“…Additional multivariable analyses (propensity score matching, stratifi cation, and regression adjustment) and analyses in infants in whom these medications were started early after initiation of TV substantiated this fi nding, consistent with the fi ndings from randomised trials of drugrelated respiratory depression in neonates. [45][46][47] Propensity score matching allowed the elimination of diff erences in baseline characteristics between infants who received O-SH-GA and those who did not. Potential confounders such as being on TV at NICU admission were balanced with this approach.…”