Joint Associations of Race, Ethnicity, and Socioeconomic Status With Mortality in the Multiethnic Cohort Study

Abstract

Socioeconomic status may help delineate racial and ethnic inequities in mortality. To investigate the joint associations of race, ethnicity, and neighborhood and individual socioeconomic status with mortality. This prospective analysis used data from the Multiethnic Cohort Study. A population-based sample of participants recruited from California (mainly Los Angeles County) and Hawaii from 1993 to 1996 was followed up until 2013. African American, European American, Japanese American, Latino American, and Native Hawaiian men and women were included. Participants with baseline residential addresses that could not be geocoded or who were missing information on education or adjustment variables were excluded. Data analyses were conducted from January 2018 to December 2020. Neighborhood socioeconomic status (nSES) was derived using US Census block group data on education, occupation, unemployment, household income, poverty, rent, and house values. Participants self-reported their highest education attainment. Five racial and ethnic groups, 2 states of residence, 2 nSES, and 2 education categories were combined to create a joint exposure variable. Low and high nSES were defined as quintiles 1 to 3 and 4 to 5, respectively. Low and high education levels were defined as high school or less and greater than high school graduate, respectively. All-cause, cardiovascular disease (CVD), cancer, and non-CVD and noncancer deaths were ascertained through 2013 via linkage to death certificates and the US National Death Index. Multivariable Cox proportional hazards regression analyses were conducted. Among 182 912 participants (100 785 [55.1%] women and 82 127 [44.9%] men; mean [SD] age, 60.0 [8.9] years; 31 138 African American, 45 796 European American, 52 993 Japanese American, 39 844 Latino American, and 13 141 Native Hawaiian participants) with a mean (SD) follow-up of 17 (5) years, there were 63 799 total deaths, including 23 191 CVD deaths, 19 008 cancer deaths, and 21 235 non-CVD and noncancer deaths. The lowest all-cause mortality was found among 15 104 Japanese American participants in Hawaii with high nSES and high education (eg, 2870 all-cause deaths [19.0%]), and this population served as the reference group for all regression analyses. Native Hawaiian participants in Hawaii with low nSES and low education had the highest all-cause mortality HR (2.38; 95% CI, 2.21-2.57). African American and European American participants in California with low nSES and low education had the next highest all-cause mortality HRs (2.01; 95% CI, 1.91-2.11 and 1.98; 95% CI, 1.85-2.12, respectively). Latino American participants in California with low nSES had equivalent all-cause mortality HRs regardless of education level (high education: 1.57; 95% CI, 1.48-1.66; low education: 1.57; 95% CI, 1.50-1.65). Patterns for cause-specific mortality were similar to those for all-cause mortality. For example, Native Hawaiian participants in Hawaii with low nSES and low education had highest CVD mortality HR (2.92; 95% CI, 2.60-3.27) and cancer mortality HR (2.01; 95% CI, 1.77-2.29). These results suggest that joint associations of nSES and education may further delineate racial and ethnic inequities in mortality and that future investigations of racial and ethnic inequities in mortality should consider differences by measures of socioeconomic status, especially for underserved populations.