Using digital phenotyping measures to predict the symptoms and functional outcomes in first episode of psychosis
JianLi Wang
Phil G. Tibbo
Impact
This project demonstrated that assessing people’s behaviour, physiology, and psychological state using data from personal digital devices like smartphones and wearables is both feasible and acceptable for those with first-episode psychosis (FEP), supporting the future use of digital tools in mental health research and care.
The project was very forward-looking as it demonstrated that digital tools can reliably capture symptoms, mood, behaviour, sleep and daily activities outside of clinical settings, while highlighting the potential to support earlier identification of symptom changes, improve clinical decision-making, and creating more personalized care in early psychosis intervention.
For the Nova Scotia Early Psychosis Program (NSEPP), the research provides locally relevant evidence to guide future integration of digital tools into early psychosis services and supports future research initiatives in this area.
This work lays the foundation for larger studies exploring whether digital phenotyping can help predict symptom and functional outcomes and support broader implementation of digital mental health approaches in Canada.
Strictly from my (CMHA) point of view, it has introduced a different way of thinking about AI in the community mental health space. With so many private companies developing digital mental health applications, few are involving people with lived experience and community sector stakeholders in development and ethical assessment. We are beginning to envision an empowering and active role for people to play in technological innovation.
“This award has allowed our team to explore the feasibility of digital phenotyping data collection in patients with FEP and examine the association between baseline digital phenotyping measures and outcomes at follow-up. This project has provided valuable experience about how digital phenotyping data should be collected in these patients and laid a solid foundation for future research and collaboration with other early psychosis programs in Canada and other countries.”
— JianLi Wang and Phil G. Tibbo
-
Nova Scotia
-
Grant
-
Dalhousie University
-
Nova Scotia Early Psychosis Program (NSEPP)
-
IAM
-
2023-2024
-
Digital Mental Health
-
Patients, People with Disabilities
About the Project
This project explored whether data collected through smartphones and wearable devices could help predict symptoms and daily functioning in people experiencing first-episode psychosis (FEP). The main goal of the study was to understand if digital tools could support earlier detection of changes in mental health and improve patient monitoring and care.
Methodology
Researchers recruited 40 participants aged 12–35 through the Nova Scotia Early Psychosis Program (NSEPP). Following a baseline assessment for all participants, they took part in a six-day monitoring period during which they completed brief smartphone surveys six times daily while wearing a Fitbit smartwatch. The smartphone surveys gathered information about participants’ location, daily activities, social interactions, mood, and symptoms, while the Fitbit collected data on participants’ activity, sleep, and heart rate. These data were used to describe participants’ day-to-day experiences and physical patterns and were combined with baseline assessment data to examine whether they can help predict mental health symptoms and functioning at the six-month follow-up.
The study also evaluated the feasibility and acceptability of this digital approach by examining survey completion rates, Fitbit use, reasons for missed surveys, and responses to attention-check questions. In addition, in-depth one-on-one interviews were conducted with selected patients, caregivers, and clinicians to better understand their experiences and perspectives on using smartphone and wearable device data in early psychosis care and research.
Findings
All 40 participants completed the six-day digital monitoring period, from which 36 participants (90%) completed the six-month follow-up. Interviews were also completed with 19 patients, 6 caregivers, and 6 clinicians.
The following findings were made from the full study:
Digital phenotyping using smartphones and wearable devices was found to be feasible and generally acceptable for people with first-episode psychosis (FEP).
Participants completed an average of 71% of smartphone surveys, while adherence to wearable devices was higher, with 87% completion for activity and heart rate data and 91% completion for sleep data.
Patients, caregivers, and clinicians reported generally positive experiences with the study, with participants noting increased awareness of their mental health and physical activity and clinicians highlighting the value of collecting information between clinical visits. Reported barriers included survey burden, device discomfort, limited phone access during school or work, and occasional technical issues.
Psychotic symptoms were reported in approximately 10–20% of survey responses, while positive mood scores generally ranged from 60–65 out of 100.
Participants reported being at home in 70% of survey responses and alone during nearly half of survey moments. Cannabis use was more common than alcohol use and varied across days. Fitbit data showed generally low physical activity levels, with regular sleep and heart rate patterns but noticeable day-to-day variation.
Participants who were unemployed at the baseline assessment showed higher psychiatric symptom levels, more severe depressive symptoms, poorer social functioning, and greater illness severity at follow-up compared with employed participants. Gender diverse participants showed fewer negative symptoms than men.
Overall, the findings suggest that digital phenotyping can reliably capture daily experiences, symptoms, and behaviours in people with FEP and may support future personalized approaches to mental health care and symptom monitoring.
Project Outreach
The project has not expanded beyond Nova Scotia or its current scope. Future expansion will be considered after the final analysis is completed and funding becomes available. At that stage, opportunities for collaboration with national and international partners may be explored to support further research or broader implementation.
Resources Created
Manuscript: Yang, L; Ross, K; Stewart, K; Palmer, M; Ye, Q; Olson, D; Abidi, S; Morrison, J; Tibbo, P; Wang, JL (2025) Perceptions and experiences of patients, caregivers, and clinicians on using digital phenotyping measures for outcome prediction in first episode of psychosis (FEP): A qualitative study" Early Intervention in Psychiatry (submitted Sept 15th 2025).
Poster presentation: Perceptions and experiences of patients, caregivers, and clinicians on using digital phenotyping measures for outcome prediction in first episode of psychosis (FEP): A qualitative study. Dalhousie University Psychiatry Research Day, November 2025, Halifax, Canada.
Poster presentation: Patient experiences and perspectives on using digital phenotyping measures for outcome prediction in first episode of psychosis: A qualitative study. MedInfo conference, August 2025, Taipei, Taiwan.