Obtaining informed consent can be challenging in stressful and urgent circumstances. One example is when potential participants have recently had their child removed into care; intervention is urgent and mandatory whereas participation in associated research is voluntary. Using a nested qualitative study, we examined experiences of consent processes in a randomised controlled trial of a family assessment and intervention service for maltreated young children. Some potential participants found it difficult to use information; some believed consenting might influence the return of their child. In response to these ethical challenges, we propose reversing the typical process of securing consent, so that randomisation to an intervention occurs before inviting potential participants to consider the trial. This will avoid delays, delineate research from intervention, and make it easier to consider information. We suggest that this innovation could be useful in trials across service areas that incorporate urgent and complex interventions.