The implausibility of ‘usual care’ in an open system: sedation and weaning practices in Paediatric Intensive Care Units (PICUs) in the United Kingdom (UK)

Abstract

Background: The power of the randomised controlled trial depends upon its capacity to operate in a closed
system whereby the intervention is the only causal force acting upon the experimental group and absent in the
control group, permitting a valid assessment of intervention efficacy. Conversely, clinical arenas are open systems
where factors relating to context, resources, interpretation and actions of individuals will affect implementation and
effectiveness of interventions. Consequently, the comparator (usual care) can be difficult to define and variable in
multi-centre trials. Hence outcomes cannot be understood without considering usual care and factors that may
affect implementation and impact on the intervention.
Methods: Using a fieldwork approach, we describe PICU context, ‘usual’ practice in sedation and weaning from
mechanical ventilation, and factors affecting implementation prior to designing a trial involving a sedation and
ventilation weaning intervention. We collected data from 23 UK PICUs between June and November 2014 using
observation, individual and multi-disciplinary group interviews with staff.
Results: Pain and sedation practices were broadly similar in terms of drug usage and assessment tools. Sedation
protocols linking assessment to appropriate titration of sedatives and sedation holds were rarely used (9 % and 4 %
of PICUs respectively). Ventilator weaning was primarily a medical-led process with 39 % of PICUs engaging senior
nurses in the process: weaning protocols were rarely used (9 % of PICUs). Weaning methods were variably based
on clinician preference. No formal criteria or use of spontaneous breathing trials were used to test weaning
readiness. Seventeen PICUs (74 %) had prior engagement in multi-centre trials, but limited research nurse
availability. Barriers to previous trial implementation were intervention complexity, lack of belief in the evidence and
inadequate training. Facilitating factors were senior staff buy-in and dedicated research nurse provision.
Conclusions: We examined and identified contextual and organisational factors that may impact on the
implementation of our intervention. We found usual practice relating to sedation, analgesia and ventilator weaning
broadly similar, yet distinctively different from our proposed intervention, providing assurance in our ability to
evaluate intervention effects. The data will enable us to develop an implementation plan; considering these factors
we can more fully understand their impact on study outcomes.