We agree with the thoughtful letter by Wong and colleagues up to a point. The Crisis team Optimisation and RElapse prevention (CORE) Fidelity Scale for crisis resolution teams (CRTs) was based mainly on stakeholders’ opinions rather than robust empirical evidence regarding components of effective crisis care.Reference Lloyd-Evans, Bond, Ruud, Ivanecka, Gray and Osborn¹ Some fidelity items may be more important than others, and some items may not constitute critical ingredients of effective CRTs.

The CORE service improvement programme evaluated in our trialReference Lloyd-Evans, Osborn, Marston, Lamb, Ambler and Hunter² built in a lot of flexibility and ownership for teams to choose their own goals for improving their service and plan how these would be achieved, in their local context, given their available resources. This flexibility in the programme was valued by the teams. We agree that giving CRT teams dedicated time and space to reflect on their team's performance and how this could be improved, and offering support from an experienced clinician (the CRT facilitator), are both important components of the programme.

We do not recommend that practitioners should ignore CRT model fidelity, however, for two reasons. First, the CORE CRT Fidelity Scale specifies many aspects of CRT service organisation and delivery, and the total fidelity score is a fairly blunt measure. Although our trial found no relationship between CRT total fidelity score and hospital admission or CRT patients’ readmission rates, we did find relationships between these outcomes and fidelity scale subscale scores, as reported in our paper.Reference Lloyd-Evans, Osborn, Marston, Lamb, Ambler and Hunter² Our results suggest that to avert hospital admissions requires rapid, easy access to CRT care (the access and referrals subscale); while to help CRT patients recover and avoid readmissions to acute care requires provision of good quality CRT care (the content of care, and timing and location of care subscales). This makes intuitive and clinical sense. Different fidelity items may be most important for different outcomes but are diluted in the total fidelity score.

Second, seeking to improve model fidelity was an integral part of our trial's successful CRT service improvement programme. CRT teams’ whole-team scoping day and their service improvement plans were informed by a fidelity review. Teams targeted specific items from the CRT Fidelity Scale (a median of eight items per team) as the means by which to improve their service. Our trial demonstrated that a service improvement programme, informed by a CRT fidelity review and focused on improving model fidelity, was successful in reducing hospital admissions and CRT patients’ readmissions to acute care. Wong and colleagues’ suggestion that this could be achieved just as successfully without reference to model fidelity is an untested assertion.

Our exploration of the relationship between CRT Fidelity Scale scores and outcomes involved only 25 teams in the unusual context of a trial. Further research is desirable to establish the relationship between model fidelity and outcomes, and, in time ideally, to refine the CRT Fidelity Scale to include only items demonstrated to constitute critical components of the CRT model.

In the meantime, the CORE CRT Fidelity Scale may not provide a blueprint, but does offer a helpful guide for practitioners and service planners in what an effective, high-quality CRT service looks like. As such, it is recognised as a descriptor of best practice for CRTs in current NHS England policy guidance.³