Author's reply

We welcome the UK Society for Behaviour Analysis’ interest in our work and we agree that there are important issues in considering fidelity in complex interventions. All too often, psychosocial interventions that have worked well in controlled conditions fail when tested in real-life settings.¹ This has been observed across many other interventions in mental health and raises the question about how to integrate findings from negative trials with what is known from small-scale early-phase trials, before–after or controlled studies, and n = 1 experiments. Many factors influence fidelity of a complex intervention, for example participant characteristics, intervention complexity and organisational issues.² Clearly, including implementation information in future trials of psychosocial interventions will be paramount in supporting the delivery of evidence-based care in the field of intellectual developmental disability and challenging behaviour.

In our pragmatic trial, which examined the clinical effectiveness of staff training in positive behaviour support for challenging behaviour in routine care, we made efforts to address implementation by training, mentoring, site visits, monthly teleconferences with the trainers in order to aid motivation and help the therapists problem solve. Clearly practitioner skill and competence play an important role in delivering interventions successfully; all professionals who volunteered to act as therapists in the study have had significant clinical experience in the field of intellectual disability.

There were other reasons for poor fidelity of the intervention that are stated in our report in Health Technology Assessment.³ However, our study findings highlight a more pressing question about the current level of implementation of positive behaviour support in community practice. Already a training programme at service level for National Health Service and social care staff to enable them to carry out positive behaviour support has been rolled out at a cost of over £500 000 but the long-term evaluation of its impact is unknown.⁴ It may also be necessary to consider the feasibility of positive behaviour support (as currently defined) being delivered as part of routine care; alternatively, other interventions could be explored. Without further evidence as to what is delivered and by whom, and of the real-life effectiveness of established interventions such as positive behaviour support we may fail people with intellectual developmental disabilities and their carers.

We would also like to correct an inaccuracy in the letter; indeed, paperwork was submitted for a proportion of the participants, please see our article in the BJPsych (p. 165) under ‘Fidelity of intervention and implementation’.