44. Discrete Event Simulation (DES)

This section focuses on Discrete Event Simulation (DES) models also known as Time-to-Event models, which allow you to structure models around events and continuous time rather than health states and time cycles of fixed length. Within this section, we will use DES models and Time-to-Event models somewhat interchangeably.

Discrete Event Simulation models differ significantly from Markov models in that movement through the model is driven by events whenever they occur rather than events that may or may not happen within a given cycle. TreeAge Pro uses a similar decision tree model structure for DES and Markov models, but uses specific node types and keywords for DES. It is important to consider that flow through the model is based on events and the progression of continuous time and not discrete time intervals.

TreeAge Pro supports most components of Discrete Event Simulation (DES) models; however, TreeAge Pro does not directly support resource constraints, queues and other elements less frequently used for healthcare models.

Similar to Markov models, DES models can be incorporated into a large decision tree to facilitate cost-effectiveness analysis and other mechanisms to compare treatment strategies. Unlike Markov models, DES models must be analyzed via Microsimulation, which sends individual patients through the model. Individual patient histories are then aggregated into means to generate expected values for decision analysis.

DES models must be analyzed via Microsimulation to send individual patients through the model. Cohort-level analyses are not supported.

We will use the DES tutorial example model, DES.Model.With.Health.States. This model compares two treatment strategies with a disease progression through the health states of: mild, severe and dead.