Sample size
Our sample size (n=110) was determined in order to show a hypothesized two-fold increase (80% vs. 40%) in use of the IUD at 6 weeks in the immediate compared to delayed group, with a hypothesized loss-to-follow-up of 20%, power set to 90%, and a two-sided alpha of 0.05 and to allow for a sub-analysis of the group in each arm requiring vacuum aspiration to remove placental remains, assumed to be up to 50% of all women.