Date of Award




Document Type


Degree Name

Doctor of Public Health (DrPH)


Department of Epidemiology and Biostatistics



Content Description

1 online resource (viii, 194 pages) : color illustrations, color maps.

Dissertation/Thesis Chair

Melissa Tracy

Committee Members

Jennifer Manganello, Laura Elwyn


Behavioral Health, Emergency Department Utilization, Health Literacy, Hospitalization, Health literacy, Emergency medical services, Hospitals

Subject Categories

Public Health


Behavioral health disorders are common in the United States (US) and are associated with potentially preventable high-cost health care utilization, such as inpatient hospitalization and emergency department utilization. Health literacy (HL) may be an important modifiable contributing factor to this preventable utilization, as limited health literacy is common in the US and has been linked to increased risk of hospitalization and emergency department utilization. However, few published studies have examined the association between health literacy and behavioral health disorders, and no studies have examined either the association between health literacy and utilization or the potential effects HL interventions might have on high-cost health care utilization for behavioral health disorders.Arizona, Florida, and New York were selected to represent the U.S. because of their large population size, diverse populations, and data availability. Data from the University of North Carolina’s Health literacy mapping project, which utilized the National Center for Educational Statistics 2003 National Assessment of Adult Literacy, was used to estimate health literacy at the zip code level. Hospitalization and emergency department utilization data were obtained from the Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project statewide datasets for 2006. Because data from the UNC Model was at the census block group level and the utilization data was at the zip code level, the health literacy estimates were mapped to the zip code level based on the overlap between zip codes and census block groups. A total of 2,699 zip codes were included in the analysis from these three states. The association between health literacy and healthcare utilization was analyzed using a negative binomial regression model adjusted by sociodemographic information obtained from the US Census Bureau for 2010/2011. This model was also utilized to estimate the potential impact HL interventions aimed at increasing zip code-level HL would have on utilization for behavioral health disorders. A significant and negative association was found between health literacy and high-cost healthcare utilization associated with behavioral health disorders among the zip codes analyzed. Health literacy appears to be more protective for emergency department utilization than hospitalization. There is significant inter-state variation in health literacy distribution, high-cost health care utilization, and sociodemographic characteristics. There is also a significant interaction between state and health literacy in predicting utilization rates, which may be at least partially caused by substantial differences in health systems and access to care between states. This study showed that interventions aimed at increasing health literacy at the zip code level could significantly reduce high-cost healthcare utilization and that targeting zip codes with limited health literacy may result in the greatest reduction in high-cost utilization within zip codes. More study is needed, as this is the first examination of the association between health literacy and high-cost behavioral healthcare utilization at the community level. However, health literacy appears to be a powerful tool that might prove useful in reducing health disparities seen overall and for behavioral health populations, and this study shows there is the potential to affect change at a community level.

Included in

Public Health Commons