Describir: Estimating treatment effects from a randomized controlled trial with mid-trial design changes.