McDonald, Yolanda Jane (2017-05). The Landscape Across the Continuum of Invasive Cervical Cancer Prevention: A Geospatial Approach. Doctoral Dissertation.
Thesis
In the United States (U.S.) over the past decade, approximately 40,000 women have died from a disease that is mostly preventable, invasive cervical cancer (ICC). Since the 1950's, with the introduction of the Papanicolaou (Pap) screening test, there has been a >= 60% decline in the incidence rate. Screening is highly effective in reducing the incidence of ICC, but it is only one component of the multi-step process of the "continuum of care" (i.e. screening, diagnosis, and treatment). There is increasing concern about why women in the U.S. are still dying from ICC and why minority women are disproportionately burdened with higher mortality rates. To address these concerns, a multidisciplinary approach informed by theoretical frameworks and methods in GISHealth, health geography, epidemiology, and sociology were used to examine three aims: (1) characterize the effort to manually geocode healthcare facilities that provided invasive cervical cancer preventive services; (2) examine if geographic accessibility, defined as travel time and travel distance, differs for women who live in rural areas as opposed to women who live in non-rural areas; and (3) determine if socioeconomic status, acculturation, race/ethnicity, and geography are associated with ICC through the examination of main effects and interactions using a case-control research design. I found that there was a 90% improvement in geocode quality type and a corresponding spatial shift that ranged from 0.02 to 151,368 meters after manual intervention of geocoding of healthcare facilities, which took 42 hours of processing time. The Mann-Whitney test confirmed that there was a significant (P < 0.001) difference in travel time for all services across the ICC continuum of care comparing women who resided in rural areas to women who resided in non-rural areas. Multivariable conditional logistic regression results showed that female-headed households, not having health insurance, being a Hispanic woman, being a non-Hispanic white woman, travel time to a cervical cancer screening facility, living in a rural area as opposed to a non-rural area, and low educational attainment had increased odds while living in poverty and "speaking Spanish but not speaking English well" had decreased odds with ICC (statistically significant at P < 0.050).
