of late stage at diagnosis decreased. Individuals in the highest SES category had significantly reduced odds of late stage diagnosis when compared with those in the lowest SES category (O.R. = 0.91, 95% C.I. = 0.87, 0.94). For individuals who lived in large towns and small rural towns, SES was not significantly associated with stage at diagnosis.

Table 3. The Effects of SES on Odds of Late Stage at Diagnosis within Urban/Rural Category for Colorectal Cancer Cases Ages 50 Years and Older, California, 1988–2000 (n = 167,910)*
SES levelResidential category
Urban (n = 157,114), ORLarge town (n = 5953), ORSmall rural town (n = 4843), OR
  • SES indicates socioeconomic status; OR, odds ratio; CI, confidence interval.

  • *

    Data are from the California Cancer Registry and adjusted for race, age, sex, and marital status.

  • Values in parentheses indicate 95% CI.

11.001.001.00
20.99 (0.96–1.04)0.94 (0.82–1.08)0.95 (0.82–1.11)
30.93 (0.89–0.96)0.82 (0.72–0.95)0.87 (0.73–1.03)
40.90 (0.86–0.93)0.98 (0.81–1.18)0.81 (0.63–1.04)
50.91 (0.87–0.94)1.10 (0.73–1.66)0.83 (0.51–1.34)

The results of the multivariate models stratified on SES level did not indicate any significant differences in stage at diagnosis by urban/rural category (Table 4). Although the odds ratios for small rural town were slightly elevated within the lowest three SES categories, the differences did not reach statistical significance.

Table 4. The Effects of Urban/Rural Category on Odds of Late Stage at Diagnosis by SES Level for Colorectal Cancer Cases Aged 50 Years and Older, California, 1988–2000 (N = 167,910)*
Urban/rural categorySES level
1 (n = 25,324), OR2 (n = 35,526), OR3 (n = 39,272), OR4 (n = 40,447), OR5 (n = 39,844), OR
  • SES indicates socioeconomic status; OR, odds ratio; CI, confidence interval.

  • *

    Data are from the California Cancer Registry and adjusted for race, age, sex, and marital status.

  • Values in parentheses indicate 95% CI.

Urban1.001.001.001.001.00
Large town0.99 (0.89, 1.10)0.92 (0.84, 1.01)0.88 (0.80, 0.97)1.06 (0.91, 1.24)1.19 (0.80, 1.77)
Small rural town1.08 (0.95, 1.22)1.04 (0.95, 1.14)1.04 (0.92, 1.17)0.99 (0.79, 1.23)1.01 (0.63, 1.6)

DISCUSSION

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

This study examined the impact of SES and rural–urban differences on the detection of colorectal cancers at advanced stage. We found no evidence that rural residents were any more likely than urban residents of comparable SES to have colorectal cancers detected at late stage. However, we found a significant inverse relationship between SES among urban, but not rural residents, suggesting that urban/rural status may modify the effect of SES on late stage cancer diagnosis. These results pertaining to individuals living in urban areas are consistent with previous work documenting the inverse relationship between SES and the likelihood of late stage cancer diagnosis.6, 10 To our knowledge, however, there have been no previous reports examining the impact of SES on late stage cancer diagnosis among rural residents. These findings suggest that SES may be a less important determinant of late stage diagnosis than other factors among rural residents.

Previous studies have reported that rural residents are more often diagnosed with late stage colorectal and other cancers3, 5, 30–32 and are less likely than their urban counterparts to receive preventive services including cancer screening.17, 33–35 These inequalities have been assumed to be in some part attributable to the generally lower income and education levels of rural compared to urban populations.36 However, we found no evidence of urban–rural differences in late stage colorectal cancer diagnosis at any level of SES. Possible reasons for the disagreement of our results with those of previous studies include differences in methods, i.e., small sample sizes in rural populations, different definitions of urban/rural status, absence of adjustment for SES, or the use of other SES indicators in previous studies. In addition, none of the previous studies examined a California population. It is possible that rural residents in California have different, unmeasured characteristics that may impact late stage cancer diagnosis when compared with those of other states or countries.

