Integrated Process Mining & Simulation-Optimization Framework For Healthcare Design

Farouq Halawa, Alice Gittler, Sreenath Chalil Madathila, Mohammad T. Khasawneh 

Department of Systems Science and Industrial Engineering,     

Binghamton University, New York  EwingCole, New York

INTRODUCTION

  • Simulation-optimization is an underutilized methodology to test and better inform healthcare facility programming and design1.• 17% of hospital design projects have applied simulation modeling during facility planning and design in the US2.
  • The average age of healthcare facilities in the US exceeds 30 years, and growth in new building projects is expected for both inpatient and outpatient facilities3.
  • Challenges for timely application of simulation during healthcare design projects include:
    Highly variable workflows and lack of electronic health records (EHR) data matched to locations.
  • Time required for detailed observation and data collection/processing.
  • ISE/OR methods still in nascent phase in architecture.

 

OBJECTIVE

  • Improve healthcare facility layout planning by adopting data-driven approaches to account for variability in patient pathways and volumes in the early stages of the design process.

 

Methodology

  • Three-phase framework for healthcare facility layout planning.
    • Phase 1: Probabilistic deterministic automata4 to extract significant patient pathways.
    • Phase 2: Discrete-event simulation for right-sizing and space programming.
    • Phase 3: Automated layout planning using Facility Layout Problem1 solved by Genetic Algorithm.
  • Apply framework during the programming and design of a heart and vascular clinic to validate the approach.

 

CASE STUDY

Setting: New Outpatient Heart and Vascular
Timing: Schematic Design and Spatial Program Reconciliation

 

CLIENT REQUEST

  • Determine optimal number of spaces (exam, imaging, procedural) to reach an acceptable level of utilization (60-80%) and minimize patient waiting time.
  • Optimize spatial layout to maximize efficiency of patient and staff flows.

 

RESULTS

  • Real-time decision support and collaboration with facilities, operational, and clinical leadership.
  • Data-driven approach required limited on-site timestamp documentation.
  • Process mining allowed identification of significant flows (reduced noise in the data).
  • 8 significant out of 90 variants for Heart and Vascular Clinic.
  • Discovered total cost saving of $242,250.
  • Developed optimal spatial layout informed by flow-frequency analysis and genetic algorithm.

 

CONCLUSION

  • The practice of healthcare, particularly in the clinic setting, involves highly variable patient flows, however only a few are significant and can be extracted using process mining.
  • Simulation modeling optimizes space requirements in the context of operational objectives and can identify significant cost savings.
  • Outputs of simulation and process mining can enable automated layout planning based on significant flows.
  • OR methods have great potential to ensure facility designs actively support better care quality and efficiency.
  • This research will develop further algorithms for layout automation.

 

References

  1. Vahdatzad, V., et al. (2019) ‘Improving patient timeliness of care through efficient outpatient clinic layout design using data-driven simulation and optimization Improving patient timeliness of care through efficient outpatient’, Health Systems, pp. 1–22.
  2. Health Facilities Management/ASHE 2017 Hospital Construction Survey. https://www.hfmmagazine.com/articles/2664-focus-on-efficiency.
  3. Healthcare Design Magazine. https://www.healthcaredesignmagazine.com/architecture/global-lessons-healthcare.
  4. Arnolds, I. V. and Gartner, D. (2018) ‘Improving hospital layout planning through clinical pathway mining’, Annals of Operations Research. Springer, 263(1–2), pp. 453–477.