Early Youth Engagement: service dis/engagement and implementation of digital interventions in young people’s and early psychosis community mental health services

Psychosis is the primary reason for referral to secondary mental health services, early intervention after a first episode reduces the chance of relapse, however, service disengagement is a problem for this patient group. Utilising digital technologies might help ameliorate some of the problems associated with this, but clinical implementation is challenging, and real-world evaluation is often overlooked in preliminary research. Additionally, little is known about social processes in online groups and what might be missing compared to the in-person format.

This thesis provides an up-to-date synthesis of predictors and rates of disengagement in the Early Intervention service for Psychosis (EIP). We also evaluated implementation of two real-world digital interventions, one delivering online groups for young people in mental health services, and the other evaluating the use of an online mental health companion ‘Likemind’ developed for the Early Youth Engagement (EYE-2 project); of which this PhD is affiliated. This is the first detailed implementation study of the Likemind online resources, and we used a cross sectional survey design to ask clinicians about website use, internet skills and accessibility and their attitudes towards sharing it in clinical sessions as a decision-making tool. Using the Normalisation Process theory (NPT) framework we identified barriers and facilitators of implementation.

For a second study we investigated social connection and self-stigma in online groups for young people in mental health services. We introduced a novel avatar platform (ProReal) to see if it was able to enhance the group experience. We used the NASSS framework (Non-adoption, Abandonment, Spread, Scale-up and Sustainability) to conduct a framework analysis evaluating real-world implementation of online groups with and without ProReal.

Our systematic review supports existing findings that robust predictors of disengagement are substance-use, medication non-adherence and low symptomology. From the meta-analysis, our pooled average was 15.6%. Methodological variation makes comparisons challenging. From our synthesis, we give recommendations for researchers to create a more cohesive evidence base in future research enabling more meaningful comparisons to be drawn from the data.

From our theory driven implementation studies we developed a theoretical framework of barriers and facilitators for youth mental health services to help guide preliminary research of new digital interventions in real world settings. We found clinician attitudes were generally good for digital implementation and service users welcomed the use of digital technologies. Despite this, clinicians were not committed to ongoing support of digital interventions and preferred existing practices. Key clinical barriers were identified around organisational support, clinician skills, and training. Adding more novel software could help to help improve the intervention value but adds multiple complexities across other implementation domains. Staff turnover and sickness is a barrier for embedding especially where the intervention requires specialist training. Digital interventions appear to be more suited to some service users rather than others. For successful implementation, it is recommended that simpler interventions are used where possible at a national level. Where there is high value in more complex interventions, implementation might work better in dedicated, regional digital teams. Going forward testing of our model with larger sample sizes in a variety of clinical settings would help establish replicability of our findings.