Print

McLaughlin, Victoria

Mathematics and Statistics

MS

National health care research shows that more than 60% of Americans over the age of 18 suffer from at least one chronic condition (90% for those over age 65). Furthermore, estimates suggest that about nine of every ten dollars spent for medical care on American adults is spent to treat persons with chronic conditions. As a state, Vermont mirrors these national findings. Inpatient data indicate that over half of Vermont hospitalizations include a diagnosis for a chronic condition (84% for those over age 65). These hospitalizations typically last longer and cost more than hospital stays that do not include a chronic illness diagnosis. As the Vermont population continues to age, it will be critical to·manage and prevent chronic illness hospitalizations.
The most common chronic illness diagnosis in Vermont hospitalizations is for ischemic heart disease (coronary artery disease). Behavioral risk factors for ischemic heart disease, such as smoking and obesity, are well documented at the individual and community level. Emerging research suggests that certain socioeconomic factors may also playa role in predicting this disease. Studying behavioral or socioeconomic risk factors separately does not account for associations between these different types of factors. The current research seeks to determine significant risk factors for ischemic heart disease hospitalizations at a county level by combining behavioral, socioeconomic, and demographic risk factors in a single regression model.
This study uses public-access health care datasets to construct a Poisson regression model for ischemic heart disease hospitalizations in Vermont counties. When behavioral, socioeconomic, and demographic risk factors are combined in a single regression model, the results indicate that tobacco use and education status are the best predictors of hospitalizations. In particular, the model indicates that the county rate for hospitalizations involving a primary diagnosis for ischemic heart disease almost doubles with an increase of 5% in the proportion of current smokers in the county (p=0.01). These results suggest that resources applied to tobacco cessation and prevention programs at the county level may have the greatest return on investment in reducing ischemic heart disease hospitalizations.
Because of the availability of data, this study is limited to risk factor analysis at the county level, meaning that significant results pertain to county hospitalization rates as opposed to individual risk. Additionally, the data used to obtain hospitalization counts has the potential to include multiple hospitalizations for a single patient (re-hospitalization). The impact of re-hospitalizations is explored in depth, with the conclusion that their effect is minimal for the analysis of hospitalizations with a primary diagnosis for ischemic heart disease. Analysis of risk factors for hospitalization with a non-primary diagnosis for this condition is more seriously impacted by potential re-hospitalizations.