Conos Normales, Regulares y Completamen te Regulares

As stated previously, the significance of these research findings results from the establishment of the complex interaction of variables dependent on place. Prior to the accommodation of spatial variations, it is important to first consider how the variables interact with each other in a place. This involves reassessing the conceptual model from the beginning. Figure 8.1 displays the revised concept model for cancer disparities based on this research. Following the diagram is a description of each component and the interactions.

After looking back through the results of the research it became apparent that establishing predetermined relationships among the factors would likely have a negative impact on future endeavors to assess cancer disparities. The idea behind running a spatial analysis was to, in part, establish the variability of relationships between societal

characteristics and the MIR. Having shown this variability to exist at the U.S. scale, it would not make any sense to then create a conceptual model implying specific

144

Figure 8.1. Revised concept map based on results presented.

An inductive approach does appear to be the best approach to determine the strength an order of relationships that best describe the societal constructs leading to cancer disparities in a place. Based on this thought, the best model is one in which all variables are assumed as equal contributors. A correlation between the variables is also assumed to exist.

The new model does not propose any starting point in the determination of cancer disparities. While there is some evidence from the path analysis to support placing financial and medical access first in the model, there is no guarantee that the same relationship will occur at different scales. In addition, there is a good deal of evidence presented in the spatial analysis that suggests different relationships between the factors and the MIR that are dependent on place. Taking this into consideration, a more suitable

145

model should refrain from the use of directionality. Instead of having a definitive starting point, each of the factors in this model is displayed as having a direct influence on cancer disparities. The strength of the connection is not implied because it may change.

A benefit of this model for future research is its ability to be operationalized. There is a great deal of potential to use this model to test high and low MIR counties against each other, for example. The relationships among factors and the MIR can be assessed to determine specific alterations that may exist, leading to the disparities. Using this proposed model to establish strengths of relationships and to run path analyses, models specific to a place could be created. This research can serve as an example of this model specification. At the U.S. level, there were specific interactions evident among the factors and the MIR that could be used to create a model explicit to this data set and geography. An example of this directional model is presented in Figure 8.2.

Figure 8.2. Conceptual Model created using U.S. county-level data.

The relationships found in this U.S. level research support a certain level of model directionality as well as specific correlations between the factors. Financial and medical

146

access characteristics would most likely form the base of this U.S. specific model along with social characteristics as determined through the structural equation modeling. There are strong factor correlations between the health and behavioral characteristics, the financial and medical access characteristics, and the social characteristics that suggest a close relationship among the three. This relationship was confirmed through analysis of factor correlations as well as through visual analysis using hot spot maps of the factors. The community and environmental characteristics shift in the new model as well and act as a filter through which the cancer disparities are ultimately defined. The

evidence for the movement of this factor to the last position in the model comes primarily from the spatial analysis. The hot spot maps revealed a pattern in the MIR that coincided with the urban/rural divide, which is the most significant indicator within the community factor. Despite this spatial correlation the community factor did not predict the MIR well at all. Not all rural areas corresponded to higher cancer fatalities even though they are easily distinguishable on the map. Based on this evidence, it appears as though the community factor is highly dependent on the characteristics present in the other three factors.

Testing the function of this urban/rural divide as a filter for the other factors is the next step in better defining the drivers of cancer disparities. The characteristics of a place have proven to influence the patterns of interaction between variables that lead to cancer fatalities. This precludes any future large-scale research and demands a shift to regional analysis. Urban and rural areas should be studied independent of each other with no a priori assumptions as to the influence of variables on cancer outcomes. A PCA run on all known indicators of the MIR will identify the variables most likely contributing to the

147

disparities and inform a path analysis to derive the underlying cause of the higher or lower fatality rate. Regional analysis based on the urban/rural divide should lead to two different sets of models, each modified to reflect the influence of variables in that place. The two resulting models can be compared in order to clearly establish unique patterns in each location. In the end, the goal is to eradicate the disparities in cancer deaths and ensure that these rates drop equitably for all groups. Identifying a location where rates are significantly different is only half of the battle, and should only be useful to establish the causes of the difference. If the reasons change based on location, as was proven, the use of a place-based model is essential in order to first establish the significant variables and then predict how they will influence the outcomes.

8.6. Caveats

There are some limits to note in this research based on the available data and the geography. First, the county level analysis conducted in this research will not reflect trends in smaller geographic areas. There is a lot of potential variability within counties, and this certainly causes smaller scale trends to be hidden in the data. Running smaller scale analysis in the future would allow for this problem to be circumvented. Data availability for many of the variables was responsible for the county level measure. The goal from the beginning was to establish a method and model that would allow scaling. This would be dependent on data availability and the needs of the local or regional medical staff or other individuals concerned with caner vulnerability.

