TreeAge Pro Articles

Please use the following text to cite TreeAge Pro within a publication.

  • TreeAge Pro 2021, R1. TreeAge Software, Williamstown, MA; software available at

Note that example citation above refers specifically to TreeAge Pro 2021, R1.
You should reference the software version used in your project.


The following sites include publications citing TreeAge Pro:

Selecting appropriate cycle length of the Markov model involves a trade off between the accuracy of the results and simulation run time.  The attached paper draws on an analogy with sampling analog signals, which is grounded in the work of Bell Labs engineers Harry Nyquist and Claude Shannon in 1930s.  Their work found practical applications for establishing sampling frequencies for first Music CD standard.  Our conjecture is that similar approach applies to selection of Markov cycle length.  The paper and attached TreeAge Pro models demonstrate how to build Markov models so that the cycle length is parametrized making it easy to perform comparisons with varying cycle lengths, before deciding on the appropriate cycle length to be used.

Click here to access the paper.

Click here to access the example models.

TreeAge Software and University of Sheffield collaborated on a poster presentation studying the relative bias between Markov models and Discrete Event Simulation models. The poster materials were presented at ISPOR Dublin in November, 2013.

  • Click here to download the poster.
  • Click here to download the models implemented with TreeAge Pro 2013.
  • Click here to download a sample model implemented with new DES and Time Nodes in TreeAge Pro 2014.
NICE released the following document.
The document includes guides for simulation models using TreeAge Pro – both Markov and DES models.

In 2014 TreeAge Software Inc. undertook a comprehensive study of key statistical distribution implementations in TreeAge Pro software.

In collaboration with Department of Mathematics & Statistics of Williams College, Williamstown, MA the random number generator and following distributions: exponential, Weibull, normal, gamma, Rayleigh, Erlang, Bernoulli, binomial, Poisson and discrete uniform were subjected to various statistical tests.

Click here to see the summarized results of the analysis.

For raw test results click here.

TreeAge Software Inc. in collaboration with Department of Mathematics & Statistics of Williams College, Williamstown, MA created a reference document describing corresponding parametrization of selected distributions between TreeAge Pro, STATA, SAS and R.

Please click here to download the PDF document.

For users of R survival analysis packages SURVREG and FLEXSURVREG please click here for specific parameterization examples.

TreeAge Pro models can be calibrated using Scientific Python (SciPy) libraries. An example of a calibration of a Markov 3 state progression model with Weibull transition hazards using Nelder-Mead optimization algorithm can be found by following this link. The compressed archive includes instructions for installing Python, as well as overview of the calibration process, TreeAge Pro model and working Python script. There is also a video showing the sample calibration session.

A research poster detailing results obtained by using different optimization algorithms was presented at ISPOR conference in Tokyo on September 11, 2018. The poster is available by following this link.

TreeAge Pro Partitioned Survival models (PartSA) can be converted to Markov model using built in a built in function. The resulting Markov models can be further explored and calibrated to study the equivalence between these two different modeling techniques.

Please find the detail white paper here.

The associated example models can be found here.


Learn More about How Decision Tree Analysis Reduces Litigation Uncertainty and Facilitates Good Settlements by Marc B. Victor

Please click here to download a PDF.