Price GD, Heinz MV, Nemesure MD, McFadden J, Jacobson NC. Predicting symptom response and engagement in a digital intervention among individuals with schizophrenia and related psychoses. Front Psychiatry. 2022;13:807116. doi:10.3389/fpsyt.2022.807116
This study utilized data from a trial delivering a psychosocial smartphone app designed for patients with psychosis (App4Independence or A4i) to better understand personalized markers of digital intervention engagement and response. Machine learning models were applied to baseline data, app use data, and semi-structured interview data to predict response to change in symptoms, level of engagement, and qualitative impressions of the A4i app. Thirty-eight participants received the A4i app intervention for one month. Machine learning models were capable of moderately predicting participant engagement and experience with the app (r=0.39) as well as changes in psychosis symptom severity (r=0.32). Participants with high baseline interpersonal sensitivity, versus low, benefitted more from the A4i intervention in reducing symptom severity. Additionally, participants with lower baseline psychotic and obsessive-compulsive traits were predicted to benefit more. Higher baseline depression predicted both higher engagement and satisfaction with the app. These findings demonstrate the potential of predicting response to a digital intervention for psychosis using unique patient factors. This study further highlights the need to investigate individual response to and engagement with digital-based mental health interventions. Future research should consider how individual demographic characteristics influence engagement with a digital intervention.