To say that an outcome is "unidimensionally a good thing" does not imply that all interventions that have that outcome as part of their outcomes, are good interventions.

(to say nothing as to the question of whether the outcome in question actually is unidimensionally a good thing)