The use of all-cancer rates also presents some potential issues in the interpretation of results. Not all cancers have similar incidence or mortality rates, and therefore certain cancer types may dominate the data. For example, breast cancer and prostate cancer

148

higher incidence rates, yet relatively low mortality rates. Because of this relationship, these two cancers will drive many of the trends visible in this research. As stated previously, however, the goal of this research was to establish overall cancer disparities, regardless of cancer type. Breast cancer and prostate cancer can be fatal if not treated properly or diagnosed too late. Smaller scale studies may benefit from breaking out specific cancers, but this should be done after an initial analysis of the cancer prevalence in the area and the societal characteristics present.

There exists an assumption in this research that geographic variability stems directly, and solely, from the factors being studied. In reality, there may be an element of chance, or general variability present in the outcomes that has no connection to the

factors. This variability potential is greater in this study due to the size of the data set and the spatial extent of the data, and is not accounted for.

Another limitation of this research lies in the use of some transformed variables. Although the data is intended as a comparative analysis, looking at the data values with respect to the scores of others in the set, the process of transformation can cause some issues in the interpretability of the data. It was deemed more important to have all linear variables and be able to conduct a linear regression model than to adapt the model to the presence of non-linear variables and lose a good deal of the predictive power.

Lastly, the inference of a causal relationship between the identified drivers and the outcomes does not guarantee its actual existence. There are a multitude of

possibilities that could exist. Aggregating up to the county level can allow for ecological fallacies, meaning that many of the county level assumptions may not apply to smaller scale patterns. Individual-level assumptions cannot be made from this research. Also,

149

inherent in geographic research of this variety is the assumption of treatment in a local area. Mobility of the population cannot be fully accounted for, and people could easily cross county lines for medical services. Data inaccuracies can also play a role in the accuracy of the research. Registry data, for instance, relies on proper diagnosis codes and accuracy in reporting could create problems.

Another major consideration that must be accounted for in future research is the cancer type. This study used the all-cancer incidence and mortality rates for each county. The predominant cancer types in each county undoubtedly sway the data analysis. Breast cancer may have higher prevalence in a specific area and result in an inaccurate

assumption that mammograms are highly correlated, for instance. Adding cancer type to the analysis will be pertinent to the regional analysis. It is likely that certain regions will possess one or two predominant cancer types. In this case, a model should be

constructed for each, as it is probable that different variables will have different contributions dependent on the cancer type in addition to the place.

150

REFERENCES

Adger, W.N. 2006. “Vulnerability.” Global Environmental Change-Human and Policy

Dimensions. 16(3): 268-281.

Allison, P.D. 1998. Multiple Regression: A Primer. SAGE Publications, Inc. American Cancer Society. 2010. Cancer Facts & Figures 2010. Atlanta: American

Cancer Society.

American Cancer Society. 2012. Cancer Facts & Figures 2012. Atlanta: American Cancer Society.

Amodeo, G., Camera, A., and G. Caimi. 2010. “Physical activity and cancer.” Medicina

Dello Sport. 63(4): 591-600.

Ananthakrishnan, A.N., Hoffmann, R.G., and K. Saeian. 2010. “Higher Physician Density is Associated with Lower Incidence of Late-stage Colorectal Cancer.”

Journal of General Internal Medicine. 25(11): 1164-1171.

Anderson, M.B. 2000. ‘‘Vulnerability to Disaster and Sustainable Development: A General Framework for Assessing Vulnerability,’’ in Storms. ed. Roger A. Pielke, and Roger A. Pielke. London: Routledge.

Aneja, S., Gross, C.P., Soulos, P.R., and J.B. Yu. 2011. “Geographical Information Systems: Applications and Limitations in Oncology Research.” Oncology - New

York. 25(12): 1221-1225.

Armitage, P., and R. Doll. 1954. “The Age Distribution of Cancer and a Multi-stage theory of carcinogenesis.” British Journal of Cancer. 8(1): 1-12.

Auchincloss, A.H., Gebreab, S.Y., Mair, C., and A.V.D. Roux. 2012. “A Review of Spatial Methods in Epidemiology, 2000-2010.” In Annual Review of Public

Health. Vol. 33. edited by J. E. Fielding. Palo Alto: Annual Reviews.

Banegas, M.P., and C.I. Li. 2012. “Breast cancer characteristics and outcomes among Hispanic Black and Hispanic White women.” Breast Cancer Research and

Treatment. 134(3): 1297-1304.

Beaulac, J., Kristjansson, E., and S. Cummins. 2009. A systematic review of food deserts, 1966-2007. Prevention of Chronic Disease. 6(3): A105.

