Honors Project On-Campus Access Only
Using information contained in the electronic health records of asthma patients, we use an automated method to characterize the state of asthma— controlled or exacerbated—for these patients over time. We attempt to model patterns in asthma exacerbations using a semi-Markov model that allows us to consider the impact of seasonality on patterns of asthma control. Then, we identify subpopulations of asthma patients based on seasonal patterns of exacerbations. We show that clustering patients into subpopulations allows us to better model patterns in asthma exacerbations and that seasonal patterns, as well as exacerbated state durations, appear to influence the creation of these clusters. We attempt to expand modeling of asthma exacerbations to include socioeconomic and environmental factors. We assess associations between these variables and asthma exacerbations using logistic regression and evaluate the predictive value of these variables using regression trees.
Abbott, Madeline R., "Incorporating seasonality and the environment into asthma exacerbation modeling" (2018). Mathematics, Statistics, and Computer Science Honors Projects. 39.
© Copyright is owned by author of this document