Skip to main content
NCBI home page
Health Services Research logoLink to Health Services Research
. 2013 Feb 28;48(4):1433–1449. doi: 10.1111/1475-6773.12047

Radiation Therapy Resources and Guideline-Concordant Radiotherapy for Early-Stage Breast Cancer Patients in an Underserved Region

Nengliang Yao 1, Stephen A Matthews 2, Marianne M Hillemeier 3, Roger T Anderson 4
PMCID: PMC3703497  NIHMSID: NIHMS457926  PMID: 23445478

Abstract

Objective

To examine the relationship between radiation therapy resources and guideline-concordant radiotherapy after breast-conserving surgery (BCS) in Kentucky.

Data Sources

The SEER registry and Area Resource File provided county-level data describing cancer care resources and socioeconomic conditions of Kentucky residents.

Study Design

The outcome variable was rate of BCS without radiotherapy in each county for 2000–2007. Eight-year weighted average rates of radiation therapy providers and hospitals per 100,000 residents were explanatory variables of interest. Exploratory spatial data analyses and spatial econometric models were estimated.

Principal Findings

Appalachian counties in Kentucky had significantly fewer radiation oncologists, hospitals with radiation therapy facilities, and surgeons per 100,000 residents than non-Appalachian counties. The likelihood of BCS without radiation was significantly higher among Appalachian compared to non-Appalachian women (42.5 percent vs. 29.0 percent, p < .001). Higher proportions of women not receiving recommended radiotherapy after BCS were clustered in Eastern Kentucky around Lexington. This geographic disparity was partially explained by significantly fewer radiation therapy facilities in Appalachian Kentucky in adjusted analyses.

Conclusions

Scarce radiation therapy resources in Appalachian Kentucky are associated with disparities in receipt of guideline-concordant radiotherapy, suggesting that policy action is needed to improve the cancer treatment infrastructure in disadvantaged mountainous areas.

Keywords: Underserved, cancer treatment resource, guideline-concordant care, spatial analysis

The National Cancer Institute has designated Appalachia as a priority area (Zerhouni and Ruffin 2002; National Cancer Institute 2010), based on a growing body of research demonstrating cancer health disparities among Appalachian residents (Lengerich et al. 2005; Wingo et al. 2008). Overall breast cancer mortality rates have declined less rapidly in Appalachian counties (17.5 percent) in recent years compared with non-Appalachian counties in Appalachian states (30.5 percent), and compared with non-Appalachian U.S. counties as a whole (28.3 percent) (Yao, Lengerich, and Hillemeier 2012). Researchers have suggested that lack of cancer care resources plays a part in Appalachian cancer disparities (Friedell et al. 2001; Armstrong et al. 2004; Lengerich et al. 2004; Mayo et al. 2004; Coughlin et al. 2006; Lyttle and Stadelman 2006; Kelly et al. 2009; McAlearney et al. 2010; Yao, Lengerich, and Hillemeier 2012), but few have specifically analyzed the availability of those resources in the region. The availability of medical resources within Appalachia has received some attention (Susi and Mascarenhas 2002; Halverson, Ma, and Harner 2004; Wang and Luo 2005; Denham, Wood, and Remsberg 2010; U.S. Department of Health and Human Services 2010); however, cancer care-related specialists and facilities were not examined in this research. Insufficient research has been done to describe the availability of cancer care resources in the Appalachia region to examine how resource availability may relate to cancer outcomes.

The availability of radiation therapy-related resources is important for treatment of women with early-stage age breast cancer, as breast-conserving surgery (BCS) with radiation therapy has been recommended in practice guidelines for stage I/II breast cancers since 1990 (National Institute of Health 1991; National Cancer Institute 2012; National Comprehensive Cancer Network 2012). Recent research has shown that the likelihood of BCS without radiation is significantly higher in Appalachian patients compared to their non-Appalachian counterparts (Freeman, Huang, and Dragun 2012). This is concerning because all histologic types of early-stage invasive breast cancer are candidates for treatment with BCS plus radiation therapy (Weiss et al. 1992), and omitting radiation therapy is associated with higher local recurrence rates and a higher mortality risk (Early Breast Cancer Trialists' Collaborative Group 1995; Veronesi et al. 2001; Fisher et al. 2002; Vinh-Hung et al. 2003). Rates of BCS with radiation vary significantly across the country (Farrow, Hunt, and Samet 1992; Smith et al. 2009), and radiation has been shown to be more often omitted in certain groups of patients after BCS including minority women, younger women, women living in isolated areas, and the uninsured or under-insured (Bickell et al. 2006; Enger et al. 2006; Anderson et al. 2008; Smith et al. 2009). Patients may choose to forego radiation because of unavailability or inconvenience of cancer treatment resources or perceived unavailability/inconvenience of these resources (Anderson et al. 2008).

