International Journal of Environmental Research and Public Health
Background: U.S. rural populations have been disproportionately affected by the syndemic of opioi... more Background: U.S. rural populations have been disproportionately affected by the syndemic of opioid-use disorder (OUD) and the associated increase in overdoses and risk of hepatitis C virus (HCV) and human immunodeficiency virus (HIV) transmission. Local health departments (LHDs) can play a critical role in the response to this syndemic. We utilized two geospatial approaches to identify areas of discordance between LHD service availability and disease burden to inform service prioritization in rural settings. Methods: We surveyed rural Illinois LHDs to assess their OUD-related services, and calculated county-level opioid overdose, HIV, and hepatitis C diagnosis rates. Bivariate choropleth maps were created to display LHD service provision relative to disease burden in rural Illinois counties. Results: Most rural LHDs provided limited OUD-related services, although many LHDs provided HIV and HCV testing. Bivariate mapping showed rural counties with limited OUD treatment and HIV servic...
IMPORTANCE An association between social and neighborhood characteristics and health outcomes has... more IMPORTANCE An association between social and neighborhood characteristics and health outcomes has been reported but remains poorly understood owing to complex multidimensional factors that vary across geographic space. OBJECTIVES To quantify social determinants of health (SDOH) as multiple dimensions across the continental United States (the 48 contiguous states and the District of Columbia) at a small-area resolution and to examine the association of SDOH with premature mortality within Chicago, Illinois. DESIGN, SETTING, AND PARTICIPANTS In this cross-sectional study, census tracts from the US Census Bureau from 2014 were used to develop multidimensional SDOH indices and a regional typology of the continental United States at a small-area level (n = 71 901 census tracts with approximately 312 million persons) using dimension reduction and clustering machine learning techniques (unsupervised algorithms used to reduce dimensions of multivariate data). The SDOH indices were used to estimate age-adjusted mortality rates in Chicago (n = 789 census tracts with approximately 7.5 million persons) with a spatial regression for the same period, while controlling for violent crime. MAIN OUTCOMES AND MEASURES Fifteen variables, measured as a 5-year mean, were selected to characterize SDOH as small-area variations for demographic characteristics of vulnerable groups, economic status, social and neighborhood characteristics, and housing and transportation availability at the census-tract level. This SDOH data matrix was reduced to 4 indices reflecting advantage, isolation, opportunity, and mixed immigrant cohesion and accessibility, which were then clustered into 7 distinct multidimensional neighborhood typologies. The association between SDOH indices and premature mortality (defined as death before age 75 years) in Chicago was measured by years of potential life lost and aggregated to a 5-year mean. Data analyses were conducted between July 1, 2018, and August 30, 2019. RESULTS Among the 71 901 census tracts examined across the continental United States, a median (interquartile range) of 27.2% (47.1%) of residents had minority status, 12.1% (7.5%) had disabilities, 22.9% (7.6%) were 18 years and younger, and 13.6% (8.1%) were 65 years and older. Among the 789 census tracts examined in Chicago, a median (interquartile range) of 80.4% (56.3%) of residents had minority status, 10.2% (8.2%) had disabilities, 23.2% (10.9%) were 18 years and younger, and 9.5% (7.1%) were 65 years and older. Four SDOH indices accounted for 71% of the variance across all census tracts in the continental United States in 2014. The SDOH neighborhood typology of extreme poverty, which is of greatest concern to health care practitioners and policy advocates, comprised only 9.6% of all census tracts across the continental United States but characterized small areas of known public health crises. An association was observed between all SDOH indices and age-adjusted premature mortality rates in Chicago (R 2 = 0.63; P < .001), even after accounting for violent crime and spatial structures. (continued) Key Points Question How do social determinants of health vary across multiple dimensions and geographic space? Findings In this cross-sectional study of 71 901 census tracts with approximately 312 million persons across the continental United States, multivariate social determinants of health measures were reduced to 4 indices reflecting advantage, isolation, opportunity, and mixed immigrant cohesion and accessibility and were clustered into 7 neighborhood typologies that included an extreme poverty group. Social determinants of health indices were associated with premature mortality rates in Chicago, Illinois. Meaning The use of multidimensional geospatial approaches to quantify social determinants of health rather than the use of a singular deprivation index may better capture the complexity and spatial heterogeneity underlying these determinants.
