Data Visualization In Models for Healthcare Workflow Improvement

Open Access
Chen, Yining
Graduate Program:
Industrial Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
Committee Members:
  • Harriet Black Nembhard, Thesis Advisor
  • Data Visualization
  • Healthcare
  • workflow
  • Tuberculosis
  • Optimization
Workflow optimization in the healthcare system has been analyzed and discussed by many researchers in recent years. The increasing difficulties in dealing with complex data and the increasing expectations from physicians and patients for more reliable information have necessitated improvements in the health care decision making process. Data visualization (DV) is useful approach for dealing with multifaceted data and presenting information in a user friendly way. Much research has been done to develop useful DV tools that have made information more accessible to patients as well as physicians. However, very little research has been done to implement the DV techniques effectively into the different stages of a healthcare delivery workflow. In this thesis, we study the question domains that patients and physicians must address in the treatment encounter, followed by a possible solution set for each domain with the DV techniques that may be involved. The workflow is investigated and developed from the input stage to the treatment cycle stage, focusing on steps in the decision making process and their sequences in the workflow. The DV techniques are then analyzed in detail based on their characteristics, functions and advantages. For each stage, the DV display is shown with a focus on how it may help patients and the physicians better communicate. An optimization model is formulated based on the workflow to identify the DV technique for each stage so that the overall objectives of quality, efficiency, and cost are addressed. Goal programming is used to combine different criteria into one single overall criterion. In order to evaluate the workflow with DV interventions, a workflow model is built in Simio to simulate the current and proposed workflow. Analysis of the simulation results provides insight on comparisons of the performance of the workflow with DV intervention and without DV intervention. A case study is presented for tuberculosis (TB) to show how DV techniques can help TB patients and the physicians better communicate, understand and use the information to make informed decisions. Based on this study, several future research opportunities are identified.