The main objective of this article is to use data from Kentucky, an Appalachian state that participates in the SEER Cancer registry, to describe the distribution of radiation therapy resources and examine the relationship between radiation therapy resources and guideline-concordant radiotherapy after BCS. This analysis addressed the association of radiation resources with breast cancer outcomes in the Appalachian region, an area that is of considerable interest to cancer researchers and policy makers (Freeman et al. 2012; Hall et al. 2002; Halverson, Ma, and Harner 2004; Kelly et al. 2009; Lengerich et al. 2005; McAlearney et al. 2010; National Cancer Institute; Wingo et al. 2008). A county-level ecological analysis such as this can shed light on relationships between cancer care resources and treatment receipt likely to exist in other medically underserved areas.

Data and Methods

Data

Surveillance, Epidemiological, and End Result (SEER) data for the years from 2000 through 2007 were used to measure receipt of guideline-concordant radiotherapy. Kentucky has participated in the SEER registry since 2000. County-level data on socio-demographic characteristics, cancer care infrastructure, and treatment facilities came from the Area Resource File (ARF) (US Department of Health and Human Services 2011). The county boundary files (or shape files) for deriving spatial weights matrices that were utilized in our spatial analysis are extracted from the County and Equivalent Map (Census 2000) downloaded from the website of the Geography Division of the US Census Bureau. County was used as the unit of analysis for several reasons. The administrative geographic hierarchy for much of policy making is made up of federal, state, and county governments, and many policy decisions do not extend to governmental units below the county level (Allen 2001). Also, data from the U.S. decennial census and other governmental agencies such as HRSA (e.g., Area Resource File) are readily available at the county level.

Outcome Variable

Guideline-concordant radiation therapy was the outcome of interest in this study. The outcome variable was defined as the percentage of patients omitting radiation therapy after BCS among all patients who had BCS during 2000–2007 in each Kentucky county (N = 120), stratified by location as either within or outside the Appalachian region. Patients were assigned to their county of residence at diagnosis. A total of 54 Kentucky counties are in the Appalachian region as defined by the Appalachian Regional Commission (ARC) (Appalachian Regional Commission 2010). Rates of BCS without radiation were computed based on county-level data from the SEER registry from 2000 to 2007 for Kentucky. We included all women aged 18 years or older who were diagnosed with a primary breast cancer that was American Joint Committee on Cancer (AJCC) stage I or II who underwent primary BCS and whose cancer histology (Appendix 1) was likely to be treated according to standard local therapy guidelines. Women undergoing partial mastectomy not otherwise specified (NOS), less than total mastectomy NOS, lumpectomy, and segmental mastectomy were defined as having received BCS.

Independent Variables

County-level data on cancer care infrastructure and treatment facilities came from the ARF. Breast cancer care providers of interest included surgeons and radiation oncologists. Treatment facilities considered in this study include hospitals with oncology services and hospitals with radiation therapy facilities. Independent variables of interest are the number of radiation oncologists and hospitals with radiation therapy facilities per 100,000 people in each county. Other covariates were selected based on conceptual models of equitable access to cancer services and of access to health care (Andersen et al. 1983; Mandelblatt, Yabroff, and Kerner 1999). These covariates include Appalachian status, poverty rate, the number of surgeons, and oncology service hospitals per 100,000 people, percentage of stage II cases among early-stage breast cancer patients, the Medicare managed care penetration rate in 2000 in each county, and a dummy variable to identify whether a county contains an interstate highway (a proxy measure for access and interaction between counties). Other socio-demographic characteristics, including educational level, percentage of the population that is white, and percentage of the population that is uninsured were initially included in the model, but subsequently dropped because of multicollinearity.

Analytical Methods

Descriptive analyses included calculation of weighted means and frequencies and associated standard errors. Aspects of care, including the average number of breast cancer care providers and hospitals per 100,000 residents, were weighted by the total population during 2000–2007. Exploratory spatial data analysis was conducted to identify overall spatial patterning of resource distribution. We used Moran's I to measure global spatial autocorrelation based on a first-order queen weights matrix, which defines a location's neighbors as those with either a shared border or vertex (Moran 1950). We compared the first-order Queen spatial weights matrix against other spatial weight specifications (e.g., Rook case), and we selected first-order queen weight matrix for the spatial analysis. Moran's I values are based on the z-scores of the variable in question and in our case is a correlation-like measure between the attribute of interest within a county and the average value of the attribute of interest in all neighboring counties defined by the weights matrix. The resultant value of a Moran's I for a variable of interest ranges from approximately −1 to +1; values nearer −1 indicate spatial dispersion, those near +1 suggest spatial clustering, and values close to 0 indicate a random spatial pattern. Permutation tests are used to compare the observed distribution to alternative distributions and can be used to indicate statistical significance. Local indicators of spatial association (LISA) analyses were also performed to produce cluster maps. Although Moran's I can help identify spatial structure and the presence of spatial clustering in the data, the LISA is a “local” measure that identifies the specific location of clusters (Anselin, #b500).

