Abstract

Objective: This study examined the associations between long-term exposure to ambient particle components and total psychiatric hospital admission risk among the U.S. general population.  Methods: Using Health Cost and Utilization Project (HCUP) State Inpatient Databases (SIDs), we analyzed the relative risk of psychiatric hospitalization associated with increased exposure to 14 particle components in eight U.S. states over the period of 2002-2018. The particle components data was based on super-learning models. Exposure data was matched with each residential address based on ZIP code of residence and admission year. Single component models were conducted via fitting zero-inflated negative binomial regression on counts of admission per ZIP code per year for each component separately. A multi-components regression was modeled by including the four main mass components (organic carbon (OC), elemental carbon (EC), nitrate and sulfate). We also applied Nonnegative Matrix Factorization (NMF) to identify particle source factors and obtain the source-specific effect estimates. Generalized Weighted Quantile Sum (gWQS) Regression was applied to obtain an overall mixture effect.  Results: Sulfate, Fe, Pb and Zn were associated with the largest risk increases in single-component models. For each interquartile (IQR) increase in OC and sulfate, we observed a relative risk (RR) of 1.10 (95% CI, 1.07 - 1.12) and 1.14 (95% CI, 1.12 - 1.16) respectively. Significant harmful associations were observed for multiple particle sources with the metal industry source (high loadings of Pb and sulfate) being the most influential. For one quartile increase in the WQS score of the components mixture, we observed an increased RR of 1.24 (95% CI, 1.21 - 1.26) adjusting for covariates.  Conclusions: Living in areas with higher annual average levels of particle components was associated with increased risk of psychiatric hospitalization among residents in eight U.S. states. Certain components (i.e. Pb, sulfate) and sources (metal industry) were the most related.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call