Super learning to hedge against incorrect inference from arbitrary parametric assumptions in marginal structural modeling

Erroneous IPW inference about clinical effectiveness because of arbitrary and incorrect modeling decisions may be avoided with SL.