Author

Chao Wang

Date of Award

2022

Document Type

Dissertation

Degree Name

PhD in Business

Department

Department of Mathematical Sciences: Business Analytics

First Advisor

Mingfei Li

Second Advisor

Dhaval Dave

Third Advisor

Dennis Lin

Abstract

Healthcare systems globally face multiple challenges in the face of population growth and changes in disease pathology. With regard to the rising demand of the healthcare and the global threats of the pandemic, the medical datasets can be trained further to develop preventive methods. Meanwhile, policy reforms of health systems could be a critical aspect to deal with the public crisis and concerns. However, two basic problems must be addressed first: identification of key factors on a priority basis and evaluation of changes.

Thus, the paper presents a series of trials on the application of data analytics to health-related problems, including an investigation into the significance of associations of factors in relation to certain symptoms, a demonstration of evaluating health policy interventions via a series of strategic models, and an innovation designed exploration to improve existing models into AI applicability. These perspectives are addressed in three distinct projects.

For the first project, we apply the Random Forest method to study a number of factors pertinent to children’s sleep quality based on data from the Children’s Burns Outcome Questionnaire (CBOQ). For this study, we focus on babies and children from birth to 4 years old, and rank the importance of the leading factors in relations to the quality of the sleep during the recovery period.

In the second project, the impacts of the California 1115 Social Security Waiver of 2005 on uncompensated care and Medi-Cal costs from the designated public hospitals (DPHs) that participated in the waiver, are reconsidered. Through difference-in-difference models, we assess the vii impact of this state law on the inpatient care utilization. Based on the project results, our findings can provide suggestive information for a future path of instituting health reforms.

In the third project, we focus on improving existing methods based on Bayesian theory. Following Lee’s work in 1977, we discuss the practical application of the method, which we explore with both normal prior and exponential prior. Specifically, we extend the previous research idea and develop a new Bayesian method for real-time data applications, thereby contributing to a movement towards effective use of artificial intelligence.

Download

Share

COinS