Probabilistic sensitivity analysis with multiple payoffs
Hi,
I have developed a simple decision model comparing two types of diagnostic tests. The model has been set up to produce expected costs for each strategy and the number of patients identified as true positive, false positive, true negative and false negative at the terminal nodes. So, I have 4 payoffs - expected costs and expected number of TP, FP, TN, FN. When the model is rolled back results are produced fine, although the values at the TP, FP, TN and FN nodes need to be added together to produce the total number of each (TP, FP, TN and FN) for each strategy.
I also have a dirichlet distribution in the model and would like to run the model probabilistically. When I run a probabilistic analysis the model produces probabilistic results for the costs. However, I am unsure how to get probabilistic results for the TP, FP, TN and FN data. Can you help me with this? Thank you.
Comments
There was an issue in this model related to the payoff entries for counting outcomes like TP, FP, TN, FN.
To count outcomes like this, you should enter a 1 into the appropriate payoff set representing that outcome at that terminal node. The 1 is then multiplied by the probability of reaching that terminal node in the model. If there are multiple terminal nodes reflecting the same outcome, the total reported value for a strategy will be the sum of the probabilities of reaching each of those terminal nodes.
Within the context of a Markov model, the 1 would more likely be placed in transition rewards at specific events which may not be terminal nodes.
Please sign in to leave a comment.