![[ Visit Client Website ]](images/banner.png)
BACKGROUND: The 2011 and 2012 Behavioral Risk Factor Surveillance System (BRFSS) results demonstrate significant disparities in diabetes and pre-diabetes by race-ethnicity in Hawaii, with highest rates among Japanese (26.1%) and lowest among Caucasians (12.2%) and Pacific Islanders (13.6%). However, disparities in blood sugar screening confound this association; Pacific Islanders and Filipinos report lower rates of receiving a blood sugar test within the past 3 years. This trend persists among older adults in whom triennial testing for blood sugar is recommended. This analysis evaluates whether racial and ethnic disparities in the prevalence of pre-diabetes and diabetes among older adults persist after adjusting for variability in screening and other confounders.
METHODS: The analysis included Hawaii 2011-2012 BRFSS respondents over age 45. Univariate logistic regressions evaluated associations between being pre-diabetic or diabetic with sex, age, poverty status, race-ethnicity, health coverage, checkup within the past 2 years, blood sugar test within the past 3 years, and BMI. Significant predictors were input into a multivariate regression using a stepwise approach. Self-reported race-ethnicities were parsed into seven categories (Caucasian, Native Hawaiian, Japanese, Chinese, Filipino, Pacific Islander, and Other). “Other” was excluded from analysis. BMI categories derived from standard and race-specific BMI cutoffs were evaluated separately.
RESULTS: Sex and poverty level were not associated with being pre-diabetic or diabetic. In the final multivariate models, not having health coverage and not receiving a checkup within the past 2 years were not associated with pre-diabetes and diabetes, whereas age and weight categories demonstrated consistent associations. Not having a blood sugar test within the past 3 years was a source of detection bias, with substantially lower awareness of pre-diabetes or diabetes status among those who had not been screened (OR = 0.27 (CI:0.22-0.34)). Using race-specific BMI cutoffs decreased adjusted variability in pre-diabetes and diabetes prevalence by race-ethnicity.
CONCLUSIONS: Disparities in pre-diabetes and diabetes by race-ethnicity are explained in part by population differences in age and weight status. Low rates of blood sugar screening confound differences in rates of pre-diabetes and diabetes by race-ethnicity. Assigning weight categories using standard BMI cutoffs fails to identify individuals at risk among Asian sub-populations and exaggerates the true relationship between race-ethnicity and metabolic disorders. Race-specific BMI cutoffs should be used in clinical settings to identify individuals at risk for metabolic conditions. Interventions to increase blood sugar testing that emphasize the reduction of disparities in screening are needed.