Authors: Kate Lester*, University of North Texas
Topics: Medical and Health Geography, Ethnicity and Race, Spatial Analysis & Modeling
Keywords: Suicide, Race, Ethnicity, Mental Health, United States, Regression
Session Type: Paper
Presentation File: No File Uploaded
Suicide is a top ten cause of death in the United States, yet there has been little progress explaining its spatial variation. I argue that the inability to find robust correlations with suicide is due to decades of ignoring the central role of race and ethnicity. Suicide rates vary widely among racial/ethnic groups (from 6.7 per 100,000 among African Americans to 18.9 per 100,000 among non-Hispanic whites). When these differences are ignored, race/ethnicity becomes a major confounding variable, warping spatial patterns of distress and resulting in misleading correlations. This paper introduces one possible method to remedy this issue. Based on common age-adjustment methods, I will calculate both crude suicide rates and indirectly race-adjusted suicide rates at the county level (CDC data, 1999-2017). Because of small numbers in the American Indian/Alaska Native and Asian/Pacific-Islander categories, a few different stratification methods will be tested using statistical simulation. Then, using the most appropriate adjustment scheme, I will calculate indirectly adjusted suicide rates and standardized mortality ratios. A regression analysis will be performed on the race-adjusted data set and the crude rates using the same independent variables. Models will be compared for both statistical and theoretical fit. Results show that race-adjusting suicide data reduces the impact of rurality and poverty, two variables with a strong relationship to race/ethnicity. Ultimately, race-adjusted rates result in stronger effect sizes and more theoretically robust conclusions, when compared to crude or age-adjusted rates.