I think the problem isn't mistakes in the methodology, nor is it utilitarianism in general, but one specific feature of DALYs: it is multiplicative instead of additive.
Since DALYs is the years of life lost times the disease burden (where the disability burden is a scalar determined by surveys), the marginal utility of a disabled persons years of life is multipled by the disability burden.
If instead you had a unit that was years of life lost minus disease burden (where the disability burden is a number of years that can determined by same surveys), the marginal utility of a disabled person's years of life and a healthy person's years of life is the same!
So the problem isn't trying to price every intervention in utils, it's that DALYs specifically are an incorrect way to weigh years of life v.s. disease burden.