What is the best way to interpret / explain the contribution analysis visualization in the results of my simulations?
Hello, thanks for reaching us.
Let's use an example to illustrate an interpretation of a contribution analysis.
Given some simulation results, one may want to analyse which patient descriptors impact the most the volume of a tumor at the end of the simulation, tumorVolume.tend, for a virtual population of 100 patients.
The resulting plot is the following:
On the left side are the 10 most impactful parameters.
The first one: logExtraLivingCellsWithAngio.tmin has the highest contribution.
Moreover, given this plot, tumorVolume.tend and logExtraLivingCellsWithAngio.tmin are positively correlated: the subpopulation containing the 50% of highest value for logExtraLivingCellsWithAngio has a median for tumorVolume.tend 360% higher than in the overall population. This can be deduced from the yellow horizontal bar.
As per the second most impactful parameter, logKdCellsByIsHuman.tmin, the correlation is negative: the sub population containing the 50% of highest value for logKdCellsByIsHuman has a median for tumorVolume.tend 98% lower than in the overall population.
If you need more details on how to run a contribution analysis and theoretical background, I encourage you to have a look at this specific How-to