151

Berger, F., Doussau, A., Gautier, C., Gros, F., Asselain, B., and F. Reyal. 2012.“Impact of socioeconomic status on stage at diagnosis of breast cancer.” Revue D

Epidemiologie Et De Sante Publique. 60(1): 19-29.

Belsley, D.A. 1991. Conditioning diagnostics: Collinearity and weak data in regression. New York: Wiley.

Block, G., Patterson, B., and A. Subar. 1992. “Fruit, vegetables, and cancer prevention-a review of the epidemiologic evidence.” Nutrition and Cancer-an International

Journal. 18(1): 1-29.

Bouchardy, C., Schuler, G., Minder, C., Hotz, P., Bousquet, A., Levi, F., Fisch, T., Torhorst, J., and L. Raymond. 2002. “Cancer risk by occupation and

socioeconomic group among men - a study by The Association of Swiss Cancer Registries.” Scandinavian Journal of Work Environment & Health. 28: 1-88. Brawley, O.W. 2012. “Prostate cancer epidemiology in the United States.” World

Journal of Urology. 30(2): 195-200.

Burton, I., Kates, R.W., and G.F. White. 1993. The environment as hazard. New York: Guilford Press.

Carbone, D. 1992. “The Effects of Cigarette Smoking: A Global Perspective.” The

American Journal of Medicine. 93(1): S13-S17.

Calle, E.E., and M.J. Thun. 2004. Obesity and cancer. Oncogene. 23: 6365–6378

Chamie, K., Kwan, L., Connor, S.E., Zavala, M., Labo, J., and M.S. Litwin. 2012. “The impact of social networks and partnership status on treatment choice in men with localized prostate cancer.” British Journal of Urology International. 109(7): 1006-1012.

Chen, A.Y., DeSantis, C., and A. Jemal. 2011. “US Mortality Rates for Oral Cavity and Pharyngeal Cancer by Educational Attainment.” Archives of Otolaryngology-

Head & Neck Surgery. 137(11): 1094-1099.

Clarke, K.C., McLafferty, S.L., and B.J. Tempalski. 1996. “On epidemiology and geographic information systems: A review and discussion of future directions.”

Emerging Infectious Diseases. 2(2): 85-92.

Coleman, D., and R. Rowthorn. 2011. “Who's Afraid of Population Decline? A Critical Examination of Its Consequences.” Population and Development Review. 37: 217-248.

152

Cook, M.B., McGlynn, K.A., Devesa, S.S., Freedman, N.D., and W.F. Anderson. 2011. “Sex Disparities in Cancer Mortality and Survival.” Cancer Epidemiology

Biomarkers & Prevention. 20(8): 1629-1637.

Cronbach, L.J. 1951. “Coefficiant Alpha and the Internal Structure of Tests.”

Psychometrika. 16(3): 297-334.

Cutter, S.L., Mitchell, J.T., and M.S. Scott. 2000. “Revealing the Vulnerability of

People and Places: A Case Study of Geogretown County, South Carolina.” Annals

of the Association of American Geographers. 90(4): 713-737.

Cutter, S.L. 1996. “Vulnerability to environmental hazards.” Progress in Human

Geography. 20(4): 529-539.

Cutter, S.L., Barnes, L., Berry, M., Burton, C., Evans, E., Tate, E., and J. Webb. 2008. “A place-based model for understanding community resilience to natural

disasters.” Global Environmental Change-Human and Policy Dimensions. 18(4): 598-606.

Cutter, S.L., Boruff, B.J., and W.L. Shirley. 2003. “Social vulnerability to environmental hazards.” Social Science Quarterly. 84(2): 242-261.

Cutter, S.L., Burton, C.G., and C.T. Emrich. 2010. “Disaster Resilience Indicators for benchmarking baseline conditions.” Journal of Homeland Security and

Emergency Management. 7(1): Article 51.

Davis, A.M., Vinci, L.M., Okwuosa, T.M., Chase, A.R., and E.S. Huang. 2007. “Cardiovascular health disparities - A systematic review of health care interventions.” Medical Care Research and Review. 64(5): 29S-100S.

Do, D.P., Frank, R., and B.K. Finch. 2012. “Does SES explain more of the black/white health gap than we thought? Revisiting our approach toward understanding racial disparities in health.” Social Science & Medicine. 74(9): 1385-1393.

Eakin, H., A.L. Leurs. 2006. “Assessing the vulnerability of social-environmental systems.” Annual Review of Environment and Resources. 31: 365-394 Ell, K., Nishimoto, R., Mediansky, L., Mantell, J., and M. Hamovitch. 1992. “Social

relations, social support and survival among patients with cancer.” Journal of

Psychosomatic Research. 36(6): 531-541.