Empirical work in regional science has seen a growing interest in causal inference, leveraging in... more Empirical work in regional science has seen a growing interest in causal inference, leveraging insights from econometrics, statistics, and related fields. This has resulted in several conceptual as well as empirical papers. However, the role of spatial effects, such as spatial dependence (SD) and spatial heterogeneity (SH), is less well understood in this context. Such spatial effects violate the so-called stable unit treatment value assumption advanced by Rubin as part of the foundational framework for empirical treatment effect analysis. In this article, we consider the role of spatial effects more closely. We provide a brief overview of a number of attempts to extend existing econometric treatment effect evaluation methods with an accounting for spatial aspects and outline and illustrate an alternative approach. Specifically, we propose a spatially explicit counterfactual framework that leverages spatial panel econometrics to account for both SD and SH in treatment choice, treatm...
Background The purpose of this study was to examine the longitudinal association between rising v... more Background The purpose of this study was to examine the longitudinal association between rising violent crime and elevated blood pressure (BP). Methods We analyzed 217,816 BP measurements from 17,783 adults during a temporal surge in violent crime in Chicago (2014-2016). Serial observations were abstracted from the EHR at an academic medical center and paired to the City of Chicago Police Data Portal. The violent crime rate (VCR) was calculated as the number of violent crimes per 1000 population per year for each census tract. Longitudinal multilevel regression models were implemented to assess elevated BP (systolic BP≥140 mmHg or diastolic BP≥90 mmHg) as a function of the VCR, adjusting for patient characteristics, neighborhood characteristics, and time effects. Secondary dependent measures included elevated heart rate, obesity, missed outpatient appointments, all-cause hospital admissions, and cardiovascular hospital admissions. Results At baseline, the median VCR was 41.3 (IQR, 1...
This map shows areas with significantly high and significantly low prevalence of diabetes and pre... more This map shows areas with significantly high and significantly low prevalence of diabetes and prediabetes in a health center population in Chicago. Prevalence was determined by ICD (International Classification of Diseases) codes and measured hemoglobin A1 c (HbA 1c). The map highlights regional clusters and isolated areas of diabetes prevalence that could be targeted with interventions to improve health outcomes. Diagnoses determined by ICD codes are shown in colors as hot and cold spot cluster cores corresponding to "high-high" (HH) and "low-low" (LL) LISA (local indicator of spatial autocorrelation) statistics, where selected census tracts and neighboring tracts both have high rates (HH) or both have low rates (LL) of diabetes. Hot spot outliers have high diabetes rates compared with neighboring tracts ("high-low" [HL]), whereas cold spot outliers have low diabetes rates compared with neighboring tracts ("low-high" [LH]). Hot spots of prediabetes and diabetes determined by measured HbA 1c levels are also shown. LISA significance set at P < .05. Supermarket data are from Kolak et al (1). Census tract and community area boundary data are from the Chicago Data Portal (2,3). Basemap imagery is from OpenStreetMap, Leaflet, and Carto.
IMPORTANCE Little is known about the distribution of life-saving trauma resources by racial/ethni... more IMPORTANCE Little is known about the distribution of life-saving trauma resources by racial/ethnic composition in US cities, and if racial/ethnic minority populations disproportionately live in US urban trauma deserts. OBJECTIVE To examine racial/ethnic differences in geographic access to trauma care in the 3 largest US cities, considering the role of residential segregation and neighborhood poverty. DESIGN, SETTING, AND PARTICIPANTS A cross-sectional, multiple-methods study evaluated census tract data from the 2015 American Community Survey in Chicago, Illinois; Los Angeles (LA), California; and New York City (NYC), New York (N = 3932). These data were paired to geographic coordinates of all adult level I and II trauma centers within an 8.0-km buffer of each city. Between February and September 2018, small-area analyses were conducted to assess trauma desert status as a function of neighborhood racial/ethnic composition, and geospatial analyses were conducted to examine statistically significant trauma desert hot spots. MAIN OUTCOMES AND MEASURES In small-area analyses, a trauma desert was defined as travel distance greater than 8.0 km to the nearest adult level I or level II trauma center. In geospatial analyses, relative trauma deserts were identified using travel distance as a continuous measure. Census tracts were classified into (1) racial/ethnic composition categories, based on patterns of residential segregation, including white majority, black majority, Hispanic/Latino majority, and other or integrated; and (2) poverty categories, including nonpoor and poor. RESULTS Chicago, LA, and NYC contained 798, 1006, and 2128 census tracts, respectively. A large proportion comprised a black majority population in Chicago (35.1%) and NYC (21.4%), compared with LA (2.7%). In primary analyses, black majority census tracts were more likely than white majority census tracts to be located in a trauma desert in Chicago (odds ratio [OR], 8.48; 95% CI, 5.71-12.59) and LA (OR, 5.11; 95% CI, 1.50-17.39). In NYC, racial/ethnic disparities were not significant in unadjusted models, but were significant in models adjusting for poverty and race-poverty interaction effects (adjusted OR, 1.87; 95% CI, 1.27-2.74). In comparison, Hispanic/Latino majority census tracts were less likely to be located in a trauma desert in NYC (OR, 0.03; 95% CI, 0.01-0.11) and LA (OR, 0.30; 95% CI, 0.22-0.40), but slightly more likely in Chicago (OR, 2.38; 95% CI, 1.56-3.64). CONCLUSIONS AND RELEVANCE In this study, black majority census tracts were the only racial/ ethnic group that appeared to be associated with disparities in geographic access to trauma centers.
