13.3 Distribution sampling options

There are three sampling options which can be selected when you create distributions and these can be used to make the distributions more flexible in how they are used in the models. The sampling rates are:

  • Re-sample per EV/group of trials: sample possible values of uncertain parameters for probabilistic sensitivity analysis.

  • Re-sample per Markov stage: sample values from the distribution per Markov stage.

  • Re-sample per individual trial: sample for every trial/patient in a Patient Level Simulation model.

To change the sampling rate of a distribution:

  • Choose Views > Distributions from the toolbar.

  • Select a distribution from the list.

  • Click the "edit" toolbar button to open the Add/Change Distribution dialog (as below).

  • Select the Sampling rate option.

Re-sample per EV/group of trials

By default, distributions are set to sample once per EV recalculation of the model. This provides a different sample from the distribution for each calculation of the whole model per EV calculation.

In a two-dimensional simulation ("trials and samples"), this re-sampling would be per “group of trials”. By default, a simple microsimulation re-samples only distributions set to sample per 1st-order trial or Markov stage.

Resample per individual trial

While most distributions are used to sample possible values of uncertain parameters for probabilistic sensitivity analysis, distributions can also be used to instead represent individual variability/patient characteristics. You can identify these two different classes of distributions via the TreeAge Pro distribution’s sampling rate property. This is particularly important if a model includes both types of distributions.

This option will be of interest in specific models:

  • Markov models using tracker variables to follow detailed patient history, including discrete event models.

  • Trees using some distributions to represent variability among individuals, or patient characteristics.

These distributions are set to sample once per 1st-order trial or patient.

Re-sample per Markov state

It is also possible to set a distribution’s sampling frequency to generate a new sample value at each Markov cycle/stage during first-order trials and/or cohort/EV calculations. The DistForce() syntax can be used to re-sample more frequently.

Sampling during non-simulation, EV calculations

What happens to distributions when you are not running patient level simulation or PSA? During non-simulation analyses, such as roll back and sensitivity analysis, a reference to a distribution returns the mean value every time it is referenced. If 'Override mean' is used, this value would be the override value for the calculated mean value. The Dist() function can, however, override this behavior and return a randomly sampled value from the referenced distribution during any EV calculation.

To cause a random distribution sample to be returned by the Dist() function during expected value (EV) calculations, simply add a second parameter to the function with a value of 1. The formula DistForce(n), or Dist(n;1), will sample a new value from distribution number n each time the distribution is referenced in a tree calculation.

This also means that at each place in the tree where the distribution is referenced, a different sample value will be returned.