Using Passive Sensing to Predict Psychosis Relapse: An In-Depth Qualitative Study Exploring Perspectives of People With Psychosis.

Abstract

Background: Relapses result in negative consequences for individuals with psychosis and considerable health service costs. Digital remote monitoring (DRM) systems incorporating "passive sensing" (sensor data gathered via smartphones/wearables) may be a low-burden method for identifying relapses early, enabling prompt intervention and potentially averting the consequences of full relapse.

Objective: This study examined detailed views from people with psychosis about using passive sensing in this context.

Study design: Qualitative interviews, analyzed using reflexive thematic analysis. Setting: Secondary care mental health services across the United Kingdom. An advisory group with relevant lived experience was involved throughout, from developing the topic guide to analysis. Participants: Clinician confirmed diagnosis of schizophrenia-spectrum psychosis (n = 58).

Study results: Four overarching themes were developed. Theme 1 outlined participants' polarized feelings about passive sensing, highlighting specific challenges relating to privacy, especially regarding location data. Theme 2 examined participants' fears that clinicians might judge their movements or routines, creating a sense of pressure to modify their actions and undermining their autonomy. Theme 3 described potential solutions: offering users choice about what data are shared, when, and with whom. Theme 4 outlined specific benefits that participants valued, including intended functions of passive sensing within DRM (ease of use, early identification of relapse, and relevance of sleep monitoring) and novel uses.

Conclusions: Our findings underline the importance of fully informed consent, choice, and autonomy. Given the potential privacy impacts, individuals are unlikely to engage with passive sensing unless they perceive clear personal benefits. Prospective DRM users need clear, accessible information about passive data collection and its relevant costs and benefits.