We expect the likelihood of accessing radiation therapy after BCS in a county may be affected by the supply of radiation therapy in the county of residence and in adjacent counties, and as such these can be modeled as spatial effects. The spatial lag model is arguably the most common and most useful way to think about spatial dependence (Ward and Gleditsch 2008). A spatial lag exists when the dependent variable y in place i is affected by the dependent variables in both place i and j. The spatial lag model incorporates the influence of unmeasured independent variables and the additional effect of neighboring attribute values, for example, the lagged dependent variable (Baller et al. 2001).

A statistically significant Moran's I for the dependent variable and also for independent variables can indicate the presence of spatial effects. The potential for spatial effects exists in ecological data defined for geographic units such as counties. Ignoring spatial dependence and the existence of spatial structure can lead to biased estimates and thus an ordinary least squares (OLS) regression model can lead to misspecification. In this article, OLS was used for model comparison purposes only, though we utilize OLS model diagnostics to help guide our analytical strategy based on spatial regression models. Specifically, we used OLS regression diagnostics to determine the form of spatial regression that best fits the data, using a Lagrange Multiplier (LM) test (Anselin 1988). In our case, the regression diagnostics indicates that a spatial model would improve on an OLS specification and that the spatial lag model was a better fit than the spatial error model. Thus, we used a spatial lag regression model to examine the relationship between radiation therapy resources and guideline-concordant radiotherapy after BCS. All statistical analysis was based on GeoDa, a specialized ESDA and spatial regression program developed by Anselin (Anselin, Syabri, and Kho 2006), and all maps were produced in ArcGIS 10 (ESRI, #b501).

Results

Descriptive statistics

A total of 17,227 women in Kentucky age ≥18 years were diagnosed with a first breast cancer that was AJCC stage I or II during the period 2000–2007, and 8,719 of them received BCS. Of those women diagnosed with early-stage breast cancer, 2,832 (32.5 percent) received BCS without radiation therapy. The likelihood of BCS without radiation was higher among Appalachian patients (42.5 percent vs 29.0 percent, p < .001). Early-stage breast cancer patients in Appalachian Kentucky counties were more likely to be diagnosed at stage II than those in other counties (46.9 percent vs 42.2 percent, p < .001) (Table 1). Appalachian counties had significantly fewer radiation oncologists, hospitals with radiation therapy facilities, and surgeons per 100,000 residents than non-Appalachian Kentucky (Table 1). The number of hospitals with oncology services per 100,000 people did not differ significantly between Appalachian and non-Appalachian counties. The average poverty rate of Appalachian counties was about twice the rate in non-Appalachian counties, and Appalachian counties were less likely to contain an interstate highway. HMO Medicare penetration rates were higher in non-Appalachian counties than in Appalachian counties.

Table 1.

Sample Statistics for County-Level Variables

Variables Non-Appalachian Region Appalachian Region Moran's I‡
Number of counties 66 54
Cancer diagnosis and care receipt
 % of cases diagnosed at stage II among early- stage breast cancer patients 42.2 (0.8) 46.9*** (0.8) 0.1194†
 % Receiving BCS without radiation therapy 29.0 (0.5) 42.5*** (0.9) 0.5893†
Cancer care resources
 Radiation oncologists per 100,000 residents 1.6 (1.58) 0.5* (1.1) −0.0223
 Surgeons per 100,000 residents 60.5 (45.30) 23.4*** (24.6) −0.0853
 Hospitals with radiation therapy facilities per 100,000 residents 0.6 (0.58) 0.3*** (0.7) −0.0013
 Hospitals with oncology services per 100,000 residents 1.3 (1.01) 0.8 (1.3) −0.0896
Other county-level characteristics
 % in poverty 11.1 (2.5) 20.9*** (5.2) 0.7610†
 Medicare HMO penetration rate 6.3 (7.51) 3.0** (4.1) 0.4343†
 % counties not containing an interstate highway 53.0 (6.14) 77.8** (5.7) N/A
Open in a new tab

Note. Values are means/percentages (standard errors). The p-values refer to t-test or chi-square comparisons of Appalachian counties versus non-Appalachian counties.

*p < .05; **p < .01; ***p < .001.

†

Moran's I is significant.

‡

Values nearer −1 indicate spatial dispersion, those near +1 suggest spatial clustering, and values close to 0 indicate a random spatial pattern.

Figure 1a,b shows that higher proportions of patients who did not receive radiation therapy after BCS were clustered in counties around the city of Lexington; most of these counties are east and south of the city. Figure 2a,b reveals that radiation therapy resources were scattered across Kentucky. A significant Moran's I test indicates nonrandom geographic distribution and spatial autocorrelation.

Figure 1.

