The treatment of newborns with safe and effective medicines is of critical importance for their outcome and subsequent quality of life. Despite this, it is still a common practice to prescribe medicines to neonates outside the label, extrapolating from dosing regimens and indications validated in older populations and based on non-neonatal pathophysiology. In a recent meta-analysis (2015) evaluating 829 (1994-2012) studies on prescribing practices in pediatric hospital care, off-label and unlicensed medicines prescriptions ranged from 12 to 71 % and 0.2 to 48 %. These authors hereby reconfirmed that (pre)term neonates were still most commonly exposed to off-label and unlicensed medicines [12].The federal US legislation and similar European initiatives have resulted in a relevant increase in pharmacological studies in children, with a subsequent significant increase in label changes. Unfortunately, too few included drug label changes specific to neonates. To further illustrate this, there were 406 pediatric label changes (Food and Drug Administration, 1997-2010), but only 23 drugs resulted in 11 labeled indications (e.g., linezolid, rocuronium, remifentanil, sevoflurane, stavudine, nevirapine) for neonates. The absence of label change was most commonly because of unproven efficacy, despite the fact that these compounds (e.g., paracetamol, caspofungin, valganciclovir) are likely relevant for neonates [14]. This may largely reflect the difficulties to proof efficacy due to heterogeneity in patients, variability in outcome variables, and uncertainty on biomarkers, in addition to demonstrating safety (risk of adverse drug events) in neonates. Despite the significant improvements in the knowledge on pharmacotherapy in older children, neonates still remain therapeutic orphans [2,14].In this issue of the journal, Campino et al. report on the rate of errors (calculation or accuracy) in intravenous medicine preparations when performed either bedside in 10 Spanish neonatal intensive care units (NICUs) or in hospital pharmacy (HP) services. The study group hereby evaluated the impact of a structured intervention (protocol standardization, education) on the number of errors and documented significant improvements in both calculation errors (1.35 to 0 %) and accuracy (54.7 and 38.3 % to 38.3 and 14.6 % in NICU and HP, respectively) [6]. This intervention paper follows on a previous paper which described a pre-intervention phase [5] and illustrates the potential impact of preventive strategies on the extent of drug errors.Development of a roadmap to improve neonatal pharmaceutical care should aim to improve the knowledge on clinical pharmacology, clinical pharmacy, and approaches to improve the knowledge and use of neonatal pharmacovigilance initiatives [1]. Another keystone component of such a roadmap should focus on neonatal drug development, i.e., drug development driven by neonatal pathophysiology (Fig. 1).