Our finding that SES was not a significant determinant of late stage cancer risk in rural residents can be explained by several possible factors. Fewer medical resources in rural areas means that residents of every SES level must travel longer distances to receive preventive services.37 This is supported by findings that late stage colorectal cancer rates are higher among those traveling the longest distance to their place of diagnosis,38 and in areas with fewer primary care physicians.39 Studies in which distance to screening facilities was used as a measure of urban/rural status have found that individuals living further away from a cancer center have less chance of a firm diagnosis before death and/or later stage at diagnosis.3, 5 Other factors such as differences in physician characteristics in urban and rural areas may also play a role in individual decisions to seek medical care and receipt of preventative services. These factors have been examined in relation to breast cancer screening, with studies indicating that physician recommendation is the most influential determinant of mammography among rural women,19, 40, 41 and that the likelihood of physician recommendation may vary among rural and urban physicians.3, 35

It is also possible that individuals in rural areas share characteristics independent of SES level (e.g. attitudes/knowledge about preventative services, characteristics of rural physicians, etc.) that are associated with health seeking behavior, or lack thereof. It may be that rural residents, who have been found to be more fatalistic, independent, and self-reliant than urban residents,36 have a higher threshold for seeking health care. A study of breast cancer screening and attitudes of rural and urban women found that while rural women exhibited levels of knowledge about breast cancer similar to urban women, they held more pessimistic views about breast cancer prognosis, which influenced their likelihood to undergo mammography screnning.35 These differences remained upon adjustment for SES. We were not able to include direct measures of these variables in our study, as our dataset lacked this information.

An important methodologic consideration in any study of rurality is the definition of rural and urban. Although rurality is often presented as a dichotomous variable, such categorization may mask important differences and misclassify a substantial proportion of a population. Other researchers have called for finer distinctions of this variable.34, 42 Because of the large number of cases available to us, we were able to categorize urban/rural status into three groups, instead of the dichotomous category used in most studies. Our finding that the inverse relationship between stage and SES did not hold in large towns may bear further study. Large towns often do have more primary care providers than isolated rural areas, so geographic access to care may not be as much of a factor as characteristics of providers or residents in these areas.

The present study has several strengths, including the use of a large, population-based dataset. The CCR represents the largest, geographically contiguous cancer registry in the world. In addition, California is a large state which has population diversity in terms of race/ethnic groups, SES and urban/rural areas. The ability of our registry to geocode most addresses to the block group level also enhances our ability to produce ecologic SES data for our cases. We also addressed the association between SES and urban/rural status in our study by producing separate models for each urban/rural group. Previous studies have not done this.

Despite these strengths, this study has a few limitations that are important to consider in the interpretation of the results. First, we did not have data on individual SES level. The use of an aggregate SES measure may or may not represent individual level SES, although work by Kreiger et al. concludes that a census-based methodology is a valid and useful approach to overcoming the absence of socioeconomic data in epidemiologic studies.43 Second, the numbers of individuals in the highest SES categories in large and small towns may have limited our ability to detect associations between SES and stage at diagnosis. Another possible limitation is the use of the RUCA classification scheme to define urban/rural areas in California. Since the RUCA scheme was developed for national use, it may or may not be the most appropriate classification scheme for designating urban/rural areas in California. For example, one California-specific analysis found that the RUCA scheme substantially under-identified the number of rural areas, leaving over 20% of the current rural health providers ineligible for certain federal rural health programs.44 This potential misclassification of rural residence may dilute the ability to identify significant differences between classification areas. Moreover, with the federal RUCA classification scheme, there were some changes in definitions of urban and rural areas between the 1990 and 2000 census. An additional reason for misclassification of rural areas may be related to rapid population growth in many areas of state in the past decade. However, the advantage of using the RUCA classification is that it is a standardized scheme that allows for comparison across states. Finally, we were not able to include measures (individual or ecologic), regarding health care access, health insurance penetration, or utilization in our study, as our registry does not routinely collect these variables.

In summary, although we found no evidence that rural California residents were at higher risk of late stage colorectal cancer diagnosis than urban residents of comparable SES, the relationship between SES and colorectal cancer stage at diagnosis differed among urban and rural residents. Future studies need to focus on examining the specific characteristics of both individuals and the environment that may affect access to and utilization of care in rural areas. In addition, methodologic studies that compare the use of different urban/rural classification schemes would be useful in deciding which measures to use in future studies of this type.