ObjectiveYoung black men who have sex with men (YBMSM) in the USA represent a subgroup that has t... more ObjectiveYoung black men who have sex with men (YBMSM) in the USA represent a subgroup that has the highest HIV incidence among the overall population. In the USA, pre-exposure prophylaxis (PrEP) is an effective prevention intervention to prevent HIV acquisition when taken regularly. Neighbourhood and network factors may relate to PrEP awareness, but have not been studied in YBMSM. This study aimed to examine the relationship of neighbourhood and network characteristics with PrEP awareness among YBMSM.MethodsWe used data collected from a sample of 618 YBMSM in Chicago (2013–2014). Home addresses were collected for participants and enumerated network members. Administrative data (eg, 2014 American Community Survey, Chicago Department of Public Health) were used to describe residence characteristics. Network member characteristics were also collected (eg, sexual partners’ sex-drug use, confidant network members who were also MSM). Multilevel analysis was performed to examine the relat...
Longitudinal analysis of supermarkets over time is essential to understanding the dynamics of foo... more Longitudinal analysis of supermarkets over time is essential to understanding the dynamics of foodscape environments for healthy living. Supermarkets for 2007, 2011, and 2014 for the City of Chicago were curated and further validated. The average distance to all supermarkets along the street network was constructed for each resident-populated census tract. These analytic results were generated with GIS software and stored as spatially enabled data files, facilitating further research and analysis. The data presented in this article are related to the research article entitled &quot;Urban foodscape trends: Disparities in healthy food access in Chicago, 2007-2014&quot; (Kolak et al., 2018).
We investigated changes in supermarket access in Chicago between 2007 and 2014, spanning The Grea... more We investigated changes in supermarket access in Chicago between 2007 and 2014, spanning The Great Recession, which we hypothesized worsened local food inequity. We mapped the average street network distance to the nearest supermarket across census tracts in 2007, 2011, and 2014, and identified spatial clusters of persistently low, high or changing access over time. Although the total number of supermarkets increased city-wide, extremely low food access areas in segregated, low income regions did not benefit. Among black and socioeconomically disadvantaged residents of Chicago, access to healthy food is persistently poor and worsened in some areas following recent economic shocks.
We analyzed community area differences in teen births in Chicago, Illinois, from 1999 to 2009. We... more We analyzed community area differences in teen births in Chicago, Illinois, from 1999 to 2009. We analyzed the association between changes in teen birth rates and concurrent measures of community area socioeconomic and demographic change. Mean annual changes in teen birth rates in 77 Chicago community areas were correlated with concurrent census-based population changes during the decade. Census measures included changes in race/ethnicity, adult high school dropouts, poverty or higher-income households, crowded housing, unemployment, English proficiency, foreign-born residents, or residents who moved in the last five years. We included non-collinear census measures with a p<0.1 bivariate association with change in teen births in a stepwise multiple linear regression model. Teen birth rates in Chicago fell faster than the overall birth rates, from 85 births per 1,000 teens in 1999 to 57 births per 1,000 teens in 2009. There were strong positive associations between increases in th...
