decision tree

[David Kantor]

Hello All,

Does anyone know of software, in Stata or otherwise, to create and manage decision trees?

I am aware of TreeAge, but I find it difficult to understand. So if there is anything in Stata, I 'd be pleased to know about it. I am tempted to write something myself, but I thought I'd ask before I launch into it.

What I need is to be able to...
1: set up a tree structure;
2: assign values at terminal nodes;
3: assign probability values to each branch at the internal nodes;
4: evaluate the mean value of a given measure (from step 2), weighting it according to the probabilities at the branches, as it filters back to the root;
5: repeat steps 2-4 (or maybe just 3 & 4) for a variety of scenarios having alternative values and probabilities.

Thanks if anyone has a suggestion.
--David

PS: I've been away from the Forum for a while, so I don't know if this has been covered.

[Carlo Lazzaro]

David:
I deal with health economic models pretty frequently and I use spreadsheets (as they are the easiest way to provide the model).
The software you mention is basically a spreadsheet add-on.
I'm not aware of any Stata community-contributed programme to handle this issue.

[David Kantor]

Hello Carlo,
Thank you for your quick reply.

Yes, we have been doing this already in Excel; so it's nice to know that we are not alone in doing it that way. And we, too, are doing it for health economics.

The issue is the multitudes of alternative scenarios. I have been dreaming up ways to do it in Excel and Stata.

Thanks again,
--David

[Carlo Lazzaro]

David:
instead of considering tons of deterministic scenarios (that do not take into account joint parameters uncertainty properly), why not running a probabilistic sensitivity analysis?
Years ago I learnt a lot (and it's still my favourite source of inspiration) from Andy Briggs and coll textbook (and related stuff): https://www.herc.ox.ac.uk/downloads/...mic-evaluation.