Elliott, T.E., Elliott, B.A., Renier, C.M. and I.V. Haller. 2004. “Rural-urban differences in cancer care: results from the Lake Superior Rural Cancer Care Project.”

153

Fay, M.P. 1991. “Approximate confidence intervals for rate ratios from directly standardized rates with sparse data.” Communications in Statistics-Theory and

Methods. 28(9): 2141-2160.

Frank, S.A. 2004. “Age-specific acceleration of cancer.” Current Biology. 14(3): 242- 246.

Freeman, H.P. 2007. “Patient Navigation: A Community Based Strategy to Reduce Cancer Disparities.” Journal of Urban Health. 83(2): 139-141.

Galea, S. and D. Viahov. 2005. “Urban health: Evidence, challenges, and direction.”

Annual Review of Public Health. 26: 341-365.

Gold, D.R. and R. Wright. 2005. “Population disparities in asthma.” In Annual Review of

Public Health. Palo Alto: Annual Reviews.

Groenewegen, P.P., Van den Berg, A.E., Maas, J., et al. 2012. “Is a Green Residential Environment Better for Health? If So, Why?” Annals of the Association of

American Geographers. 102(5): 996-1003.

Grammich, C., Hadaway, K., Houseal, R., Jones, D.E., et al. 2012. “2010 U.S. Religion Census: Religious Congregations & Membership Study.” Association of

Statisticians of American Religious Bodies.

Grimmett, C. 2011. “Exercise and cancer survivorship.” British Journal of Hospital

Medicine. 72(4): 196-199.

Harper, S. and J. Lynch. 2010. “Methods for Measuring Cancer Disparities: Using Data Relevant to Healthy People 2010 Cancer-Related Objectives,” in NCI Cancer

Surveillance Monograph Series, Number 6. Bethesda, MD: National Cancer

Institute.

Hebert, J.R., Daguise, V.G., Hurley, D.M., et al. 2009. “Mapping Cancer Mortality-to- Incidence Ratios to Illustrate Racial and Sex Disparities in a High-Risk

Population.” Cancer. 115(11): 2539-2552.

Hecht, S.S. 2012. “Lung carcinogenesis by tobacco smoke.” International Journal of

Cancer. 131(12): 2724-2732.

Helquist, M. 2010. "Effectiveness of population-based service screening with mammography for women ages 40 to 49 years." Cancer. 118(4): 1169-1169. Hewett, K. and I. Burton. 1971. The Hazardousness of Place: A Regional Ecology of

154

Hiscock, R., Bauld, L., Amos, A., Fidler, J.A., and M. Munafò. 2012. “Socioeconomic status and smoking: a review.” Annual New York Academy of Science. 1248: 107- 23.

Howlader, N., Noone, A.M., Krapcho, M., et al (eds). 2010. SEER Cancer Statistics

Review, 1975-2008. National Cancer Institute. Bethesda, MD.

Howlader, N., Noone, A.M., Krapcho, M., et al. (eds). 2012. SEER Cancer Statistics

Review, 1975-2009 (Vintage 2009 Populations). National Cancer Institute.

Bethesda, MD.

Huckle, T., You, R.Q. and S. Casswell. 2010. “Socio-economic status predicts drinking patterns but not alcohol-related consequences independently.” Addiction. 105(7): 1192-202.

Jilcott, S.B., Moore, J.B., Shores, K.A., Imai, S. and D.A. McGranahan. 2011.

“Associations between natural amenities, physical activity, and body mass index in 100 North Carolina counties.” American Journal of Health Promotion. 26(1): 52-55.

Jew, S., AbuMweis, S.S. and P.J.H. Jones. 2009. “Evolution of the Human Diet: Linking Our Ancestral Diet to Modern Functional Foods as a Means of Chronic Disease Prevention.” Journal of Medicinal Food. 12(5): 925-934.

Kan, S., Howard, D., Kim, J., Payne, J.S., et al. 2010. “English language proficiency and lifetime mental health service utilization in a national representative sample of Asian Americans in the USA.” Journal of Public Health. 32(3): 431-439.

Kasperson, J.X., Kasperson, R.E. and B.L. Turner. 1995. Regions at Risk: Comparison of

Threatened Environments. New York: United Nations University Press.

Koball, H.L., Moiduddin, J.H., Goesling, B. and M. Besculides. 2010. “What do we know about the link between marriage and health? Journal of Family Issues. 31(8): 1019-1040.

Koroukian, S.M., Bakaki, P.M. and D. Raghavan. 2012. “Survival disparities by Medicaid status An Analysis of 8 Cancers.” Cancer. 118(17): 4271-4279.