Limits...
Ozone and childhood respiratory disease in three US cities: evaluation of effect measure modification by neighborhood socioeconomic status using a Bayesian hierarchical approach

View Article: PubMed Central - PubMed

ABSTRACT

Background: Ground-level ozone is a potent airway irritant and a determinant of respiratory morbidity. Susceptibility to the health effects of ambient ozone may be influenced by both intrinsic and extrinsic factors, such as neighborhood socioeconomic status (SES). Questions remain regarding the manner and extent that factors such as SES influence ozone-related health effects, particularly across different study areas.

Methods: Using a 2-stage modeling approach we evaluated neighborhood SES as a modifier of ozone-related pediatric respiratory morbidity in Atlanta, Dallas, & St. Louis. We acquired multi-year data on emergency department (ED) visits among 5–18 year olds with a primary diagnosis of respiratory disease in each city. Daily concentrations of 8-h maximum ambient ozone were estimated for all ZIP Code Tabulation Areas (ZCTA) in each city by fusing observed concentration data from available network monitors with simulations from an emissions-based chemical transport model. In the first stage, we used conditional logistic regression to estimate ZCTA-specific odds ratios (OR) between ozone and respiratory ED visits, controlling for temporal trends and meteorology. In the second stage, we combined ZCTA-level estimates in a Bayesian hierarchical model to assess overall associations and effect modification by neighborhood SES considering categorical and continuous SES indicators (e.g., ZCTA-specific levels of poverty). We estimated ORs and 95% posterior intervals (PI) for a 25 ppb increase in ozone.

Results: The hierarchical model combined effect estimates from 179 ZCTAs in Atlanta, 205 ZCTAs in Dallas, and 151 ZCTAs in St. Louis. The strongest overall association of ozone and pediatric respiratory disease was in Atlanta (OR = 1.08, 95% PI: 1.06, 1.11), followed by Dallas (OR = 1.04, 95% PI: 1.01, 1.07) and St. Louis (OR = 1.03, 95% PI: 0.99, 1.07). Patterns of association across levels of neighborhood SES in each city suggested stronger ORs in low compared to high SES areas, with some evidence of non-linear effect modification.

Conclusions: Results suggest that ozone is associated with pediatric respiratory morbidity in multiple US cities; neighborhood SES may modify this association in a non-linear manner. In each city, children living in low SES environments appear to be especially vulnerable given positive ORs and high underlying rates of respiratory morbidity.

Electronic supplementary material: The online version of this article (doi:10.1186/s12940-017-0244-2) contains supplementary material, which is available to authorized users.

No MeSH data available.


Related in: MedlinePlus

Annual mean ED visit rates by neighborhood SES for each ZCTA in each city. Respiratory disease ED rates are reported per 1000 children (5–18 years old) and were calculated for each ZCTA by dividing the annual total number of respiratory disease ED visits by annual estimates of the 5–18 year old population for each year in the study period. Annual ED Rates were then averaged over the study period of each city. ED visit rates for each ZCTA are represented by the “+” symbol and shown in a by percentage (%) of the adult population (≥25 years old) with less than a 12th grade education (% < 12th grade), in b by % of households living below the federal poverty line (% below poverty), and in c by the Neighborhood Deprivation Index (NDI). The solid black line represents local polynomial regression using weighted least squares to fit a line through the data. The dotted gray lines represent the 1st, 2nd, and 3rd quartile values of each SES indicator. In each panel and city, neighborhood SES decreases from left to right. Abbreviations: ED, Emergency Department; NDI, Neighborhood Deprivation Index; RDAS, respiratory disease ED visits; SES, socioeconomic status; ZCTA, ZIP Code Tabulation Area
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC5382444&req=5

Fig5: Annual mean ED visit rates by neighborhood SES for each ZCTA in each city. Respiratory disease ED rates are reported per 1000 children (5–18 years old) and were calculated for each ZCTA by dividing the annual total number of respiratory disease ED visits by annual estimates of the 5–18 year old population for each year in the study period. Annual ED Rates were then averaged over the study period of each city. ED visit rates for each ZCTA are represented by the “+” symbol and shown in a by percentage (%) of the adult population (≥25 years old) with less than a 12th grade education (% < 12th grade), in b by % of households living below the federal poverty line (% below poverty), and in c by the Neighborhood Deprivation Index (NDI). The solid black line represents local polynomial regression using weighted least squares to fit a line through the data. The dotted gray lines represent the 1st, 2nd, and 3rd quartile values of each SES indicator. In each panel and city, neighborhood SES decreases from left to right. Abbreviations: ED, Emergency Department; NDI, Neighborhood Deprivation Index; RDAS, respiratory disease ED visits; SES, socioeconomic status; ZCTA, ZIP Code Tabulation Area

Mentions: Another plausible reason for having observed weaker associations in low SES populations may be due to our use of multiplicative models and the mathematical scale of effect measures. While multiplicative models are used in the vast majority of air pollution-health studies, [3, 51] the true nature of the effect of ozone on ED visits may be additive. In our own data, we observed a marked increase in ED rates from high SES to low SES in each city and for each SES indicator (Fig. 5). Assuming additive effects, low baseline risk could explain stronger relative effects of ozone in the highest SES populations and apparent weaker relative effects in the lowest SES populations [10, 27]. However, in many analyses we observed strong, positive associations in low SES areas, which may reflect supra-additive effects of SES and ozone [27]. While there are methods for estimating additive interaction based on results of multiplicative models (e.g. the Relative Excess Risk due to Interaction (RERI) and the Synergy Index), these methods cannot be straightforwardly applied to our models, and the validity of applying these methods to models with multiple covariates and a continuous exposure is uncertain.Fig. 5


