Next: , Previous: Study/Model, Up: Top


10 Transitions

Transitions connect states/processes within a study/model. Below, they are indicated by solid arrows and indicate the paths that subjects can take through the model. Transitions also hold progression rate information for studies or transition probabilities for models.



10.1 Transitions for Models

10.1.1 Creating Transitions for Models

  1. From the main window, click 'Transitions' on the left navigation pane. This form can also be reached from a Study/Model form by selecting the transition button for a certain Study/Model.



  2. This form shows the transitions for a given Study/Model. Select the Model to be used from the drop-down box (A). Note: if the Transitions page was opened from the Study/Model page, that Study/Model will be selected and the combo box will be grayed out.
  3. Click the 'Add' button (B).
  4. Select the origin state for the transition (D).
  5. Select the destination state for the transition (E).
  6. Enter the transition probability in the box (F) as an expression, which may or may not include a Parameter. Parameters can be directly selected as the expression using the drop-down list.
  7. Close the form to save the entry.



For a video demonstration on creating Transitions, and a model, click here.

10.1.2 Removing Transitions

Identify the row that will be removed, and click the 'X' button (C) in that row. This may require deletion of other entities and may be difficult if the deletion candidate was extensively used.

10.2 Transitions for Studies

Defining study transitions is related to parameter Estimation. Study transitions may mean different things depending on the cases described below.

10.2.1 Creating Transitions for Studies

  1. From the main window, click 'Transitions' on the left navigation pane.



  2. This form shows the transitions for a given Study/Model. Select the Study to be used from the drop-down box (A). Note: if the Transitions page was opened from the Study/Model page, that Study/Model will be selected and the combo box will be grayed out.
  3. Click the 'Add' button (B).
  4. Select the origin state for the transition (C).
  5. Select the destination state for the transition (D). Leave blank to indicate initial population count of the study.
  6. (Optional) Enter the Initial Population Count / Incident Count in the box (E). This field should be left blank for regression studies and filled for studies that report incident counts. Note that studies will in many cases require two transitions, the first for indicating initial count of people starting at a certain state, and the second for indicating the number of people ending at the end state at the end of the study. Both transitions will start at the same state, the first will have no end state indicating initial population count and the latter will define the ending state indicating the incident count reaching this end state.
  7. (Optional) Enter a 'Regression Parameters Vector' (F). This field should be left blank for studies that provide results as incident counts. For a regression study it must be filled with a vector of the form [param0, param1, ...] that contains the reported regression parameters. If a bias term is reported, it is represented as the value 1 rather than a parameter name. This vector is multiplied by the Regression Coefficient Vector when creating the regression equation reported in the study.
  8. (Optional) Enter a 'Regression Coefficients Vector' (H). This field should be left blank for studies that provide results as incident counts. For a regression study it must be filled with a vector of the form [coef0, coef1, ...] that contains the reported regression coefficients associated with the Regression Parameters Vector. Note that both vectors should be of the same size. This vector is multiplied with the Regression Parameters Vector when creating the regression equation reported in the study.
  9. (Optional) Enter a 'Covariance Matrix' (I) or select one defined as a parameter from the drop-down list. This field should be left blank for studies that provide results as incident counts. For a regression study it must be filled with a matrix that contains the reported covariance matrix associated with the Regression Parameters Vector. Note that the Matrix should be of a size compatible with both vectors.
  10. Close the form or move to the next record to save the entry. This will trigger validity checking of the data entered and if no error message is displayed, then the data has been saved to memory. Note that the information is not yet saved to a file.

10.2.2 Removing Transitions

Identify the row that will be removed, and click the 'X' button (C) in that row. This may require deletion of other entities and may be difficult if the deletion candidate was extensively used.

10.2.3 Copying Transitions from Another Study/Model

It may be desirable to build or modify a model/study by copying transition information from another model/study. The systems support this sort of copy using the following steps:

  1. Initiate the transitions copy by pressing the Copy From Study/Model button (K). The following form will appear:



  2. Select Study/Model you wish to copy the transitions from by clicking on it.
  3. Press OK to initiate the copy or Cancel to abort the copy operation.
  4. If the operation was not aborted, the system will bring a message indicating how many transitions were successfully copied. Dismiss this dialog box by pressing OK and the transitions form will display the copied transitions.

Note that this operation is useful while creating variations of a specific model that change the state/sub-process hierarchy. The copy transitions operation will try to copy all the transitions from the source model to the destination model. However, some transitions may not be copied as these may violate validation rules. Here are examples of transitions that will not be copied: