Geography matters: state-level variation in children's oral health care access and oral health status.

OBJECTIVES: To ascertain differences across states in children's oral health care access and oral health status and the factors that contribute to those differences. STUDY DESIGN: Observational study using cross-sectional surveys. METHODS: Using the 2007 National Survey of Children's Health, we examined state variation in parents' report of children's oral health care access (absence of a preventive dental visit) and oral health status. We assessed the unadjusted prevalences of these outcomes, then adjusted with child-, family-, and neighbourhood-level variables using logistic regression; these results are presented directly and graphically. Using multilevel analysis, we then calculated the degree to which child-, family-, and community-level variables explained state variation. Finally, we quantified the influence of state-level variables on state variation. RESULTS: Unadjusted rates of no preventive dental care ranged 9.0-26.8% (mean 17.5%), with little impact of adjusting (10.3-26.7%). Almost 9% of the population had fair/poor oral health; unadjusted range 4.1-14.5%. Adjusting analyses affected fair/poor oral health more than access (5.7-10.7%). Child, family and community factors explained ∼» of the state variation in no preventive visit and ∼½ of fair/poor oral health. State-level factors further contributed to explaining up to a third of residual state variation. CONCLUSION: Geography matters: where a child lives has a large impact on his or her access to oral health care and oral health status, even after adjusting for child, family, community, and state variables. As state-level variation persists, other factors and richer data are needed to clarify the variation and drive changes for more egalitarian and overall improved oral health.

Levels of analysis for the study of environmental health disparities.

Reducing racial/ethnic and socioeconomic environmental health disparities requires a comprehensive multilevel conceptual and quantitative approach that recognizes the various levels through which environmental health disparities are produced and perpetuated. We propose a conceptual framework that incorporates the micro level, contained within the local level, which in turn is contained within the macro level. We discuss the utility of multilevel techniques to examine environmental level (both physical and social) and individual-level factors to appropriately quantify and improve our understanding of environmental health disparities. We discuss the reasoning and the methodological approach behind multilevel modeling, including differentiating between individual and contextual influences on individual outcomes. Next we address the questions and principles that guide the choice of levels or geographic units in multilevel studies. Finally, we address the ways in which different data sources can be combined to produce suitable data for multilevel analyses. We provide some examples of how such data sources can be linked to create multilevel data structures, and offer suggestions to facilitate the integration of multilevel techniques in environmental health disparities research and monitoring.

Choosing area based socioeconomic measures to monitor social inequalities in low birth weight and childhood lead poisoning: The Public Health Disparities Geocoding Project (US).

STUDY OBJECTIVES: : To determine which area based socioeconomic measures can meaningfully be used, at which level of geography, to monitor socioeconomic inequalities in childhood health in the US. DESIGN: Cross sectional analysis of birth certificate and childhood lead poisoning registry data, geocoded and linked to diverse area based socioeconomic measures that were generated at three geographical levels: census tract, block group, and ZIP code. SETTING: Two US states: Massachusetts (1990 population=6,016,425) and Rhode Island (1990 population=1,003,464). PARTICIPANTS: All births born to mothers ages 15 to 55 years old who were residents of either Massachusetts (1989-1991; n=267,311) or Rhode Island (1987-1993; n=96 138), and all children ages 1 to 5 years residing in Rhode Island who were screened for lead levels between 1994 and 1996 (n=62,514 children, restricted to first test during the study period). MAIN RESULTS: Analyses of both the birth weight and lead data indicated that: (a) block group and tract socioeconomic measures performed similarly within and across both states, while ZIP code level measures tended to detect smaller effects; (b) measures pertaining to economic poverty detected stronger gradients than measures of education, occupation, and wealth; (c) results were similar for categories generated by quintiles and by a priori categorical cut off points; and (d) the area based socioeconomic measures yielded estimates of effect equal to or augmenting those detected, respectively, by individual level educational data for birth outcomes and by the area based housing measure recommended by the US government for monitoring childhood lead poisoning. CONCLUSIONS: Census tract or block group area based socioeconomic measures of economic deprivation could be meaningfully used in conjunction with US public health surveillance systems to enable or enhance monitoring of social inequalities in health in the United States.

Using aggregate geographic data to proxy individual socioeconomic status: does size matter?

OBJECTIVES: This study assessed whether aggregate-level measures of socioeconomic status (SES) are less biased as proxies for individual-level measures if the unit of geographic aggregation is small in size and population. METHODS: National Health Interview Survey and census data were used to replicate analyses that identified the degree to which aggregate proxies of individual SES bias interpretations of the effects of SES on health. RESULTS: Ordinary least squares regressions on self-perceived health showed that the coefficients for income and education measured at the tract and block group levels were larger than those at the individual level but smaller than those estimated by Geronimus et al. at the zip code level. CONCLUSIONS: Researchers should be cautious about use of proxy measurement of individual SES even if proxies are calculated from small geographic units.

The effect of income inequality on the health of selected US demographic groups.

OBJECTIVES: This study assessed whether documented effects of income inequality on health are consistent across demographic subgroups of the US population. METHODS: Data from the National Health Interview Survey on White and Black non-Hispanics were used. Logistic regression models were estimated with SUDAAN software. Perceived health was the outcome variable. RESULTS: The results of the multivariate analysis, in which individual family income and county-level poverty rates were included, were not consistent with existing research. In the presence of covariates, the conditional effects of inequality were restricted to Whites aged 18-44 years in the 2 highest income inequality quartiles and middle-aged Whites in counties with the highest level of income inequality. The health of Blacks of all ages, elderly Whites, and middle-aged Whites outside of the areas of highest inequality was unaffected when controls for individual characteristics and county-level poverty were in place. CONCLUSIONS: For the United States, the independent and direct contribution of income inequality to the determination of self-perceived health net of individual income and county income levels is restricted to certain demographic groups.

Aggregation and the measurement of income inequality: effects on morbidity.

This is a cross-sectional study using records from the National Health Interview Survey linked to Census geography. The sample is restricted to white males ages 25-64 in the United States from three years (1989-1991) of the National Health Interview Survey. Perceived health is used to measure morbidity. Individual covariates include income-to-needs ratio, education and occupation. Contextual level measures of income inequality, median household income and percent in poverty are constructed at the US census county and tract level. The association between inequality and morbidity is examined using logistic regression models. Income inequality is found to exert an independent adverse effect on self-rated health at the county level, controlling for individual socioeconomic status and median income or percent poverty in the county. This corresponding effect at the tract level is reduced. Median income or percent poverty and individual socioeconomic status are the dominant correlates of perceived health status at the tract level. These results suggest that the level of geographic aggregation influences the pathways through which income inequality is actualized into an individuals' morbidity risk. At higher levels of aggregation there are independent effects of income inequality, while at lower levels of aggregation, income inequality is mediated by the neighborhood consequences of income inequality and individual processes.