Evaluating Adjusted Per-Protocol Effect Estimators in Pragmatic Trials to Address Treatment Non-Adherence