Thesis Title: Economic Decision Analysis for Healthcare Service: Theory and Practice

Advisors: Dr. Eva Lee

Committee: Dr. Shijie Deng, Dr. Dave Goldsman, Dr. Nikhil Chanani (School of Medicine, Emory University and Children’s Healthcare of Atlanta), Dr. Adam Skelton (U.S. Centers for Disease Control and Prevention)

Date and Time: Friday, August 19, 2016, 10:30 AM.

Location: Groseclose 402 (Advisory Boardroom)

Abstract:

Health care service system is complex both in its complicated delivery design and in scientific and regulatory requirement. It is high cost, high risk, dynamic, and unpredictable.  On one hand, healthcare service providers are challenged by high R&D cost, inefficient process management, variability in quality and necessity of changes to avoid the unanticipated consequences of (new) regulations. On the user end, they are overwhelmed by the complex service pricing and government regulations, as they often make decisions with asymmetric information with little understanding of what’s best and available in the market. To help guide providers and users and to advance healthcare service research, we carry out theory and practice studies to better understand current issues and provide reasonable decision making framework.

We first develop theoretical framework for a two-sided market framework to model competition between a Preferred Provider Organization (PPO) and a Health Maintenance Organization (HMO). Both health plans compete to attract policyholders and providers. Our game-theoretical framework examines the consequences of this risk segmentation on providers and the network effect on policyholders based on market information and network size. The outcome of competition mainly depends on two effects: a market share effect and an adverse selection effect, captured by policyholders’ surplus expectations on both policies and copayments. If the adverse selection effect is strong enough, the HMO plan takes advantage on competition. On the contrary, if the market share effect dominates, the PPO profit is higher in spite of the unfavorable risk segmentation and higher premium. Next, we expand our analysis to investigate proverse selection and network externality and its effect on consumers and providers in a generalized transaction market. Our framework establishes consumer’s utility function and provider’s profit function in terms of network indirect externality. We test the responses of both sides using an agent-based model; and establish the empirical design to enhance a “balanced” level of consumer’s utility and provider’s profit. We highlight a decision support framework for the consumers, the providers, along with the market regulators that enables an optimal social welfare.

In the second topic, we aim to analyze patient care process variation to establish new clinical guidelines for substantial process and quality improvement. In the post-operative care study of congenital heart surgery, we identify and summarize numerous practice variations within the post-operative care process. Specifically, we pinpoint various process/decision points along the post-op care continuum in which difference in care coordination, management, resources and practice may lead to significant impact and reasons for the difference in the length of stay among the five clinical sites. Using clustering and machine learning approaches, we rank the significance of these factors in predicting and influencing the length of stay of the patients. We compare current process to improved one using simulation model to demonstrate evidence in LOS improvement. Next, we generalize the study to establish and quantify the safety and efficacy of large dose needle-based epidural technique in obstetric anesthesia. The time and dose required to achieve the desired sensory level is quantified. We establish a safe and quickly effective epidural dose that can be administered through the epidural needle prior to the insertion of the epidural catheter. Based upon these clinical findings, safe dosage parameters for injections as large as 20 milliliters through the epidural needle are quantified. Physician preference for medication and dosing levels are contrasted. Understanding the causes and effects of such variations can help providers and healthcare organizations in avoiding practices that negatively impact outcomes. These findings facilitate the establishment of new epidural practice guidelines and delivery improvement.

In the last topic, we describe the first in-silico drug design system model to accelerate drug discovery. Our model spans preclinical research, the IND review, clinical research, and the NDA review. We identify the global process for any drug discovery pathway with timeline along the entire system process. Next, we identify bottlenecks, and perform system optimization that offers a holistic view of discovery pathways. The integration of bottlenecks into possible candidate tasks which can be conducted simultaneously highlights critical paths for the accelerated development process. We define the critical paths as parallel model for the new drug development. Our generalized parallel model allows not only rapid development but also processes that minimize risk, cost, and time.