Smartphone-Based Self-Monitoring in First Episode Psychosis: Mixed-Methods Study of Barriers and Facilitators to Engagement.

Background: Despite growing interest in smartphone technologies to monitor patient outcomes, only a limited number of studies have investigated the long-term use of smartphone applications in patients with psychosis. Therefore, we have a limited understanding of facilitators and barriers to sustained engagement with these technologies in this population. Objective: This study aimed to examine the impacts of demographic and clinical characteristics on engagement with smartphone-based assessments in people who experienced a first episode of psychosis (FEP); to assess acceptability, barriers, and motivators influencing long-term engagement; to understand how to improve users' experience in future smartphone-based research; and to better understand participants' attitude toward the use of digital technologies in clinical settings. Methods: This study used a mixed-methods design. Linear regression models were used to assess the association between demographics and clinical variables (global assessment of functioning [GAF], global functioning [GF], Positive and Negative Syndrome Scale [PANSS], and Hamilton Rating Scale for depression [HAM-D]) and the proportion of completed app-based assessments during the 12-month period. Qualitative data were analyzed using a hybrid deductive and inductive thematic approach. Results: A total of 274 participants were recruited. The average smartphone-assessment completion rate was 14.5% (53/365 daily assessments completed over a 1-year period). No statistically significant associations were found with demographic characteristics, while there was a positive association with the baseline level of functioning (GAF disability: coefficient 0.173; 95% CI 0.022-0.324), GAF symptoms (coefficient 0.155, 95% CI 0.011-0.298), and GF social (coefficient 2.27, 95% CI 0.474-4.061). In contrast, the completion rate was negatively associated with PANSS general subscales (coefficient, -0.589, 95% CI -0.923 to -0.254) and depression (coefficient, -0.524, 95% CI -0.874 to -0.174). A total of 20 participants completed the qualitative sub-study. Participants acknowledged the ease of using the app and the user-friendly interface. They also valued the repetitive nature of the assessments, noting that it encouraged self-reflection. However, daily assessments felt time-consuming and repetitive over time, and some participants found it challenging to remain engaged due to competing life priorities. While some participants felt the app lacked tangible benefits, others reported indirect benefits from tracking their daily experiences. The opportunity to contribute to research was highly valued, as well as the importance of human connection with the researchers in sustaining long-term engagement. Participants made several suggestions: reducing the frequency of smartphone assessments to 1-2 times per week, integrating passive and active monitoring tools, and incorporating individualized features. Most participants supported the idea of using an app as part of their clinical care, provided data security and privacy were adequately addressed. Conclusions: Our study suggests that smartphone-based technologies are an acceptable monitoring tool for patients with first episode of psychosis. While participants expressed interest in using smartphone apps as a clinical tool, they emphasized that this should complement, rather than replace, the therapeutic relationship. Patients experiencing more severe symptoms may require additional support and incentives to sustain engagement over time.