For those doing the TreeAge version of HW5 (Markov): I got a couple of comments that the model was not computing properly, and specifically that you were getting the same number of QALYs gained for both arms. Try this:
Make sure “_stage” = “totalCycles’ in both Tx1 and Tx2.
In the Tx1 arm make sure your probability of death among advanced cancer patients is “pAdvancedToDeath”. This is in Indolent è Die
Set your probability of progressing to advanced cancer (“pIndolentToAdvanced”) in the original Tx1 node to “pIndolentToAdvanced1” and to pIndolentToAdvanced2 for Tx1 and Tx2 arms, respectively. This will work if the original probability to “pIndolentToAdvanced” is listed. This is provided in the HW5 TreeAge file. See the variable list Windowsè View è Variables.
Explanation: You must use “pIndolentToAdvanced” in Tx1 rather than “pIndolentToAdvanced1 because Tx1 is cloned in Tx2. If you specify “pIndolentToAdvanced1 in Tx1, that same value will be used in Tx2. But if “pIndolentToAdvanced is set equal to “pIndolentToAdvanced1” in the variable specifications for Tx1, and to “pIndolentToAdvanced2” in the specifications for Tx2, the differing values for “pIndolentToAdvanced” are correctly implemented in the respective arms.
You can also forgo “pIndolentToAdvanced” completely and go straight to “pIndolentToAdvanced1” in Tx1. Then you can set “pIndolentToAdvanced1”=”pIndolentToAdvanced2” in Tx2
Either approach works.
For those of you who ran into trouble, I invite you to resubmit corrected homework if you want. However - This is a bit tricky if you don't have previous TreeAge experience. I will not mark you down if you made a good faith effort to get this right the first time.
- Elliot