Figure 1

Open in a new tab

(a,b) Rate of Breast-Conserving Surgery without Radiation in Kentucky: 2000–2007Note. According to Anselin (2004), “the high-high and low-low locations (positive local spatial autocorrelation) are typically referred to as spatial clusters, while the high-low and low-high locations (negative local spatial autocorrelation) are termed spatial outliers. Spatial clusters shown on the cluster map only refer to the core of the cluster. The cluster is classified as such when the value at a location (either high or low) is more similar to its neighbors (as summarized by the weighted average of the neighboring values, the spatial lag) than would be the case under spatial randomness. Any location for which this is the case is labeled on the cluster map. However, the cluster itself likely extends to the neighbors of this location as well.”

Figure 2.

Figure 2

Open in a new tab

(a,b) Availability of Radiation Therapy in Kentucky: 2000–2007

Table 1 shows a random spatial pattern of treatment resources and spatial correlation of poverty rates and Medicare HMO penetration rate. It also shows a spatial correlation in the proportion of stage II diagnosis. The Moran's I value for rates of BCS without radiation therapy is 0.5893 in Kentucky during 2000–2007 for early-stage breast cancer patients, indicating that clustering of rates of omitting radiation therapy after BCS exists in Appalachian Kentucky.

Regression Analyses

An OLS regression model with rates of BCS without radiation as the dependent variable had an adjusted R-squared 0.29 (Table 2, column 1). The multicollinearity condition number is 17.90, indicating that multicollinearity is a not serious concern in the analysis. The GeoDa regression diagnostics revealed the presence of spatial structure. Following the model specification steps outlined in Anselin (2004), a spatial lag model was identified as the most appropriate model form. As shown in Table 2, both the LM-Lag and LM-Error are significant, with the Robust LM-Lag statistic more significant than the Robust LM-Error statistic (p < .001 compared to p = .04). The spatial lag model increases the log likelihood from −509.95 (for OLS) to −481.79 (for the spatial lag model). Similarly, compensating for the added spatially lagged dependent variable the AIC (from 1039.90 to 985.59) and Schwartz Criterion (from 1067.77 to 1016.25) tests both decrease relative to the OLS model, again suggesting an improvement of fit for the spatial lag model specification. The spatial autoregressive coefficient is estimated as 0.71 and is highly significant (p < .001). The Likelihood Ratio Test compares the OLS model to the alternative spatial lag model. The value of 56.31 confirms the strong significance of the spatial autoregressive coefficient.

Table 2.

Predictors of Rates of Breast-Conserving Surgery without Radiation Therapy (N = 120)

Variables OLS Model Spatial Lag Model
Spatially lagged rate of BCS without radiation 0.71*** (0.07)
Constant 36.94*** (10.54) 12.51 (7.81)
Appalachian status 7.40** (2.35) 2.26 (1.92)
Radiation oncologists per 100,000 residents 2.74 (2.35) 2.59 (1.66)
Hospitals with radiation therapy facilities per 100,000 residents −16.06*** (3.87) −8.70** (2.76)
Surgeons per 100,000 residents 0.19 (0.10) 0.03 (0.07)
Hospitals with oncology services per 100,000 residents 0.17 (1.29) 0.03 (0.92)
% of cases diagnosed at stage II among early stage breast cancer patients 0.16 (0.18) −0.05 (0.13)
% in poverty −0.27 (0.73) 0.03 (0.30)
Medicare HMO penetration rate −0.72 (0.40) −0.16 (0.28)
% counties not containing an interstate highway −5.82 (3.88) −1.48 (2.75)
R-squared (adjusted R-squared) 0.29 (0.23)
Multicollinearity condition number 17.90
AIC 1039.90 985.59
Log likelihood −509.95 −481.79
Schwarz criterion 1067.77 1016.25
Diagnostics for spatial dependence
 Lagrange multiplier (lag) 63.75***
 Robust LM (lag) 21.63***
 Lagrange multiplier (error) 46.04***
 Robust LM (error) 3.91*
Likelihood ratio test for spatial dependence 56.31***
Open in a new tab

Note. Values are coefficients (Std. error).

*p < .05; **p < .01; ***p < .001.

A consistent significant negative association is seen between the number of radiation therapy hospitals per 100,000 residents and rates of omitting radiation after BCS, controlling for Appalachian location and other variables in the model. That is, women are more likely to receive radiation after BCS when there are more radiation therapy hospitals available that provide radiation therapy. However, the number of radiation therapy hospitals per 100,000 residents in the spatial lag model is less significant than in the OLS model (p < .0001). The magnitude of the estimated coefficients decreased in absolute value (from −16.06 to −8.70). The Appalachian designation and the constant variable are no longer significant in the spatial lag model. The explanatory power of these variables that was attributed to their in-county values in the OLS model is reduced in the spatial lag model, where the spatial spillover effect is captured by the coefficient of the spatially lagged dependent variable.

