Of the 17,795 respondents to the long version of survey that included the discrimination measures, there were a total of 16,463 respondents to the general discrimination question and 12,151 respondents to the health care discrimination question. Respondents to the health care discrimination question were well distributed across the major racial/ethnic groups of interest: 20% Black, 20% Latino, 12% Asian, 11% Filipino, and 30% White, with 6% reporting another race/ethnicity. Respondents not answering the discrimination questions were more likely to report Latino, Asian, Filipino, or Other race/ethnicity, and be older, less educated, lower income, not born in the U.S., with poor health literacy and LEP.
The majority (66%) of respondents reported no general discrimination in the past year, and an overwhelming majority (95%) reported no discrimination from their health care provider in the past year (). However, after accounting for non-proportional sampling fractions (specifically, the oversampling of minorities), the proportions reporting general and health care discrimination in the overall Kaiser Diabetes Registry population were 20% and 3%. Furthermore, the reports of discrimination were significantly related to one another – that is, individuals who reported general discrimination were more likely also to report health care provider discrimination (p<.01, ).
Reports of Racial/Ethnic Discrimination, By Race/Ethnicity
Patient Reports of General Discrimination and Health care Discrimination
Whites reported discrimination substantially less frequently than any racial/ethnic minority group (9% for general discrimination and 2% for health care discrimination). Blacks reported general discrimination most frequently (52%), and Filipinos reported health care provider discrimination most frequently (8%). Among the minority group respondents, Latinos reported general discrimination least frequently (35%), and Asians reported health care discrimination the least frequently (4%). presents the unadjusted associations between the patient characteristics and reported health care discrimination. The majority of individual-level characteristics were significantly related to both discrimination outcomes in bivariate comparisons. Minority race/ethnicity, younger age, lower income, disability status, poorer health literacy, immigrant status, LEP, lower subjective social status, higher emotional instability, higher external locus of control, higher perceived stress, worse physical functioning, and depressive symptoms were significantly associated with higher reports of health care provider discrimination. However, gender and education were not significantly associated with health care discrimination in unadjusted analyses.
Unadjusted Associations with Reports of Health care Discrimination
reports the adjusted logistic regression model results for health care discrimination, presenting the series of models across the columns. When only including demographic and socioeconomic variables in the first model, all racial/ethnic minority groups were more likely than Whites to report discrimination from their health care provider. Black, Latino, Asian, Filipino, and Other race/ethnicity respondents had 3.99 (95% CI: 2.78-5.73), 3.03 (95% CI: 2.02-4.53), 2.53 (95% CI: 1.61-3.98), 4.60 (2.96-7.14), and 3.15 (95% CI: 1.80-5.51) times the likelihood (respectively) of reporting health care discrimination as Whites. In addition, younger age (OR=1.85 comparing <50 vs. ≥65; 95% CI: 1.28-2.69) and poorer health literacy (OR=1.16 for a 1-point change; 95% CI: 1.11-1.21) were associated with increased reports of discrimination. There were no significant, adjusted associations between income, education, or subjective social status with reported health care discrimination.
Series of Adjusted Logistic Regression Models Examining Reported Health care Discrimination
In the next stage, acculturation variables of immigrant status and LEP were added to the logistic regression model. Those with poor English skills were more likely than those without English limitations to report discrimination from their health care provider (OR=2.03, 95% CI: 1.42-2.91). In addition, younger age, poor health literacy, and minority race/ethnicity remained significantly associated with discrimination in this model.
Psychosocial scales of locus of control, emotional instability, and perceived stress were added next. Perceived stress (OR=1.11 for a 1-point increase, 95% CI: 1.06-1.15) and emotional instability (OR=1.18 for a 1-point increase, 95% CI: 1.00-1.39) were significant in the adjusted model. Depression, self-reported physical health, and disability status were added in the following stage. Depression was significantly associated with reporting health care provider discrimination (OR=1.67, 95% CI: 1.20-2.33), but neither the physical component score nor disability status reached significance. With the addition of these covariates, the associations of younger age and emotional instability with reporting health care discrimination became statistically insignificant, but poorer health literacy and LEP remained associated with reported health care discrimination.
Finally, the binary indicator of reporting general discrimination was added to the model. General discrimination was significantly associated with health care discrimination in this fully adjusted model (OR=2.42; 95% CI: 1.88, 3.12), and the associations of LEP (OR=1.91, 95% CI: 1.32-2.78), depression (OR=1.53, 95% CI: 1.10-2.13), and poorer health literacy (OR=1.10, 95% CI: 1.04-1.16) remained significantly related to reporting health care discrimination. Furthermore, minority race was significantly associated with reports of health care discrimination in all stages of adjustment. However, in this fully adjusted model, the odd ratios of reporting health care discrimination for Blacks, Latinos, Asians, Filipinos, and Other race/ethnicity decreased substantially to 2.86 (95% CI: 1.93-4.22), 2.17 (95% CI: 1.40-3.37), 1.95 (95% CI: 1.20-3.16), and 2.82 (95% CI: 1.72-4.62), and 2.42 (95% CI: 1.35-4.34) (respectively) compared with Whites – representing 36%, 22%, 24%, 32%, and 22% decreases in the ORs without adjustment for reports of general discrimination.