International Journal of Environmental Research and Public Health
Background: U.S. rural populations have been disproportionately affected by the syndemic of opioi... more Background: U.S. rural populations have been disproportionately affected by the syndemic of opioid-use disorder (OUD) and the associated increase in overdoses and risk of hepatitis C virus (HCV) and human immunodeficiency virus (HIV) transmission. Local health departments (LHDs) can play a critical role in the response to this syndemic. We utilized two geospatial approaches to identify areas of discordance between LHD service availability and disease burden to inform service prioritization in rural settings. Methods: We surveyed rural Illinois LHDs to assess their OUD-related services, and calculated county-level opioid overdose, HIV, and hepatitis C diagnosis rates. Bivariate choropleth maps were created to display LHD service provision relative to disease burden in rural Illinois counties. Results: Most rural LHDs provided limited OUD-related services, although many LHDs provided HIV and HCV testing. Bivariate mapping showed rural counties with limited OUD treatment and HIV servic...
IMPORTANCE An association between social and neighborhood characteristics and health outcomes has... more IMPORTANCE An association between social and neighborhood characteristics and health outcomes has been reported but remains poorly understood owing to complex multidimensional factors that vary across geographic space. OBJECTIVES To quantify social determinants of health (SDOH) as multiple dimensions across the continental United States (the 48 contiguous states and the District of Columbia) at a small-area resolution and to examine the association of SDOH with premature mortality within Chicago, Illinois. DESIGN, SETTING, AND PARTICIPANTS In this cross-sectional study, census tracts from the US Census Bureau from 2014 were used to develop multidimensional SDOH indices and a regional typology of the continental United States at a small-area level (n = 71 901 census tracts with approximately 312 million persons) using dimension reduction and clustering machine learning techniques (unsupervised algorithms used to reduce dimensions of multivariate data). The SDOH indices were used to estimate age-adjusted mortality rates in Chicago (n = 789 census tracts with approximately 7.5 million persons) with a spatial regression for the same period, while controlling for violent crime. MAIN OUTCOMES AND MEASURES Fifteen variables, measured as a 5-year mean, were selected to characterize SDOH as small-area variations for demographic characteristics of vulnerable groups, economic status, social and neighborhood characteristics, and housing and transportation availability at the census-tract level. This SDOH data matrix was reduced to 4 indices reflecting advantage, isolation, opportunity, and mixed immigrant cohesion and accessibility, which were then clustered into 7 distinct multidimensional neighborhood typologies. The association between SDOH indices and premature mortality (defined as death before age 75 years) in Chicago was measured by years of potential life lost and aggregated to a 5-year mean. Data analyses were conducted between July 1, 2018, and August 30, 2019. RESULTS Among the 71 901 census tracts examined across the continental United States, a median (interquartile range) of 27.2% (47.1%) of residents had minority status, 12.1% (7.5%) had disabilities, 22.9% (7.6%) were 18 years and younger, and 13.6% (8.1%) were 65 years and older. Among the 789 census tracts examined in Chicago, a median (interquartile range) of 80.4% (56.3%) of residents had minority status, 10.2% (8.2%) had disabilities, 23.2% (10.9%) were 18 years and younger, and 9.5% (7.1%) were 65 years and older. Four SDOH indices accounted for 71% of the variance across all census tracts in the continental United States in 2014. The SDOH neighborhood typology of extreme poverty, which is of greatest concern to health care practitioners and policy advocates, comprised only 9.6% of all census tracts across the continental United States but characterized small areas of known public health crises. An association was observed between all SDOH indices and age-adjusted premature mortality rates in Chicago (R 2 = 0.63; P < .001), even after accounting for violent crime and spatial structures. (continued) Key Points Question How do social determinants of health vary across multiple dimensions and geographic space? Findings In this cross-sectional study of 71 901 census tracts with approximately 312 million persons across the continental United States, multivariate social determinants of health measures were reduced to 4 indices reflecting advantage, isolation, opportunity, and mixed immigrant cohesion and accessibility and were clustered into 7 neighborhood typologies that included an extreme poverty group. Social determinants of health indices were associated with premature mortality rates in Chicago, Illinois. Meaning The use of multidimensional geospatial approaches to quantify social determinants of health rather than the use of a singular deprivation index may better capture the complexity and spatial heterogeneity underlying these determinants.
