markov cohort model for a screening test
Dear TP modelers:
I am new to TP, working on a project comparing two screening test strategies. There is not a lot about screening modeling. I am stuck at how to combine decision tree and markov nodes incorporating True positie/false positive/true negative/false negative etc.
Wondering if anyone has experience? could you share a general framework of your model? or if not too much trouble, your .trex file?
many thanks. susan
Comments
Susan
From the TreeAge point of view I can share with you information you can find in the Help files (Help > Help Documentation) in regards to setting up screening - especially if you happen to be doing Bayes Revision. There is a Bayes Revision tool which is useful.
Take a look at Chapter 29: Bayes Revision in Decision Trees. This walks you through how to set up the model. (There are example models which are referred to in the help files. The models are in the Files view).
You can always contact support@treeage.com to get one-on-one assistance with your questions.
Thanks, Belinda
There is a key question regarding the structure of the model. Is the screening repeated over time?
If screening is repeated, then the screening must be performed within the Markov model after certain states where screening should be performed.
For example, you might have undiagnosed healthy and undiagnosed sick states that should both be screened on specific intervals. In such a case, you can use the sensitivity and specificity values directly in probabilities without Bayes' Revision.
Andrew
thank you so much, Belinda and Andrew for your quick replies. yes, the screening test is repeated every 2 years. So, seems like I can only do it in the Markov model, not in a decision tree. i have submitted a support ticket and attached my .trex, with a few specific questions. many thanks!
susan
Please sign in to leave a comment.