Children in intensive care are being kept on breathing machines longer than necessary because doctors lack a standardised, evidence-based plan for safely reducing sedation and removing the ventilator. This matters because every extra day on a ventilator raises the risk of infection, muscle weakness, and psychological trauma for the child, while also consuming scarce intensive care beds and driving up NHS costs. The SANDWICH trial tests a new protocol that gives nurses a greater role in sedation decisions, uses a structured tool to assess how deeply sedated each child is, and triggers a spontaneous breathing trial at a higher threshold than usual—so the ventilator is removed as soon as the child is ready, rather than after a slow, gradual reduction of support. If the protocol works, it could shorten ventilation by a clinically meaningful amount, reduce complications such as unplanned extubation and post-extubation stridor, and save the NHS money by freeing up intensive care capacity. The trial also includes a health economic analysis to calculate the cost per complication avoided, ensuring any change in practice is both clinically effective and financially sustainable.
View original technical description
DESIGN: Cluster-randomised stepped wedge clinical and cost-effectiveness trial with an internal pilot and a process evaluation. SETTING: PICUs in the UK with a case mix typical of UK critical care practice. TARGET POPULATION: Invasively mechanically ventilated children admitted to UK PICUs. HEALTH TECHNOLOGY BEING ASSESSED: A protocol-based intervention incorporating co-ordinated care with greater nursing involvement; patient-relevant sedation plans linked to regular assessment using the COMFORT sedation tool; regular assessment of ventilation parameters with a higher than usual trigger for undertaking an extubation readiness test; a spontaneous breathing trial (SBT) on low levels of respiratory support to test extubation readiness. COMPARATOR: Usual care that is non-protocol-based and primarily medically-driven. Sedation levels are assessed using the COMFORT tool. Ventilator weaning generally involves slow, gradual reduction of pressure support to very low levels to test readiness for extubation. MEASUREMENT OF COSTS AND OUTCOMES: The primary outcome is duration of IMV. Secondary outcomes are: unplanned extubations; reintubation; total duration of IMV; post-extubation use of non-invasive ventilation; any adverse events (e.g. unplanned removal of any invasive tube); post-extubation stridor; tracheostomy insertion; PICU and hospital length of stay; mortality (PICU and hospital); cost per complication avoided at 28-days. Outcomes will be measured from patient admission up to 90 days or discharge (whichever is earlier); or a maximum of 28-days after the 20-month enrolment period. The process evaluation will evaluate dose, fidelity and reach of the intervention, acceptability of staff delivering the intervention and factors promoting sustainability. HEALTH ECONOMIC EVALUATION: A within-trial economic evaluation will be undertaken to measure cost-effectiveness of the intervention with standard care. The perspective of the analysis will be the hospital. The incremental cost-effectiveness ratio (ICER) is a measure of the additional cost per additional unit of effect produced by one intervention compared with another. We will calculate the cost per complication avoided at 28-days. SAMPLE SIZE: To inform the power calculation we used PICU admissions data for the years 2014-2016 from 18 units participating in the trial to determine parameters to inform the sample size calculation. In this trial, duration of ventilation is censored at the point of transitioning from the control to the training period, discharge to another hospital, at 90-days, death, and receiving a tracheostomy so applying censoring to this dataset provided us with a homogeneous population that more accurately reflected the trial population. The mean duration of mechanical ventilation was 5.8 (SD 9.6) days and an ICC (95% CI) of 0.005 (0.001 – 0.01). It is postulated that a reduction of one day on ventilation is both clinically important and achievable. The app https://clusterrcts.shinyapps.io/rshinyapp/ was used to update the sample size calculation given this revised information38. Using this app and for the actual design of the trial (using the actual information on the number of clusters and number of steps, rather than approximated values and using the following assumptions: no. clusters per sequence=1, ICC=0.005 (with consideration across the range 0.001-0.01), an exchangeable correlation structure, mean difference=1, SD=9.6, at 5% significance level, the power is approximately 80% for a cluster size of 28. The calculation is based on a standardised effect size (mean and SD) rather than the Hazard Ratio because the Shinyapp does not yet accommodate survival outcomes. This is a conservative approach meaning that it should have slightly underestimated the power not having allowed for the time to event nature of the data. The expected sample size is 9520 based on an average cluster size of 28 patients per block. PROJECT TIMETABLE: The trial duration is 36 months that includes a 9-month start-period, 20-month trial period and 7-month close-down period to allow for follow-up, cost effectiveness, data analysis and reporting. EXPERTISE IN TEAM: The team comprises multi-professionals with relevant skills to successfully conduct and complete the trial. Expertise includes: PICU clinical experience; conduct of multi-centre clinical trials, stepped wedge design and statistics; health economics; epidemiology; and ethnography.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know