Empirical work in regional science has seen a growing interest in causal inference, leveraging in... more Empirical work in regional science has seen a growing interest in causal inference, leveraging insights from econometrics, statistics, and related fields. This has resulted in several conceptual as well as empirical papers. However, the role of spatial effects, such as spatial dependence (SD) and spatial heterogeneity (SH), is less well understood in this context. Such spatial effects violate the so-called stable unit treatment value assumption advanced by Rubin as part of the foundational framework for empirical treatment effect analysis. In this article, we consider the role of spatial effects more closely. We provide a brief overview of a number of attempts to extend existing econometric treatment effect evaluation methods with an accounting for spatial aspects and outline and illustrate an alternative approach. Specifically, we propose a spatially explicit counterfactual framework that leverages spatial panel econometrics to account for both SD and SH in treatment choice, treatm...
Background The purpose of this study was to examine the longitudinal association between rising v... more Background The purpose of this study was to examine the longitudinal association between rising violent crime and elevated blood pressure (BP). Methods We analyzed 217,816 BP measurements from 17,783 adults during a temporal surge in violent crime in Chicago (2014-2016). Serial observations were abstracted from the EHR at an academic medical center and paired to the City of Chicago Police Data Portal. The violent crime rate (VCR) was calculated as the number of violent crimes per 1000 population per year for each census tract. Longitudinal multilevel regression models were implemented to assess elevated BP (systolic BP≥140 mmHg or diastolic BP≥90 mmHg) as a function of the VCR, adjusting for patient characteristics, neighborhood characteristics, and time effects. Secondary dependent measures included elevated heart rate, obesity, missed outpatient appointments, all-cause hospital admissions, and cardiovascular hospital admissions. Results At baseline, the median VCR was 41.3 (IQR, 1...
This map shows areas with significantly high and significantly low prevalence of diabetes and pre... more This map shows areas with significantly high and significantly low prevalence of diabetes and prediabetes in a health center population in Chicago. Prevalence was determined by ICD (International Classification of Diseases) codes and measured hemoglobin A1 c (HbA 1c). The map highlights regional clusters and isolated areas of diabetes prevalence that could be targeted with interventions to improve health outcomes. Diagnoses determined by ICD codes are shown in colors as hot and cold spot cluster cores corresponding to "high-high" (HH) and "low-low" (LL) LISA (local indicator of spatial autocorrelation) statistics, where selected census tracts and neighboring tracts both have high rates (HH) or both have low rates (LL) of diabetes. Hot spot outliers have high diabetes rates compared with neighboring tracts ("high-low" [HL]), whereas cold spot outliers have low diabetes rates compared with neighboring tracts ("low-high" [LH]). Hot spots of prediabetes and diabetes determined by measured HbA 1c levels are also shown. LISA significance set at P < .05. Supermarket data are from Kolak et al (1). Census tract and community area boundary data are from the Chicago Data Portal (2,3). Basemap imagery is from OpenStreetMap, Leaflet, and Carto.
IMPORTANCE Little is known about the distribution of life-saving trauma resources by racial/ethni... more IMPORTANCE Little is known about the distribution of life-saving trauma resources by racial/ethnic composition in US cities, and if racial/ethnic minority populations disproportionately live in US urban trauma deserts. OBJECTIVE To examine racial/ethnic differences in geographic access to trauma care in the 3 largest US cities, considering the role of residential segregation and neighborhood poverty. DESIGN, SETTING, AND PARTICIPANTS A cross-sectional, multiple-methods study evaluated census tract data from the 2015 American Community Survey in Chicago, Illinois; Los Angeles (LA), California; and New York City (NYC), New York (N = 3932). These data were paired to geographic coordinates of all adult level I and II trauma centers within an 8.0-km buffer of each city. Between February and September 2018, small-area analyses were conducted to assess trauma desert status as a function of neighborhood racial/ethnic composition, and geospatial analyses were conducted to examine statistically significant trauma desert hot spots. MAIN OUTCOMES AND MEASURES In small-area analyses, a trauma desert was defined as travel distance greater than 8.0 km to the nearest adult level I or level II trauma center. In geospatial analyses, relative trauma deserts were identified using travel distance as a continuous measure. Census tracts were classified into (1) racial/ethnic composition categories, based on patterns of residential segregation, including white majority, black majority, Hispanic/Latino majority, and other or integrated; and (2) poverty categories, including nonpoor and poor. RESULTS Chicago, LA, and NYC contained 798, 1006, and 2128 census tracts, respectively. A large proportion comprised a black majority population in Chicago (35.1%) and NYC (21.4%), compared with LA (2.7%). In primary analyses, black majority census tracts were more likely than white majority census tracts to be located in a trauma desert in Chicago (odds ratio [OR], 8.48; 95% CI, 5.71-12.59) and LA (OR, 5.11; 95% CI, 1.50-17.39). In NYC, racial/ethnic disparities were not significant in unadjusted models, but were significant in models adjusting for poverty and race-poverty interaction effects (adjusted OR, 1.87; 95% CI, 1.27-2.74). In comparison, Hispanic/Latino majority census tracts were less likely to be located in a trauma desert in NYC (OR, 0.03; 95% CI, 0.01-0.11) and LA (OR, 0.30; 95% CI, 0.22-0.40), but slightly more likely in Chicago (OR, 2.38; 95% CI, 1.56-3.64). CONCLUSIONS AND RELEVANCE In this study, black majority census tracts were the only racial/ ethnic group that appeared to be associated with disparities in geographic access to trauma centers.
