Age Problem in Markov Modeling
Annie
Dear TP modelers:
I am working on a project comparing screening test strategies. I am stuck on how to introduce different age cohorts in a Markov model. I'd like to introduce several hypothetical cohorts with different start ages in the model, and I wonder if I do this by establishing several decision nodes, with each node representing a single age cohort, is appropriate. And if it is viable, how can I incorporate the results if I seperate them at the beginning?
Thank you,
Best Regards
Comments
You could run a deterministic two-way sensitivity analysis on start age and stop age, This would give you the model's result for every possible cohort age range (i.e. a grid of values where each point represents a different combination of start age and stop age). If you only want to vary start age and have a fixed stop age, then you could run a one-way DSA on start age.
If you want the average across age ranges or start age, and each are weighted equally, you could export the text data from the DSAs to Excel and average overage over the results.
There are fancier ways of doing this using the _sample key word and tables that will calculate an average within Treeage. I can describe those if you'd like.
You might take a simple approach of running the model several times with different start ages. I would consider this like running the model under different scenarios. Then you could get full model results for different age cohorts.
You may find that different strategies are more cost-effective for different cohorts.
Andrew
The problem has been solved. Thank you so much for your help!
Please sign in to leave a comment.