Purdue University Graduate School

File(s) under embargo





until file(s) become available


posted on 2021-12-10, 17:33 authored by Carolina Vivas ValenciaCarolina Vivas Valencia

Health care has become a data-intense industry with billions of data generated at numerous healthcare encounters on a daily basis through wearable devices, medical images, and from biomedical and biobehavioral research studies. These data are essential to driving improvements in population health to address issues related to cost, quality, and outcomes, as well as to promote personalized medicine and precision health. Even though significant progress has been made in harnessing the digitized administrative data, there is still a long way to ensure excellence in precision population health, especially on service utilization and delivery. This Ph.D. dissertation focuses on developing innovative data-driven systems analytics solutions to understand opioid use disorder and its associated harms, and to synthesize today’s evidence to improve tomorrow’s prevention and treatment outcomes on opioid use disorder.

Over the past 20 years, the opioid epidemic has been responsible for numerous cases of overdose deaths among Americans and worldwide. The economic burden of the opioid crisis to individuals and society is immense. Furthermore, the opioid overdose epidemic has not impacted communities equally, with increased barriers to care access in an already fragmented system. This thesis comprises five essays that aim to address unanswered questions within the public health and economic literature on opioid use disorder diagnosis and treatment. The first essay of the thesis aims to understand how physicians’ implicit (unconscious) bias contributes to opioid-based treatment disparities in Indiana. The second essay takes a closer look into health care disparities by examining the impact of social determinants on opioid user disorder diagnosis and treatment. The third essay studies how effectively the passage of Indiana Public Law 194 reduced the percentage of doctor shoppers for prescription opioids among Indiana Medicaid enrollees. With the use of network analysis, the fourth essay takes a closer look into the interactions between prescription-opioid doctor-shoppers and providers prescribing opioids under the influence of prescription opioid prescribing guidelines/regulations. Finally, the last essay compares the effect and cost-effectiveness of opioid use disorder treatments based on a self-developed individualized state transition model. These essays collectively provide a data-analytics framework to study population-level opioid use disorder outcomes, investigate the impact of current public approaches on the opioid overdose epidemic, and inform policymakers and the general community about the social and structural determinants affecting communities and populations vulnerable to the epidemic.


Degree Type

  • Doctor of Philosophy


  • Biomedical Engineering

Campus location

  • West Lafayette

Advisor/Supervisor/Committee Chair

Nan Kong

Additional Committee Member 2

Paul Griffin

Additional Committee Member 3

Nicole Adams

Additional Committee Member 4

Jacqueline Linnes

Additional Committee Member 5

Denny Yu