43.4 Time Node Type

Time nodes are the key to DES models. Time nodes function in a similar way to chance nodes in that there are expressions under each of its branches. However, each expression represents the time-to-event for that event rather than a probability of that event occuring within the Markov cycle. The patient will move from the time node to the event with the shortest time-to-event.

Consider the time node Mild in the DES.Model.With.Health.States, as shown below.

Note there are two events, Mild to Severe and Die from Mild, and each has a time expression beneath it, representing the time passed before each event will occur.

• The expression Time_MildToDead refers to an Exponential distribution, parameterized with a fixed rate.

• The expression Time_MildToSevere is a variable that references the appropriate Exponential distribution for that strategy. This is done to support cloning.

Whether directly or indirectly, both time-to-event expressions reference distributions. When a patient reaches the Mild state, that patient's sampled values from those distributions are compared, and the patient will move to the node with the shortest time-to-event sample.

The second time node Severe only has one event which can happen, which is dying from the Severe condition. Therefore, only that event will occur with the time-to-event sample determining how long it takes for that event to occur.