Austin Parker and Dr. Ruth Kerry, Geography
Several medical geography studies have identified Utah as an outlier in terms of various illnesses with far lower rates than other states in the country. Gooaverts (2009) noticed small clusters of low cervical cancer rates in Utah compared to other states, although he wasn’t specifically studying the heavy LDS influence on Utah’s population. Lower cervical cancer rates in Utah could be attributed to the LDS “Law of Chastity”, which discourages sexual relationships outside of marriage as up to 99% of cases of cervical cancer have been linked to the human papilloma virus, a sexually transmitted disease (http://www.cancerrecovery.org.uk/cervical-cancer). A goal of this research is to show that the LDS “Law of Chastity” is an important principle that has significant health benefits through its adherence, and that avoiding behavior like sexual relationships outside of marriage will be beneficial to an individual’s health, regardless of their religious beliefs.
Because there is a heavy LDS population in the western United States (Figure 1), most of the analysis was performed for states west of Texas. LDS membership data for the year 2000 was gathered from the Association of Religious Data Archives (thearda.com) and age-adjusted cervical cancer mortality rates came for 1990-2004 from http://ratecalc.cancer.gov. Because the cancer rate data are at the county level, Poisson Kriging was used to account for the small numbers problem that can result in unreliable rates in sparsely populated areas.
Four different types of analysis were performed: a Mann Whitney U test to compare cancer mortality rates among populations with high (>5%) and low (<5%) concentrations of LDS members; a correlation and multiple linear regression analysis to show both the relationship between cervical cancer mortality and LDS membership as well as the other significant variables that might influence cervical cancer mortality rates; a bivariate Local Moran’s I to show if there are any statistically significant clusters of low mortality rates where LDS populations are high; and a geographically weighted regression to show the relationship between cervical cancer mortality and LDS membership varies spatially. Figure 2 is a split histogram showing the frequency of the range of cervical cancer mortality rates based for counties with high (>5%) and low LDS (<5%) populations.
The mean cervical cancer mortality rate for the high LDS group, as indicated by the dashed line, is lower than the Low LDS group, as indicated by the solid line. The Mann-Whitney U test confirmed this, showing that there are significantly lower cervical cancer mortailty rates in high LDS areas at p=0.05.
As predicted, a correlation analysis showed that there is a significant negative correlation between cervical cancer mortality rates and LDS membership rates. As LDS membership increaeses, cervical cancer mortality tends to decrease. A multiple linear regression (MLR) was then performed to determine variables that may influence cervical cancer mortality, as shown in Table 1 with their corresponding P-values. LDS membership was a significant variable for both western counties and all counties in the country. There are, however, some variables that are more significant, such as the percent Hispanic population and fertility rate of 18-19 year-olds in western counties.
A Bivariate Local Moran’s I was used to find significant clusters of Poisson Kriged cervical cancer mortality and LDS membership. Figure 3, shows that there are significant clusters of low cervical cancer mortality rates and high LDS membership rates (pale blue counties) in what is commonly considered the LDS heartland in the west. The counties that have significant clusters of high rates and high LDS population also have a large hispanic population suggesting greater mortality due to lack of screening/access to healthcare in this generally poorer socio-economic group. A geographically weighted regression (Figure 4) supports this pattern by showing that there are relatively strong negative relationships between cervical cancer mortality and LDS membership in areas where there are high LDS populations. Hoever, there are some counties in eastern Utah/southwestern Wyoming area that show positive relationships and the reasons for this need to be examined further.
The results of all statistical and spatial statistical tests used here confirmed our original hypothesis that the high rates of LDS membership in some counties have a significant influence on the geography of cervical cancer in the western United States. The LDS population provide a unique, geographically concentrated group to investigate the effects of abstinence before marriage and fidelity within marriage compared to “safer sex” policies which are advocated/practised in the rest of the United States.