Implementing TMLE in the presence of a continuous outcome