Reflecting Parameter Uncertainty in Addition to Variability in Constrained Healthcare Resource Discrete Event Simulations: Worth Going the Extra Mile or a Road to Nowhere?

Hazel Y Squires, Dina Jankovic, Laura Bojke

Research output: Contribution to journalArticlepeer-review


Probabilistic sensitivity analysis (PSA) has been shown to reduce bias in outcomes of health economic models. However, only 1 existing study has been identified that incorporates PSA within a resource-constrained discrete event simulation (DES) model. This article aims to assess whether it is feasible and appropriate to use PSA to characterize parameter uncertainty in DES models that are primarily constructed to explore the impact of constrained resources.
PSA is incorporated into a new case study of an Emergency Department DES. Structured expert elicitation is used to derive the variability and uncertainty input distributions associated with length of time taken to complete key activities within the Emergency Department. Potential challenges of implementation and analysis are explored.
The results of a trial of the model, which used the best estimates of the elicited means and variability around the time taken to complete activities, provided a reasonable fit to the data for length of time within the Emergency Department. However, there was substantial and skewed uncertainty around the activity times estimated from the elicitation exercise. This led to patients taking almost 3 weeks to leave the Emergency Department in some PSA runs, which would not occur in practice.
Structured expert elicitation can be used to derive plausible estimates of activity times and their variability, but experts’ uncertainty can be substantial. For parameters that have an impact on interactions within a resource-constrained simulation model, PSA can lead to implausible model outputs; hence, other methods may be needed.
Original languageEnglish
Pages (from-to)1738-1743
Number of pages6
JournalValue in Health
Issue number12
Early online date21 Sept 2023
Publication statusPublished - 1 Dec 2023

Bibliographical note

© 2023, International Society for Pharmacoeconomics and Outcomes Research, Inc.


  • Health Economics
  • Decision Modelling
  • Resource constraints
  • Uncertainty
  • Discrete event simulation
  • Emergency Department (ED); Accident and Emergency (A&E); National Health Service (NHS); Waiting Time; Length of Stay

Cite this