Ozone and childhood respiratory disease in three US cities: evaluation of effect measure modification by neighborhood socioeconomic status using a Bayesian hierarchical approach
Annual mean ED visit rates by neighborhood SES for each ZCTA in each city. Respiratory disease ED rates are reported per 1000 children (5–18 years old) and were calculated for each ZCTA by dividing the annual total number of respiratory disease ED visits by annual estimates of the 5–18 year old population for each year in the study period. Annual ED Rates were then averaged over the study period of each city. ED visit rates for each ZCTA are represented by the “+” symbol and shown in a by percentage (%) of the adult population (≥25 years old) with less than a 12th grade education (% < 12th grade), in b by % of households living below the federal poverty line (% below poverty), and in c by the Neighborhood Deprivation Index (NDI). The solid black line represents local polynomial regression using weighted least squares to fit a line through the data. The dotted gray lines represent the 1st, 2nd, and 3rd quartile values of each SES indicator. In each panel and city, neighborhood SES decreases from left to right. Abbreviations: ED, Emergency Department; NDI, Neighborhood Deprivation Index; RDAS, respiratory disease ED visits; SES, socioeconomic status; ZCTA, ZIP Code Tabulation Area
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC5382444&req=5

Fig5: Annual mean ED visit rates by neighborhood SES for each ZCTA in each city. Respiratory disease ED rates are reported per 1000 children (5–18 years old) and were calculated for each ZCTA by dividing the annual total number of respiratory disease ED visits by annual estimates of the 5–18 year old population for each year in the study period. Annual ED Rates were then averaged over the study period of each city. ED visit rates for each ZCTA are represented by the “+” symbol and shown in a by percentage (%) of the adult population (≥25 years old) with less than a 12th grade education (% < 12th grade), in b by % of households living below the federal poverty line (% below poverty), and in c by the Neighborhood Deprivation Index (NDI). The solid black line represents local polynomial regression using weighted least squares to fit a line through the data. The dotted gray lines represent the 1st, 2nd, and 3rd quartile values of each SES indicator. In each panel and city, neighborhood SES decreases from left to right. Abbreviations: ED, Emergency Department; NDI, Neighborhood Deprivation Index; RDAS, respiratory disease ED visits; SES, socioeconomic status; ZCTA, ZIP Code Tabulation Area
Mentions: Another plausible reason for having observed weaker associations in low SES populations may be due to our use of multiplicative models and the mathematical scale of effect measures. While multiplicative models are used in the vast majority of air pollution-health studies, [3, 51] the true nature of the effect of ozone on ED visits may be additive. In our own data, we observed a marked increase in ED rates from high SES to low SES in each city and for each SES indicator (Fig. 5). Assuming additive effects, low baseline risk could explain stronger relative effects of ozone in the highest SES populations and apparent weaker relative effects in the lowest SES populations [10, 27]. However, in many analyses we observed strong, positive associations in low SES areas, which may reflect supra-additive effects of SES and ozone [27]. While there are methods for estimating additive interaction based on results of multiplicative models (e.g. the Relative Excess Risk due to Interaction (RERI) and the Synergy Index), these methods cannot be straightforwardly applied to our models, and the validity of applying these methods to models with multiple covariates and a continuous exposure is uncertain.Fig. 5

View Article: PubMed Central - PubMed

ABSTRACT

Background: Ground-level ozone is a potent airway irritant and a determinant of respiratory morbidity. Susceptibility to the health effects of ambient ozone may be influenced by both intrinsic and extrinsic factors, such as neighborhood socioeconomic status (SES). Questions remain regarding the manner and extent that factors such as SES influence ozone-related health effects, particularly across different study areas.

Methods: Using a 2-stage modeling approach we evaluated neighborhood SES as a modifier of ozone-related pediatric respiratory morbidity in Atlanta, Dallas, &amp; St. Louis. We acquired multi-year data on emergency department (ED) visits among 5&ndash;18 year olds with a primary diagnosis of respiratory disease in each city. Daily concentrations of 8-h maximum ambient ozone were estimated for all ZIP Code Tabulation Areas (ZCTA) in each city by fusing observed concentration data from available network monitors with simulations from an emissions-based chemical transport model. In the first stage, we used conditional logistic regression to estimate ZCTA-specific odds ratios (OR) between ozone and respiratory ED visits, controlling for temporal trends and meteorology. In the second stage, we combined ZCTA-level estimates in a Bayesian hierarchical model to assess overall associations and effect modification by neighborhood SES considering categorical and continuous SES indicators (e.g., ZCTA-specific levels of poverty). We estimated ORs and 95% posterior intervals (PI) for a 25&nbsp;ppb increase in ozone.

Results: The hierarchical model combined effect estimates from 179 ZCTAs in Atlanta, 205 ZCTAs in Dallas, and 151 ZCTAs in St. Louis. The strongest overall association of ozone and pediatric respiratory disease was in Atlanta (OR&thinsp;=&thinsp;1.08, 95% PI: 1.06, 1.11), followed by Dallas (OR&thinsp;=&thinsp;1.04, 95% PI: 1.01, 1.07) and St. Louis (OR&thinsp;=&thinsp;1.03, 95% PI: 0.99, 1.07). Patterns of association across levels of neighborhood SES in each city suggested stronger ORs in low compared to high SES areas, with some evidence of non-linear effect modification.

Conclusions: Results suggest that ozone is associated with pediatric respiratory morbidity in multiple US cities; neighborhood SES may modify this association in a non-linear manner. In each city, children living in low SES environments appear to be especially vulnerable given positive ORs and high underlying rates of respiratory morbidity.

Electronic supplementary material: The online version of this article (doi:10.1186/s12940-017-0244-2) contains supplementary material, which is available to authorized users.

No MeSH data available.


Related in: MedlinePlus