
48 Discrete Event Simulation for Health Technology Assessment
also enable analyses of the results for subgroups of the population, which
can be established at any time without having to redo all the work.
Storing all the outcomes individually may be interesting, but if it is not
directly relevant to the problem at hand, then the additional computer mem-
ory required and increased runtime are wasteful. For a single attribute—even
for a few—this is not going to be a noticeable problem, but it may become one
if the modeler is not careful to avoid specication of unnecessary attributes.
Moreover, most analysts do not enjoy wading through extensiv