USING DECOMPOSITION ANALYSIS TO UNDERSTAND THE IMPACT OF SODIUM ON DISPARITIES IN BLOOD PRESSURE IN THE UNITED STATES

Purpose:

High blood pressure is a leading cause of death in the US, but is significantly more prevalent among Black adults than White adults. Dietary reasons for this disparity are under investigation, particularly the role of dietary sodium. However, there is a strong nonlinear relationship between sodium intake and blood pressure, and numerous confounding factors that differ between Blacks and Whites, complicating the ability of standard regression analyses to identify how much sodium intake affects the disparities. We sought to use the method of distributional decomposition, which takes into account both confounders and nonlinear relationships between predictors and outcomes, to understand the degree to which sodium intake and related hypertension risk factors explain the disparities.

Methods:

We compared distributions of both systolic and diastolic blood pressure among non-Hispanic Whites and Blacks in the National Health and Nutrition Examination Survey (1971-2012), comparing results from the distributional decomposition method with standard adjusted logistic regressions to identify the degree to which the Black-White blood pressure disparity can be attributed to differences in the key known hypertension risk factors: sodium intake, potassium intake, body mass index, and tobacco smoking, education, and chronic stress level. The method of distributional decomposition utilizes a nonlinear kernel smoothing density reconstruction to identify how differences between groups in continuous risk factors can explain differences across continuous outcomes measures such as blood pressure.

Results:

While logistic regression results showed significant relationships between differences in all studied risk factors and the Black-White blood pressure disparity, the distributional decomposition result revealed that certain risk factors including sodium were only relevant for explaining a portion of the disparity and only for certain years of analysis. Sodium and potassium intake significantly explained disparities at lower ends of blood pressure, particularly systolic blood pressure levels below 140mmHg (Figure 1). The other major risk factors analyzed, including body mass index, had more even effects across all levels of the blood pressure disparity.

Conclusion:

The distributional decomposition method reveals that sodium intake differences between Blacks and Whites may only explain disparities in blood pressure at lower ends of the blood pressure gradient. The decomposition approach may help identify how public health interventions affect different aspects of disparities across the full distribution of continuous outcomes, which may be obscured by traditional analytical approaches.