Discussion

The ecological models estimated in this study contribute new insights by providing information on spatial relationships between counties with regard to cancer care resources and outcomes in Appalachia. Descriptive analyses of Kentucky data revealed that the likelihood of BCS without radiation was significantly higher among early-stage breast cancer patients in Appalachian counties than among their non-Appalachian counterparts. These findings are consistent with a recent study by Freeman, Huang, and Dragun (2012), who examined the use of adjuvant radiation therapy among Appalachian Kentucky residents. The Freeman et al. paper (2012), however, did not make specific comparisons between Appalachian and non-Appalachian patients. Spatial analysis shows that clustering of high percentages of patients failing to receive radiation therapy after BCS exists in Kentucky, especially in counties around the city of Lexington. This geographic variation in guideline-concordant breast cancer therapy warrants further investigation. For example, breast cancer patients could be surveyed in future research to learn more about patient characteristics, available facilities, and the communication process between the patient and her breast cancer surgeon related to potential barriers in accessing radiation therapy. This information could then be linked to cancer registry data to identify patient, tumor, provider, and facility factors that contribute to significant variation in receipt of the guideline-concordant care for breast cancer in medically underserved areas.

When other variables such as Appalachian status and surgeons per 100,000 residents are taken into account, having fewer radiation therapy resources is associated with lower rates of radiotherapy among women with BCS in Kentucky. This is consistent with previous work suggesting that more limited cancer care resources has an important influence on disparities in cancer treatment receipt (Friedell et al. 2001; Armstrong et al. 2004; Lengerich et al. 2004; Mayo et al. 2004; Coughlin et al. 2006; Lyttle and Stadelman 2006; Kelly et al. 2009; McAlearney et al. 2010). This finding may also apply to other Appalachian states where no cancer treatment data comparable to SEER are available. In supplemental analyses based on Area Resource File data from 13 states, we found that there are large areas of Appalachia with no apparent radiation oncology resources. We also found that southern Appalachian counties in Alabama, Georgia, Mississippi, and South Carolina had significant fewer hospitals with radiation therapy facilities than Northern Appalachian counterparts in New York, Ohio, and Pennsylvania (tables upon request). Our results based on Kentucky suggest that policy action is needed to improve the cancer treatment infrastructure, including the availability of radiation therapy resources in disadvantaged mountainous areas.

This study has several limitations. First, because this is an ecological study, associations between cancer treatment resources and guideline-concordant radiotherapy cannot be interpreted as demonstrating associations at the individual level. However, ecological studies such as this are often viewed as providing useful information at the population level for policy makers (Szklo and Nieto 2007; Yao, Lengerich, and Hillemeier 2012). Second, there are additional variables that may influence receipt of radiation after BCS that we were not able to include, such as the availability and use of various modes of transportation. Radiation therapy after BCS for a subset of older women with specific tumor characteristics and treatment history is controversial; however, supplemental analyses showed no evidence that differences in age distribution in the study samples influenced our results. Another limitation has to do with measuring use of radiation therapy. Registry data may miss cases of radiation therapy—especially in smaller hospitals. The missing rate in SEER data is not large (Warren, Klabunde et al. 2002). Finally, the sample size (N = 120) in the regression analysis is relatively small. The small sample size precludes a detailed analysis of patterns in Appalachia versus non-Appalachia based on spatial regimes models. Additional research with a bigger sample size (requiring more counties across a wider geographic area) and more complete information on a range of covariates is needed if researchers are to gain a more complete understanding of patterning of radiotherapy omission after BCS. Our dataset was constructed for an area based on SEER data, which unfortunately are not available for large areas of Appalachia.

In conclusion, there are large areas with no radiation oncology resources in Appalachian Kentucky, which is likely to be associated with disparities in the receipt of guideline-concordant radiotherapy. Public health intervention is warranted to improve the availability of breast cancer treatment resources in areas of scarcity in Appalachia and elsewhere. Federal and state governments may motivate community hospitals to provide radiation therapy service by assisting to pay for some of their costs. For example, the government may provide incentive to community hospitals that only provide BCS but no radiation therapy services. The public health sector can indirectly improve the availability of radiation therapy by subsidizing transportation to and from radiation therapy facility through comprehensive cancer centers in Appalachia. Radiation therapy (whole breast irradiation) usually involves a 4–7 week course of daily treatment, depending on the schedule chosen. It has been suggested that the travel burden is an important cause of lack of adherence for receiving appropriate adjuvant radiation therapy (Arthur and Vicini 2005). Direct and indirect improvement of the supply of radiation therapy facilities may help improve rates of receiving guideline-concordant care among cancer patients in medically underserved areas.

Acknowledgments

Joint Acknowledgment/Disclosure Statement: Nengliang Yao's doctoral training is partly supported by a predoctoral training grant sponsored by the Susan G. Komen Foundation. Additional support has been provided by Stephen A. Matthews at Population Research Institute, which receives core funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development Award R25–HD41025.

Disclosure: None.

Disclaimer: None.

SUPPORTING INFORMATION

Additional supporting information may be found in the online version of this article:

Appendix SA1: Author Matrix.

