Impact of environmental chemicals, sociodemographic variables, depression, and clinical indicators of health and nutrition on self-reported health status

Public health researchers ideally integrate social, environmental, and clinical measures to identify predictors of poor health. Chemicals measured in human tissues are often evaluated in relation to intangible or rare health outcomes, or are studied one chemical at a time. Using US nationally representative data from the 2003-2004 NHANES survey, this research evaluated whether self-reported health status (fair/poor or good/very good/excellent) was a reasonable proxy for health status by assessing bivariate associations between self reported health and obesity, depression, and hypertension for adults ages 20-50. Statistically significant (p-value <0.005) relationships were observed for all but hypertension (p-value <0.2). Self reported health status was then analyzed in relation to biomonitored chemicals (mercury, lead and cadmium, toluene, benzene, and cotinine), clinical indicators (C-reactive protein and glycohemoglobin), nutritional markers (vitamin D and phytoflurene), race, gender, income, and educational attainment. In unadjusted logistic regression models, chemical levels and clinical indicators of health risk were statistically significant predictors of self-reported health. After adjusting for sociodemographic variables using multivariate logistic regression, some associations were no longer statistically significant. Family income (p-value <0.001), age (p-value <0.001), education (p-value <0.001), Body Mass Index (p-value <0.005), depression (p-value <0.001), and being a current smoker (p-value <0.05) remained statistically significant predictors of poor self-reported health. Additional models investigate the impact of combined chemical exposures on self-reported health. This research moves beyond linking single chemical exposures to single health outcomes, to provide a more comprehensive understanding of the complex relationships between cumulative environmental exposures, clinical indicators, and socioeconomic determinates. This abstract does not necessarily represent EPA policy.