Sensitivity Analysis on Table Values

One-way sensitivity analysis requires that you choose a parameter variable and set a lower and upper bound to create an uncertainty range for that parameter. Sometimes, you may want to run sensitivity analysis on table values, which is more complex.

One option is to setup the table with three value columns (along with the index column). The first, second and third value columns are used for the table's low values, parameter estimates and high values, respectively. Then you can use a variable to select the proper value column, both for base case analyses and sensitivity analysis.

Model File

• There is a model attached to this article that is used to demonstrate this technique.

Table

• The model uses a table to define time-dependent probabilities using numeric values.

Table Lookups/Value Column

• The probability of progression from local cancer to metastases changes with time based on a table. The value column used for base case analysis is 2 based on the following variable definitions.
• pLocalToMetastases = tLocalToMets[_stage; tLocalToMets_valcol]
• tLocalToMets_valcol = 2

Sensitivity Analysis

• If you run one-way sensitivity analysis on the variable tLocalToMets_val col with range from 1 to 3 and 2 intervals, the analysis will evaluate the model three times, with tLocalToMets_val set to 1, 2 and 3.
• Based on the variable definition for pLocalToMetastases, the probability of progression will change from the table values in the first value column (low values), to the second (parameter estimates), then to the third (high values).

Summary

• You can use sensitivity analysis to change a set of time-based values from low to base case to high values using this technique.