Using Markov nodes in a TreeAge Pro decision tree,
it is possible to represent a variety of types of complex models,
including:
- State transition "cohort" models
– No memory; fixed-length, discrete time intervals; discrete list of states; no individuals, but cohort
can change relative size. (Learn More)
- Microsimulation models, standard (series) –
Using trackers to model individual histories, continuous "state" or time; fixed "population".
- Microsimulation models (parallel) –
Like standard microsimulation, but agents can communicate and compete; dynamic population allowed.
Discrete event simulation, in TreeAge Pro, is generally just another name for microsimulation,
given that a broad definition of discrete event simulation might be based on three characteristics of
microsimulation models:
- discrete changes in state (versus continuous flows as in differential equations), that are finite in number
- stochastic behavior (versus deterministic), variability via Monte Carlo, random numbers
- dynamic (versus static) in time
Unlike regular state transition models in TreeAge Pro, discrete event (i.e., microsimulation) models
do not need to assume a constant cycle length. Rather than using the built-in _stage (cycle) counter to
represent equal time intervals, a user-defined tracker variable keeps track of the passage of time.
In different states, different time "granularity" can be used, with the tracker incremented by the
appropriate amount for a particular health state or event.
In addition, time to next event could be sampled from 1st-order distributions (resample per trial).
This also allows for competing risks, in which the nearest sampled time to event is used to select
the event that occurs (and the time interval for that cycle/event).
In September, 2007, TreeAge Software offered a webinar on Discrete Event Simulation using TreeAge Pro.
Click here to access the materials
from the webinar.
Elsewhere in This Section
> Decision Analysis and Decision Trees
> Markov Models - State Transition Models
> Cost-Effectiveness Analysis
> Monte Carlo Simulation
Learn
how TreeAge Pro users are using decision analysis. >