Redesigning a Healthcare Delivery System: Integrated Analysis from Service Processes to System Structures

Open Access
Kang, Hyo Jung
Graduate Program:
Industrial Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
February 20, 2015
Committee Members:
  • Harriet Black Nembhard, Dissertation Advisor
  • Harriet Black Nembhard, Committee Chair
  • Paul M Griffin, Committee Member
  • Guodong Pang, Committee Member
  • Kyoungrae Jung, Committee Member
  • Colleen M Rafferty, Special Member
  • Healthcare systems engineering
  • operations research in healthcare
  • patient flow
  • emergency department performance measure
Every year, the U.S. healthcare system becomes more complex and increasingly fragmented. Inefficiency and disintegration of its structure and operations reveal serious issues in many sectors of the system. In particular, emergency departments (ED) and primary care face a growing crisis although they play significant roles in our healthcare system as a safety net for the U.S. population and as the backbone of the nation’ healthcare system. Long waiting times, increasing left without being seen rates, and frequent ambulance diversions in the ED jeopardize patient safety. The lack of care coordination between primary care and specialty care does not optimally improve outcomes of the increasing patient population with multiple chronic diseases, which leads to high mortality and excessive costs. To tackle these problems facing EDs and primary care, significant efforts have been made from streamlining operational processes to changing organizational structures to enforcing new healthcare policies. However, many of the needs that drove these attempts still remain unfulfilled. The limited success of the improvement efforts is probably because many of the approaches have focused on optimizing sub-systems in isolation. Another main reason for the gap is probably that research addressing the healthcare problems has not reached its full potential for translation into practice. To improve these challenges in the current healthcare systems, we must understand their complexity and bridge the gaps that exist within them. Therefore, the objective of this dissertation is threefold: 1) to propose innovative solutions that address the improvement needs, 2) to develop models that integrate systems engineering methodologies to analyze the solutions from a holistic perspective, and 3) to lay the groundwork for translational research that promotes the implementation of innovative solutions for an efficient and integrated healthcare system. To achieve these objectives, this dissertation investigated three different levels of the healthcare system architecture – process, system structure, and environment. First, we focused on admission processes between the emergency department (ED) and inpatient wards in a hospital to improve ED crowding problems. The objective of this study was to build a better understanding of admission process policies (APPs) in the ED and to investigate the systemic effect of APPs on patient flow. Various APPs were modeled using discrete-event simulation, and their impacts on patient flow were evaluated with respect to a set of performance measures. The simulation results of the case study at the Penn State Hershey Medical Center showed that the alternative APPs were all effective in reducing the length of stay (LOS) of admitted patients compared to the current APP. The improved flow of admitted patients affected the LOS of discharged patients and the overall LOS. This study highlighted the importance of the integrated point of view to improve patient flow by demonstrating that ED change made in the model did not yield results in a linear fashion because of the interrelated actors and processes of a system. It also showed the potential value of leveraging APPs to reduce ED crowding. Then, we extended our research scope to include organizational structures of primary care and nephrology care for patients with chronic kidney disease (CKD). The objective of this study was to develop interventions that facilitate a transformation of the current healthcare delivery system into a patient-centered medical home for CKD care and to assess potential impacts of the interventions on the system. This study applied system dynamics methodologies to build conceptual and simulation models. Then, a multi-criteria goal programming was developed to determine the required capacity of resources for a new system. Results from both qualitative and quantitative models revealed the crucial roles of primary care physicians and care coordination between care settings in improving care for CKD patients. The study results also emphasized the importance of well-balanced allocations of resources and attentions to meet multiple concurrent goals. Finally, we developed an aggregated ED performance measure that incorporates multifaceted aspects of the care system ranging from processes to environment. Using the proposed ED performance measure, this study aimed to benchmark EDs and capture the overall performance of EDs with respect to technical and scale efficiencies. This study used a two-stage data envelopment analysis (DEA) approach in which the efficiencies of EDs were estimated in the first stage and the significant exogenous factors affecting EDs’ technical efficiency were investigated in the second stage. To our best knowledge, this is the first study that examines the scale and technical efficiencies of a large number of EDs. The results of input-oriented DEA models indicated that many EDs deployed their key inputs at less than the optimal level. Further analysis of the DEA models indicated that the scale and technical efficiencies among the small, medium, and large groups are statistically different. Based on the DEA results, a multivariate logit model was constructed to investigate significant exogenous factors that impact the technical efficiency of EDs. This research is significant in that hospitals can use these models as benchmarking tools, and the findings can be a basis to redesign EDs with respect to critical hospital resources for performance improvement.