Inequalities in multiple health outcomes by education, sex, and race in 93 US counties: why we should measure them all

Abstract

Introduction: Regular reporting of health inequalities is essential to monitoring progress of efforts to reduce health inequalities. While reporting of population health became increasingly common, reporting of a subpopulation group breakdown of each indicator of the health of the population is rarely a standard practice. This study reports education-, sex-, and race-related inequalities in four health outcomes in each of the selected 93 counties in the United States in a systematic and comparable manner.

Methods: This study is a cross-sectional analysis of large, publicly available data, 2008, 2009, and 2010 Behavioral Risk Factor Surveillance System (BRFSS) Selected Metropolitan/Micropolitan Area Risk Trends (SMART) and 2008, 2009, and 2010 United States Birth Records from the National Vital Statistics System. The study population is American adults older than 25 years of age residing in the selected 93 counties, representing about 30% of the US population, roughly equally covering all geographic regions of the country. Main outcome measures are: (1) Attribute (group characteristic)-specific inequality: education-, sex-, or race-specific inequality in each of the four health outcomes (poor or fair health, poor physical health days, poor mental health days, and low birthweight) in each county; (2) Overall inequality: the average of these three attribute-specific inequalities for each health outcome in each county; and (3) Summary inequality in total morbidity: the weighted average of the overall inequalities across the four health outcomes in each county.

Results: The range of inequality across the counties differed considerably by health outcome; inequality in poor or fair health had the widest range and the highest median among inequalities in all health outcomes. In more than 70% of the counties, education-specific inequality was the largest in all health outcomes except for low birthweight.

Conclusions: It is feasible to extend population health reporting to include reporting of a subpopulation group breakdown of each indicator of the health of the population at a small jurisdictional level using publicly available data. No single group characteristic or health outcome represents the whole picture of health inequalities in a population. Examining multiple group characteristics and outcomes in a comparable manner is essential in reporting health inequalities.

Figure 1

Inequalities measured in each county. For each county, we measured several inequalities. Attribute (group characteristic)-specific inequality is education-, sex-, or race-specific inequality in each of the four health outcomes (poor or fair health, poor physical health days, poor mental health days, and low birthweight). Overall inequality is the average of these three attribute-specific inequalities for each health outcome. Summary inequality in total morbidity is the weighted average of the overall inequalities across the four health outcomes: 20% each for poor or fair health, poor physical health days, poor mental health days, and 40% for low birthweight.

Figure 2

The minimum, 25^th percentile, median, 75^th percentile, and maximum of summary inequality in total morbidity and overall inequality in each of the four health outcomes across 93 counties. Data sources: A pooled 2008, 2009, and 2010 Behavioral Risk Factor Surveillance System (BRFSS) Selected Metropolitan/Micropolitan Area Risk Trends (SMART) and a pooled 2008, 2009, and 2010 United States Birth Records from the National Vital Statistics System (NVSS). Overall inequality in each of the four health outcomes is the average of these three attribute-specific (i.e., education-, sex, or race-specific) inequalities for each health outcome. Summary inequality in total morbidity is the weighted average of the overall inequalities across the four health outcomes: 20% each for poor or fair health, poor physical health days, poor mental health days, and 40% for low birthweight.

Figure 3

The minimum, 25^th percentile, median, 75^th percentile, and maximum of attribute contributions (%) for overall inequality in each of four health outcomes in 93 counties. Data sources: A pooled 2008, 2009, and 2010 Behavioral Risk Factor Surveillance System (BRFSS) Selected Metropolitan/Micropolitan Area Risk Trends (SMART) and a pooled 2008, 2009, and 2010 United States Birth Records from the National Vital Statistics System (NVSS).