ObjectiveYoung black men who have sex with men (YBMSM) in the USA represent a subgroup that has t... more ObjectiveYoung black men who have sex with men (YBMSM) in the USA represent a subgroup that has the highest HIV incidence among the overall population. In the USA, pre-exposure prophylaxis (PrEP) is an effective prevention intervention to prevent HIV acquisition when taken regularly. Neighbourhood and network factors may relate to PrEP awareness, but have not been studied in YBMSM. This study aimed to examine the relationship of neighbourhood and network characteristics with PrEP awareness among YBMSM.MethodsWe used data collected from a sample of 618 YBMSM in Chicago (2013–2014). Home addresses were collected for participants and enumerated network members. Administrative data (eg, 2014 American Community Survey, Chicago Department of Public Health) were used to describe residence characteristics. Network member characteristics were also collected (eg, sexual partners’ sex-drug use, confidant network members who were also MSM). Multilevel analysis was performed to examine the relat...
Longitudinal analysis of supermarkets over time is essential to understanding the dynamics of foo... more Longitudinal analysis of supermarkets over time is essential to understanding the dynamics of foodscape environments for healthy living. Supermarkets for 2007, 2011, and 2014 for the City of Chicago were curated and further validated. The average distance to all supermarkets along the street network was constructed for each resident-populated census tract. These analytic results were generated with GIS software and stored as spatially enabled data files, facilitating further research and analysis. The data presented in this article are related to the research article entitled &quot;Urban foodscape trends: Disparities in healthy food access in Chicago, 2007-2014&quot; (Kolak et al., 2018).
We investigated changes in supermarket access in Chicago between 2007 and 2014, spanning The Grea... more We investigated changes in supermarket access in Chicago between 2007 and 2014, spanning The Great Recession, which we hypothesized worsened local food inequity. We mapped the average street network distance to the nearest supermarket across census tracts in 2007, 2011, and 2014, and identified spatial clusters of persistently low, high or changing access over time. Although the total number of supermarkets increased city-wide, extremely low food access areas in segregated, low income regions did not benefit. Among black and socioeconomically disadvantaged residents of Chicago, access to healthy food is persistently poor and worsened in some areas following recent economic shocks.
We analyzed community area differences in teen births in Chicago, Illinois, from 1999 to 2009. We... more We analyzed community area differences in teen births in Chicago, Illinois, from 1999 to 2009. We analyzed the association between changes in teen birth rates and concurrent measures of community area socioeconomic and demographic change. Mean annual changes in teen birth rates in 77 Chicago community areas were correlated with concurrent census-based population changes during the decade. Census measures included changes in race/ethnicity, adult high school dropouts, poverty or higher-income households, crowded housing, unemployment, English proficiency, foreign-born residents, or residents who moved in the last five years. We included non-collinear census measures with a p<0.1 bivariate association with change in teen births in a stepwise multiple linear regression model. Teen birth rates in Chicago fell faster than the overall birth rates, from 85 births per 1,000 teens in 1999 to 57 births per 1,000 teens in 2009. There were strong positive associations between increases in th...
Uploads
Papers by Marynia Kolak