Appendix SA2: Histologies Likely To Be Treated with Radiation after BCS.

hesr0048-1433-SD1.pdf (21.8KB, pdf)
hesr0048-1433-SD2.docx (15.8KB, docx)

References

  1. Allen DW. “Social Class, Race, and Toxic Releases in American Counties, 1995”. The Social Science Journal. 2001;38(1):13–25. [Google Scholar]
  2. Andersen RM, McCutcheon A, Aday LA, Chiu GY, Bell R. “Exploring Dimensions of Access to Medical Care”. Health Services Research. 1983;18(1):49–74. [PMC free article] [PubMed] [Google Scholar]
  3. Anderson RT, Kimmick GG, Camacho F, Whitmire JT, Dickinson C, Levine EA, Torti FM, Balkrishnan R. “Health System Correlates of Receipt of Radiation Therapy after Breast-Conserving Surgery: A Study of Low-Income Medicaid-Enrolled Women”. American Journal of Managed Care. 2008;14(10):644–52. [PubMed] [Google Scholar]
  4. Anselin L. Spatial Econometrics, Methods, and Models. Dordrecht, The Netherlands: Kluwer Academic; 1988. [Google Scholar]
  5. Anselin L. “Local Indicators of Spatial Association\LISA”. Geographical Analysis. 1995;27(2):93–115. [Google Scholar]
  6. Anselin L. 2004. “Exploring Spatial Data with GeoDaTM: A Workbook” [accessed on January 30, 2012]. Available at https://geodacenter.asu.edu/system/files/geodaworkbook.pdf.
  7. Anselin L, Syabri I, Kho Y. “GeoDa: An Introduction to Spatial Analysis”. Geographical Analysis. 2006;38(1):5–22. [Google Scholar]
  8. Appalachian Regional Commission. 2010. “The Appalachian Region” [accessed on 19 November, 2010]. Available at http://www.arc.gov/appalachian_region/TheAppalachianRegion.asp.
  9. Armstrong LR, Thompson T, Hall HI, Coughlin SS, Steele B, Rogers JD. “Colorectal Carcinoma Mortality among Appalachian Men and Women, 1969–1999”. Cancer. 2004;101(12):2851–8. doi: 10.1002/cncr.20667. [DOI] [PubMed] [Google Scholar]
  10. Arthur DW, Vicini FA. “Accelerated Partial Breast Irradiation as a Part of Breast Conservation Therapy”. Journal of Clinical Oncology. 2005;23(8):1726. doi: 10.1200/JCO.2005.09.045. [DOI] [PubMed] [Google Scholar]
  11. Baller RD, Anselin L, Messner SF, Deane G, Hawkins DF. “Structural Covariates of U.S. County Homicide Rates: Incorporating Spatial Effects”. Criminology. 2001;39(3):561–88. [Google Scholar]
  12. Bickell NA, Wang JJ, Oluwole S, Schrag D, Godfrey H, Hiotis K, Mendez J, Guth AA. “Missed Opportunities: Racial Disparities in Adjuvant Breast Cancer Treatment”. Journal of Clinical Oncology. 2006;24(9):1357–62. doi: 10.1200/JCO.2005.04.5799. [DOI] [PubMed] [Google Scholar]
  13. Coughlin SS, Costanza ME, Fernandez ME, Glanz K, Lee JW, Smith SA, Stroud L, Tessaro I, Westfall JM, Weissfeld JL, Blumenthal DS. “CDC-Funded Intervention Research Aimed at Promoting Colorectal Cancer Screening in Communities”. Cancer. 2006;107(5 suppl):1196–204. doi: 10.1002/cncr.22017. [DOI] [PubMed] [Google Scholar]
  14. Denham SA, Wood LE, Remsberg K. “Diabetes Care: Provider Disparities in the US Appalachian Region”. Rural and Remote Health. 2010;10(2):1320. [PubMed] [Google Scholar]
  15. Early Breast Cancer Trialists' Collaborative Group. “Effects of Radiotherapy and Surgery in Early Breast Cancer: An Overview of the Randomized Trials. Early Breast Cancer Trialists' Collaborative Group”. New England Journal of Medicine. 1995;333(22):1444–55. doi: 10.1056/NEJM199511303332202. [DOI] [PubMed] [Google Scholar]
  16. Enger SM, Thwin SS, Buist DS, Field T, Frost F, Geiger AM, Lash TL, Prout M, Yood MU, Wei F, Silliman RA. “Breast Cancer Treatment of Older Women in Integrated Health Care Settings”. Journal of Clinical Oncology. 2006;24(27):4377–83. doi: 10.1200/JCO.2006.06.3065. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. ESRI. ArcGIS Desktop: Release 10. Redlands, CA: Environmental Systems Research Institute; 2012. [Google Scholar]
  18. Farrow DC, Hunt WC, Samet JM. “Geographic Variation in the Treatment of Localized Breast Cancer”. New England Journal of Medicine. 1992;326(17):1097–101. doi: 10.1056/NEJM199204233261701. [DOI] [PubMed] [Google Scholar]