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 Faruque Ahmed for his leadership of colorectal cancer monograph project.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES
  • 1
    American Cancer Society, California Division and Public Health Institute, California Cancer Registry. California Cancer Facts and Figures. American Cancer Society, California Division, Oakland, CA, 2003.
  • 2
    Ries L, Eisner M, Kosary C, Hankey B, Miller B, Clegg L, et al. SEER Cancer Statistics Review, 1975–2000. Bethesda, MD: National Cancer Institute; 2003.
  • 3
    Campbell NC, Elliott AM, Sharp L, Ritchie LD, Cassidy J, Little J. Rural and urban differences in stage at diagnosis of colorectal and lung cancers. Br J Cancer. 2001; 84: 910914.
  • 4
    Ionescu MV, Carey F, Tait IS, Steele RJ. Socioeconomic status and stage at presentation of colorectal cancer. Lancet. 1998; 352: 1439.
  • 5
    Liff JM, Chow WH, Greenberg RS. Rural-urban differences in stage at diagnosis. Possible relationship to cancerscreening. Cancer. 1991; 67: 14541459.
  • 6
    Schwartz KL, Crossley-May H, Vigneau FD, Brown K, Banerjee M. Race, socioeconomic status and stage at diagnosis for five common malignancies. Cancer Causes Control. 2003; 14: 761766.
  • 7
    Schouten LJ, Meijer H, Huveneers JA, Kiemeney LA. Urban-rural differences in cancer incidence in The Netherlands 1989–1991. Int J Epidemiol. 1996; 25: 729736.
  • 8
    Krieger N, Quesenberry CJr, Peng T, Horn-Ross P, Stewart S, Brown S, et al. Social class, race/ethnicity, and incidence of breast, cervix, colon, lung, and prostate cancer among Asian, Black, Hispanic, and White residents of the San Francisco Bay Area, 1988–1992 (United States). Cancer Causes Control. 1999; 10: 525537.
  • 9
    Krieger N, Williams DR, Moss NE. Measuring social class in US public health research: Concepts, methodologies, and guidelines. Annu Rev Public Health. 1997; 18: 341378.
  • 10
    Mandelblatt J, Andrews H, Kao R, Wallace R, Kerner J. The late stage diagnosis of colorectal cancer: Demographic and socioeconomic factors. Am J Public Health. 1996; 86: 17941797.
  • 11
    Vernon SW, Laville EA, Jackson GL. Participation in breast screening programs: A review. Soc Sci Med. 1990; 30: 11071118.
  • 12
    Baquet CR, Commiskey P. Socioeconomic factors and breast carcinoma in multicultural women. Cancer. 2000; 88( 5 Suppl): 12561264.
  • 13
    Breen N, Figueroa JB. Stage of breast and cervical cancer diagnosis in disadvantaged neighborhoods: A prevention policy perspective. Am J Prev Med. 1996; 12: 319326.
  • 14
    Figueroa JB, Breen N. Significance of underclass residence on the stage of breast or cervical cancer diagnosis. Am Econ Rev. 1995; 85: 112116.
  • 15
    Bennett CL, Ferreira MR, Davis TC, Kaplan J, Weinberger M, Kuzel T, et al. Relation between literacy, race, and stage of presentation among low-income patients with prostate cancer. J Clin Oncol. 1998; 16: 31013104.
  • 16
    Liu L, Deapen D, Bernstein L. Socioeconomic status and cancers of the female breast and reproductive organs: A comparison across racial/ethnic populations in Los Angeles County, California (United States). Cancer Causes Control. 1998; 9: 369380.
  • 17
    Casey MM, Thiede Call K, Klingner JM. Are rural residents less likely to obtain recommended preventive healthcare services? Am J Prev Med. 2001; 21: 182188.
  • 18
    Elnicki DM, Morris DK, Shockcor WT. Patient-perceived barriers to preventive health care among indigent, rural Appalachian patients. Arch Intern Med. 1995; 155: 421424.
  • 19
    Carr WP, Maldonado G, Leonard PR, Halberg JU, Church TR, Mandel JH, et al. Mammogram utilization among farm women. J Rural Health. 1996; 12( 4 Suppl): 278290.
  • 20
    Strickland WJ, Hanson CM. Coping with the cost of prescription drugs. J Health Care Poor Underserved. 1996; 7: 5062.
  • 21
    Meden T, St John-Larkin C, Hermes D, Sommerschield S. MSJAMA. Relationship between travel distance and utilization of breast cancer treatment in rural northern Michigan. JAMA. 2002; 287: 111.
  • 22
