Cancer Capacity and Resources in Rural Arizona

Abstract

Background: Cancer is the second leading cause of death in the U.S. Rural urban cancer disparities exist nationally and in Arizona. Previous studies on the availability of rural cancer services are cancer-specific, limited to specific points along the cancer care continuum (e.g., screening, diagnosis, or treatment), or require updating to capture the current landscape as it relates to rural health. This study explored the following three aims: 1) Describe the availability of cancer capacity and services for breast, cervical, colorectal and lung cancer across the cancer care continuum in low populous counties in Arizona; 2) Evaluate the association between breast cancer capacity and resources with breast cancer incidence and mortality; and 3) Provide policy recommendations focused on increasing capacity and resources in low populous, rural counties in Arizona. Methods: For Aim 1, a cancer capacity and resources survey was developed and distributed to healthcare organizations operating outside of Arizona’s largest population centered counties, Maricopa and Pima. Numbers of healthcare providers were pulled from the Center for Medicare and Medicaid Services. Numbers of clinical sites and healthcare providers were converted to county-level per capita rates. Rural Urban Continuum (RUC) codes were used to designate county metropolitan status. County demographic information from the U.S. Census Bureau, income data from the US Bureau of Economic Analysis, and unemployment rates from the US Department of Labor were included. Descriptive statistics were used to summarize the results. A student’s t-test was used to evaluate differences between rural and urban counties. For Aim 2, the Arizona Department of Health Services Cancer Registry provided county level cancer incidence and death rates from 2010 through 2016. Linear regression was used to evaluate the association between rural status and breast cancer capacity and services with breast cancer incidence and mortality rates. For Aim 3, current US federal and state level policies focused on increasing the rural workforce were reviewed. Results: Out of Arizona’s 15 counties, 13 were represented. Six were urban (RUC codes 1 – 3) and seven were rural (RUC codes 4 – 7). Urban counties had a larger average population (216,773) than rural counties (49,507) (p-value = 0.01). Rural counties had more per capita clinical sites (20.4) than urban counties (8.9) (p-value = 0.02). Rural counties had more per capita cervical cancer screening sites (18.9) than urban counties (7) (p-value = 0.02) and rural counties had more per capita colorectal cancer screening sites (15.7) than urban counties (2.5) (p-value = 0.02). Urban counties had more per capita gastroenterologists (2.2) than rural counties (0) (p-value = 0.02) and urban counties had more per capita pathologists (1.0) than rural counties (0) (p-value = <0.01). Rural counties had zero medical oncologists. Per capita, rural counties with RUC codes 4 and 6 had hematology and oncology physicians (0.3, 2.5) and radiologists (2.8, 6.0) but those with RUC code 7 had zero. Although not significantly different, rural counties with RUC code 6 had three times as many per capita registered nurses (306.7) than urban counties (90.8). Rural county status was associated with a decrease in breast cancer incidence (β = -20.1, 95% CI: -37.2, 3.1). There was no association between breast cancer incidence and county per capita sites providing breast cancer screening (β = -8.8, 95% CI: -23.9, 6.9), diagnosis (β = -5.2, 95% CI: -22.2, 11.7), treatment (β = -6.5, 95% CI: -23.2, 10.2), and all three services (β = -8.0, 95% CI: -23.9, 7.9) or county per capita primary care physicians (β = 0.0, 95% CI: -0.54, 0.48), hematology oncology physicians (β = -0.9, 95% CI: -15.7, 13.8), medical oncology physicians (β = 35.2, 95% CI: -22.7, 93.0), OBGYN physicians (β = -0.5, 95% CI: -4.2, 3.2), radiologists (β = -0.2, 95% CI: -6.8, 6.4), and surgeons (β = 1.6, 95% CI: -3.1, 6.3). In the unadjusted model, rural RUC codes four (β = -24.1, 95% CI: -41.8, -6.4) and six (β = -32.6, 95% CI: -53.0, -12.2) were associated with breast cancer incidence. There was no association between breast cancer incidence and RUC code 7 (β = -1.8, 95% CI: -22.3, 18.6). In the model adjusted for race (percent of the county population that’s Hispanic) and Ethnicity (percent of the county population that’s American Indian and Alaska Native), RUC codes four (β = -19.0, 95% CI: -37.7, -0.4) and six (β = -32.6, 95% CI: -56.0, -7.9) were associated with breast cancer incidence. There was no association between breast cancer mortality and rural county status (β = -1.1, 95% CI: -7.7, 5.6), county per capita sites providing breast cancer screening (β = -0.2, 95% CI: -4.2, 3.8), diagnosis (β = 0.4, 95% CI: -3.8, 4.6), treatment (β = 0.4, 95% CI: -3.9, 4.6), all three services (0.2, -3.9, 4.3) or county per capita primary care physicians (β = 0.0, 95% CI: -0.1, 0.0), hematology oncology physicians (β = -1.6, 95% CI: -5.3, 2.1), medical oncology physicians (β = -0.9, 95% CI: -17.2, 15.3), OBGYN physicians (β = -0.6, 95% CI: -1.5, 0.3), radiologists (β = -0.7, 95% CI: -2.4, 1.0), and surgeons (β = -0.1, 95% CI: -1.4, 1.2). Conclusions: While rural counties may have more physical infrastructure, they lack specialists integral to providing cancer services. Non-physician clinical providers may be more prevalent in rural areas and represent opportunities for improving cancer care. Compared to urban counties, rural county status was associated with lower breast cancer incidence rates but not associated with breast cancer death. The number of sites delivering breast cancer services and physicians were not associated with breast cancer incidence or mortality at the county level. Other factors may contribute to rural urban differences in breast cancer incidence. Federal and state level policies have been effective in increasing the rural healthcare workforce. However, opportunities for improving rural cancer care through policies and programs exist. Improved data collection and availability from state level workforce data and the FDA mammography database can help improve cancer capacity research. Increased exposure to rural locations during residency, transformation of GME payment, and expansion of loan repayment and scholarship programs may help increase the number of specialists delivering cancer care in rural Arizona and nationally. Increased training opportunities and the scope of work expansions for non-physician clinicians and advanced practice providers may help improve the delivery of cancer prevention and treatment services in rural areas that lack specialist physicians. Future research should explore these factors as well as the association between cancer capacity and resources at a more local level since Arizona counties can be a heterogeneous unit of observation.