Summary. Randomized controlled trials (RCTs) can provide unbiased estimates of sample average treatment effects. However, a common concern is that RCTs may fail to provide unbiased estimates of population average treatment effects. We derive the assumptions that are required to identify population average treatment effects from RCTs. We provide placebo tests, which formally follow from the identifying assumptions and can assess whether they hold. We offer new research designs for estimating population effects that use non-randomized studies to adjust the RCT data. This approach is considered in a cost-effectiveness analysis of a clinical intervention: pulmonary artery catheterization.

HeadlineEvaluating service innovations in health care and public health requires flexibility, collaboration and pragmatism; this collection identifies robust, innovative and mixed methods to inform such evaluations.

Background: Anaesthetic medication administration errors are a significant threat to patient safety. In 2002, we began collecting data about the rate and nature of anaesthetic medication errors and implemented a variety of measures to reduce errors. Methods: Facilitated self-reporting of errors was carried out in 2002e2003. Subsequently, a medication safety bundle including 'smart' infusion pumps were implemented. During 2014 facilitated self-reporting commenced again. A barcodebased medication safety system was then implemented and the facilitated self-reporting was continued through 2015. Results: During 2002e2003, a total of 11 709 paper forms were returned. There were 73 reports of errors (0.62% of anaesthetics) and 27 reports of intercepted errors (0.23%). During 2014, 14 572 computerised forms were completed. There were 57 reports of errors (0.39%) and 11 reports of intercepted errors (0.075%). Errors associated with medication infusions were reduced in comparison with those recorded in 2002e2003 (P<0.001). The rate of syringe swap error was also reduced (P¼0.001). The reduction in error rate between 2002e2003 and 2014 was statistically significant (P¼0.0076 and P¼0.001 for errors and intercepted errors, respectively). From December 2014 through December 2015, 24 264 computerised forms were completed after implementation of a barcode-based medication safety system. There were 56 reports of errors (0.23%) and six reports of intercepted errors (0.025%). Vial swap errors in 2014e2015 were significantly reduced compared with those in 2014 (P¼0.004). The reduction in error rate after implementation of the barcode-based medication safety system was statistically significant (P¼0.0045 and P¼0.021 for errors and intercepted errors, respectively). Conclusions: Reforms intended to reduce medication errors were associated with substantial improvement.

Background: In Africa, the clinical syndrome of pneumonia remains the leading cause of morbidity and mortality in children in the post-neonatal period. This represents a significant burden on in-patient services. The targeted use of oxygen and simple, non-invasive methods of respiratory support may be a highly cost-effective means of improving outcome, but the optimal oxygen saturation threshold that results in benefit and the best strategy for delivery are yet to be tested in adequately powered randomised controlled trials. There is, however, an accumulating literature about the harms of oxygen therapy across a range of acute and emergency situations that have stimulated a number of trials investigating permissive hypoxia. Methods: In 4200 African children, aged 2 months to 12 years, presenting to 5 hospitals in East Africa with respiratory distress and hypoxia (oxygen saturation < 92%), the COAST trial will simultaneously evaluate two related interventions (targeted use of oxygen with respect to the optimal oxygen saturation threshold for treatment and mode of delivery) to reduce shorter-term mortality at 48-hours (primary endpoint), and longer-term morbidity and mortality to 28 days in a fractional factorial design, that compares: Liberal oxygenation (recommended care) compared with a strategy that permits hypoxia to SpO 2 > or = 80% (permissive hypoxia); andHigh flow using AIrVO 2 TM compared with low flow delivery (routine care). Discussion: The overarching objective is to address the key research gaps in the therapeutic use of oxygen in resource-limited setting in order to provide a better evidence base for future management guidelines. The trial has been designed to address the poor outcomes of children in sub-Saharan Africa, which are associated with high rates of in-hospital mortality, 9-10% (for those with oxygen saturations of 80-92%) and 26-30% case fatality for those with oxygen saturations <80%. Clinical trial registration: ISRCTN15622505 Trial status: Recruiting