Calculate average from effectiveness value
Boye Gricar
Hi ,
I have a Markov model using cohort analysis. What I would like to do is to create a payoff that gets the value in the eff column (the cumulative effectiveness) of the cost effectiveness text report and use this value to calculate an average by dividing it by the total number of cycles. My issue is that I have no idea how to get this cumulative effectiveness value and perform calculations on it.
Any suggestions would be appreciated!
Thanks,
Boye
Comments
Dear Boye,
Here's a solution, although it may not be the most elegant one. Other people may suggest more elegant solutions.
First, you have to set up the effectiveness variable (let's call it "EFFECT") as a payoff. To do this, go to the branch where you assign EFFECT, and access the "Markov Info" tab. If this is not already showing as one of the tabs (e.g. "Variable Properties", "Tracker Modifications" etc.), then go to "Views" on the ribbon. Click on "Markov Info" and the tab will appear at the bottom of the screen. If the branch you've chosen is a state, you will have the option of entering EFFECT as an initial value, an incremental value and a final value. In most cases, you will want to enter EFFECT just once, in the incremental value place. If the branch is a transition, you will be able to enter EFFECT just once.
Now you have the issue of calculating the average over N cycles. An easy way to do this is to allow "N_cycles" to be a variable in the model, so that you can change the time horizon at will. Then, instead of accumulating EFFECT at each cycle, change the Markov reward to EFFECT/N_cycles.
Don't forget to include discounting, if that is relevant. (TreeAge can do this automatically for you if you tell it to.)
Please sign in to leave a comment.