That's not a valid generalization. In this case, negative value of the secondary health hazards, loss of enjoyment, and discomfort experienced during fasting might exceed the cost of the drug treatment.
Well, if that's your logic then un-modeled externalities and other physiological disruptions should be taken into account, which are likely to be much more salient than in the case of simply not eating.
Fasting is, in fact, about the most reasonable and conservative 'treatment' imaginable since it's been in active deployment for the last however-many million years across diverse phyla.