  19. Fisher B, Anderson S, Bryant J, Margolese RG, Deutsch M, Fisher ER, Jeong JH, Wolmark N. “Twenty-Year Follow-Up of a Randomized Trial Comparing Total Mastectomy, Lumpectomy, and Lumpectomy Plus Irradiation for the Treatment of Invasive Breast Cancer”. New England Journal of Medicine. 2002;347(16):1233–41. doi: 10.1056/NEJMoa022152. [DOI] [PubMed] [Google Scholar]
  20. Freeman AB, Huang B, Dragun AE. “Patterns of Care With Regard to Surgical Choice and Application of Adjuvant Radiation Therapy for Preinvasive and Early Stage Breast Cancer in Rural Appalachia”. American Journal of Clinical Oncology. 2012;35(4):358–63. doi: 10.1097/COC.0b013e3182118d27. [DOI] [PubMed] [Google Scholar]
  21. Friedell GH, Rubio A, Maretzki A, Garland B, Brown P, Crane M, Hickman P. “Community Cancer Control in a Rural, Underserved Population: The Appalachian Leadership Initiative on Cancer Project”. Journal of Health Care for the Poor and Underserved. 2001;12(1):5–19. doi: 10.1353/hpu.2010.0523. [DOI] [PubMed] [Google Scholar]
  22. Hall HI, Uhler RJ, Coughlin SS, Miller DS. “Breast and Cervical Cancer Screening among Appalachian Women”. Cancer Epidemiology, Biomarkers and Prevention. 2002;11(1):137–42. [PubMed] [Google Scholar]
  23. Halverson J, Ma L, Harner E. An Analysis of Disparities in Health Status and Access to Health Care in the Appalachian Region. Washington, DC: Appalachian Regional Commission; 2004. [Google Scholar]
  24. Kelly KM, Love MM, Pearce KA, Porter K, Barron MA, Andrykowski M. “Cancer Risk Assessment by Rural and Appalachian Family Medicine Physicians”. Journal of Rural Health. 2009;25(4):372–7. doi: 10.1111/j.1748-0361.2009.00246.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lengerich E, Wyatt S, Rubio A, Beaulieu J, Coyne C, Fleisher L, Ward A, Brown P. “The Appalachia Cancer Network: Cancer Control Research among a Rural, Medically Underserved Population”. Journal of Rural Health. 2004;20(2):181–7. doi: 10.1111/j.1748-0361.2004.tb00026.x. [DOI] [PubMed] [Google Scholar]
  26. Lengerich EJ, Tucker TC, Powell RK, Colsher P, Lehman E, Ward AJ, Siedlecki JC, Wyatt SW. “Cancer Incidence in Kentucky, Pennsylvania, and West Virginia: Disparities in Appalachia”. Journal of Rural Health. 2005;21(1):39–47. doi: 10.1111/j.1748-0361.2005.tb00060.x. [DOI] [PubMed] [Google Scholar]
  27. Lyttle NL, Stadelman K. “Assessing Awareness and Knowledge of Breast and Cervical Cancer among Appalachian Women”. Preventing Chronic Disease. 2006;3(4):A125. [PMC free article] [PubMed] [Google Scholar]
  28. Mandelblatt JS, Yabroff KR, Kerner JF. “Equitable Access to Cancer Services: A Review of Barriers to Quality Care”. Cancer. 1999;86(11):2378–90. [PubMed] [Google Scholar]
  29. Mayo RM, Sherrill WW, Crew L, Watt P, Mayo WW. “Connecting Rural African American and Hispanic Women to Cancer Education and Screening: The Avon Health Connector Project”. Journal of Cancer Education. 2004;19(2):123–6. doi: 10.1207/s15430154jce1902_14. [DOI] [PubMed] [Google Scholar]
  30. McAlearney AS, Song PH, Rhoda DA, Tatum C, Lemeshow S, Ruffin M, Harrop JP, Paskett ED. “Ohio Appalachian Women's Perceptions of the Cost of Cervical Cancer Screening”. Cancer. 2010;116(20):4727–34. doi: 10.1002/cncr.25491. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Moran PAP. “Notes on Continuous Stochastic Phenomena”. Biometrika. 1950;37(1/2):17–23. [PubMed] [Google Scholar]
  32. National Cancer Institute. 2010. “Center to Reduce Cancer Health Disparities. Appalachia Community Cancer Network” [accessed on January 12, 2012]. Available at http://crchd.cancer.gov/cnp/pi-dignan-description.html.
  33. National Cancer Institute. 2012. “Breast Cancer Treatment (PDQ®): Health Professional Version 2011.” [accessed on July 9, 2012]. Available at http://www.cancer.gov/cancertopics/pdq/treatment/breast/healthprofessional/page6.