    Cancer Reporting in California: Standards for Automated Reporting. California Cancer Reporting System Standards, Volume II. California Department of Health Services, Cancer Surveillance Section, Sacramento, CA, December 1997.
  • 23
    Cancer Reporting in California: Data Standards for Regional Registries and California Cancer Registry. California Cancer Reporting System Standards, Volume III. California Department of Health Services, Cancer Surveillance Section, Sacramento, CA, December 1997.
  • 24
    Cancer Reporting in California: Reporting Procedures for Physicians. California Cancer Reporting System Standards, Volume IV. California Department of Health Services, Cancer Surveillance Section, Sacramento, CA, January 1998.
  • 25
    Cancer Reporting in California: Abstracting and Coding Procedures for Hospitals. California Cancer Reporting System Standards, Volume I. California Department of Health Services, Cancer Surveillance Section, Sacramento, CA, June 1997.
  • 26
    Fritz A, Percy C, Jack A, Shanmugaratnam K, Parkin DM, Whelan S. International Classification of Diseases for Oncology, 3rd edn. Geneva: World Health Organization, 2000.
  • 27
    YoungJLJr, RoffersSD, RiesLAG, FritzAG, HurlbutAA, eds. SEER Summary Staging Manual - 2000: Codes and Coding Instructions. Bethesda, MD: National Cancer Institute; 2001. NIH Pub. No. 01-4969.
  • 28
    Yost K, Perkins C, Cohen R, Morris C, Wright W. Socioeconomic status and breast cancer incidence in California for different race/ethnic groups. Cancer Causes Control. 2001; 12: 703711.
  • 29
    Rural Health Research Center. Rural-Urban Commuting Area Codes (RUCAs), 2003. Available at http://www.fammed.washington.edu/wwamirhrc.
  • 30
    Palmer RC, Schneider EC. Social disparities across the continuum of colorectal cancer: A systematic review. Cancer Causes Control. 2005; 16: 5561.
  • 31
    Conlisk E. Colorectal cancer in North Carolina. Risk factors, screening behaviors, incidence, stage at diagnosis, and mortality. NC Med J. 2001; 62: 298303.
  • 32
    Thomas A, Carlin BP. Late detection of breast and colorectal cancer in Minnesota counties: An application of spatial smoothing and clustering. Stat Med. 2003; 22: 113127.
  • 33
    Coughlin SS, Thompson TD, Seeff L, Richards T, Stallings F. Breast, cervical, and colorectal carcinoma screening in ademographically defined region of the southern U.S. Cancer. 2002; 95: 22112222.
  • 34
    Coughlin SS, Thompson TD, Hall HI, Logan P, Uhler RJ. Breast and cervical carcinoma screening practices among women in rural and nonrural areas of the United States, 1998–1999. Cancer. 2002; 94: 28012812.
  • 35
    Bryant H, Mah Z. Breast cancer screening attitudes and behaviors of rural and urban women. Prev Med. 1992; 21: 405418.
  • 36
    Larson SL, Fleishman JA. Rural-urban differences in usual source of care and ambulatory service use: Analyses of national data using Urban Influence Codes. Med Care. 2003; 41( 7 Suppl): III65III74.
  • 37
    Harris R, Leininger L. Preventive care in rural primary care practice. Cancer. 1993; 72( 3 Suppl): 11131118.
  • 38
    Rushton G, Peleg I, Banerjee A, Smith G, West M. Analyzing geographic patterns of disease incidence: Rates of late stage colorectal cancer in Iowa. J Med Syst. 2004; 28: 223236.
  • 39
    Roetzheim RG, Pal N, Gonzalez EC, Ferrante JM, Van Durme DJ, Ayanian JZ, et al. The effects of physician supply on the early detection of colorectal cancer. J Fam Pract. 1999; 48: 850858.
  • 40
    Rosenman KD, Gardiner J, Swanson GM, Mullan P, Zhu Z. U.S. farm women's participation in breast cancer screening practices. Cancer. 1995; 75: 4753.
  • 41
    Rauscher GH, Hawley ST, Earp JA. Baseline predictors of initiation vs. maintenance of regular mammography use among rural women. Prev Med. 2005; 40: 822830.
  • 42
    Zhang P, Tao G, Irwin KL. Utilization of preventive medical services in the United States: A comparison between rural and urban populations. J Rural Health. 2000; 16: 349356.
  • 43
    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.
  • 44
    California State Rural Health Association. An appropriate definition of rural for California, June 2002. Available at http://www.csrha.org.