It would make more sense to have the decision model check for each possible reason whether it applies in the current case and then have the resulting consequences determined by a bog-standard if-then-else program.