Cost-effectiveness decision tree expected values
Source:R/dectree_expected_default.R
dectree_expected_default.RdRoot node expected value as the weighted mean of probability and edge/node values e.g. costs or QALYS.
Arguments
- vals
Values on each edge/branch e.g. costs or QALYs (array)
- p
Transition probabilities matrix
- dat
Long node-edge value array; default:
NA
Details
The expected value at each node is calculate by
$$\hat{c}_i = c_i + \sum p_{ij} \hat{c}_j$$
The default calculation assumes that the costs are associated with the nodes. An alternative would be to associate them with the edges. For total expected cost this doesn't matter but for the other nodes this is different to assuming the costs are assigned to the nodes. The expected value would then be
$$\hat{c}_i = \sum p_{ij} (c_{ij} + \hat{c}_j)$$