  34. National Comprehensive Cancer Network. 2012. “NCCN Clinical Practice Guidelines in Oncology. Breast Cancer, Version 1, 2011” [accessed on February 9, 2012]. Available at http://www.nccn.org/professionals/physician_gls/pdf/breast.pdf.
  35. National Institute of Health. “NIH Consensus Conference. Treatment of Early-Stage Breast Cancer”. Journal of the American Medical Association. 1991;265(3):391–5. [PubMed] [Google Scholar]
  36. Smith GL, Xu Y, Shih YC, Giordano SH, Smith BD, Hunt KK, Strom EA, Perkins GH, Hortobagyi GN, Buchholz TA. “Breast-Conserving Surgery in Older Patients with Invasive Breast Cancer: Current Patterns of Treatment across the United States”. Journal of the American College of Surgeons. 2009;209(4):425–33. doi: 10.1016/j.jamcollsurg.2009.06.363. e2. [DOI] [PubMed] [Google Scholar]
  37. Susi L, Mascarenhas AK. “Using a Geographical Information System to Map the Distribution of Dentists in Ohio”. Journal of the American Dental Association. 2002;133(5):636–42. doi: 10.14219/jada.archive.2002.0239. [DOI] [PubMed] [Google Scholar]
  38. Szklo M, Nieto FJ. Epidemiology: Beyond the Basics. Sudbury, MA: Jones & Bartlett Learning; 2007. [Google Scholar]
  39. U.S. Department of Health and Human Services. 2010. “Find Shortage Areas: HPSA by State & County” [accessed on December 22, 2010]. Available at http://hpsafind.hrsa.gov/HPSASearch.aspx. [DOI] [PubMed]
  40. US Department of Health and Human Services. Area Resource File (ARF). 2009–2010. Rockville, MD: 2011. [Google Scholar]
  41. Veronesi U, Marubini E, Mariani L, Galimberti V, Luini A, Veronesi P, Salvadori B, Zucali R. “Radiotherapy after Breast-Conserving Surgery in Small Breast Carcinoma: Long-Term Results of a Randomized Trial”. Annals of Oncology. 2001;12(7):997–1003. doi: 10.1023/a:1011136326943. [DOI] [PubMed] [Google Scholar]
  42. Vinh-Hung V, Voordeckers M, Soete J, Van de Steene G, Lamote J, Storme G. “Omission of Radiotherapy after Breast-Conserving Surgery: Survival Impact and Time Trends”. Radiotherapy and Oncology. 2003;67(2):147–58. doi: 10.1016/s0167-8140(03)00002-1. [DOI] [PubMed] [Google Scholar]
  43. Wang F, Luo W. “Assessing Spatial and Nonspatial Factors for Healthcare Access: Towards an Integrated Approach to Defining Health Professional Shortage Areas”. Health and Place. 2005;11(2):131–46. doi: 10.1016/j.healthplace.2004.02.003. [DOI] [PubMed] [Google Scholar]
  44. Ward MD, Gleditsch KS. Spatial Regression Models. Thousand Oaks, CA: Sage Publications, Inc; 2008. [Google Scholar]
  45. Warren JL, Klabunde CN, Schrag D, Bach PB, Riley GF. “Overview of the SEER-Medicare Data: Content, Research Applications, and Generalizability to the United States Elderly Population”. Medical Care. 2002;40(8 suppl) doi: 10.1097/01.MLR.0000020942.47004.03. IV-3–18. [DOI] [PubMed] [Google Scholar]
  46. Weiss MC, Fowble BL, Solin LJ, Yeh IT, Schultz DJ. “Outcome of Conservative Therapy for Invasive Breast Cancer by Histologic Subtype”. International Journal of Radiation Oncology Biology Physics. 1992;23(5):941–7. doi: 10.1016/0360-3016(92)90898-r. [DOI] [PubMed] [Google Scholar]
  47. Wingo P, Tucker T, Jamison P, Martin H, McLaughlin C, Bayakly R, Bolick Aldrich S, Colsher P, Indian R, Knight K. “Cancer in Appalachia, 2001–2003”. Cancer. 2008;112(1):181–92. doi: 10.1002/cncr.23132. [DOI] [PubMed] [Google Scholar]
  48. Yao N, Lengerich EJ, Hillemeier MM. “Breast Cancer Mortality in Appalachia: Reversing Patterns of Disparity over Time”. Journal of Health Care for the Poor and Underserved. 2012;23(2):715–25. doi: 10.1353/hpu.2012.0043. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Zerhouni EA, Ruffin J. Strategic Research Plan and Budget to Reduce and Ultimately Eliminate Health Disparities. Volume I: Fiscal Years 2002–2006. Bethesda, MD: National Institutes of Health; 2002. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

hesr0048-1433-SD1.pdf (21.8KB, pdf)
hesr0048-1433-SD2.docx (15.8KB, docx)

Articles from Health Services Research are provided here courtesy of Health Research & Educational Trust

RESOURCES