Flexidelity – assessing fidelity of delivery of a complex intervention to help people live well with persistent symptoms.
Multiple Symptoms Study 3 is a randomised controlled trial of a complex intervention to help people live well with persistent physical symptoms. The intervention is delivered by an extended role GP across one 45 minute followed by three 15 minute consultations. The intervention centres around 4 main components: Recognition, Explanation, Action and Learning (REAL). The GPs received training at the start of the study and regular supervision thereafter. The intervention is highly complex: it comprises many interacting components and behaviours and a high degree of flexibility, allowing GPs to tailor the intervention to individual patients and their own personal consultation style. How can fidelity in delivery of the intervention be assessed in this case?
We assessed fidelity as part of a process evaluation nested within the trial. To do this, we specified a set of key elements of the intervention which were expected over the course of the 4 consultations. This fidelity framework was applied by a researcher to transcripts of a sample of sets of consultations per GP. For each element the researcher noted if it was present, provided an example and, where appropriate an informative comment. We used a traffic light system to code elements as ‘present’, ‘questionable’, ‘problematic’ or ‘missing’. Any other than ‘present’ were highlighted, and the Principle Investigator (PI) was alerted. The PI would then look at the element in context, and the issue would be addressed in the GP’s supervision meeting.
So far, 40 sets of transcripts have been fidelity assessed. Of these, 11 consultation sets contained one or more ‘questionable’ or ‘missing’ element’. Some consultations sets had both ‘missing’ and ‘questionable’ elements. 29 out of 40 consultation sets had 100% ‘present’ elements. These numbers indicate that teaching has been successful. The traffic light system has allowed early identification of potential problem areas, which were addressed in supervision. Assessing fidelity in this way has increased the replicability of the study. The presentation will provide examples of how key elements were identified from the consultations.
Despite variation in GP’s style of delivery, and high heterogeneity of patients, it has been possible to reliably track fidelity using this method. The GPs had very different styles, yet fidelity to key components was strong. This indicates the effectiveness of the training, which included items such as deep listening, which GPs are already familiar with, as well as more unfamiliar elements such as advanced explanation and symptom control techniques.
Assessing fidelity to flexible complex interventions is challenging. We propose the portmanteau term ‘flexidelity’ to describe one approach to this. Certain components are integral to the intervention, and identifying the presence of these within complex data. This increases the reliability of the findings, and makes it easier for the intervention to be replicated.