Digital Phenotyping in Mental Health
Digital phenotyping, as defined by Onnela and colleagues (2021), refers to the real-time assessment of individual behaviors and characteristics in everyday life using data collected from personal digital devices (Torous et al., 2016). A central feature of this approach is the continuous and passive collection of individualized data, meaning users do not need to actively report information or use additional monitoring tools. This enables the identification of objective indicators of health conditions that were previously mainly assessed through self-report or subjective measures. Due to its ability to capture real-world patterns, digital phenotyping has gained increasing attention in research. For example, smartphone-based GPS data identifying mobility and location anomalies have been associated with relapse in individuals with schizophrenia, suggesting a novel digital marker of disease progression (Barnett et al., 2018).
Beyond smartphones, wearable devices have increasingly contributed to digital phenotyping by allowing the passive collection of physiological data (De Boer et al., 2023). These devices can directly measure indicators such as heart rate, body temperature, and physical activity, which are not reliably captured through smartphones alone. Technological advancements have improved their accuracy and accessibility, supporting their broader use in health monitoring (De Boer et al., 2023; Niknejad et al., 2020). Research has shown that step count data derived from wearables are highly consistent and generally accurate in both laboratory and real-world environments (Fuller et al., 2020). At the same time, the global adoption of wearable sensors has grown rapidly, with the number of connected devices increasing from 593 million in 2018 to approximately 1,105 million in 2022, representing nearly a doubling within a few years (Business Wire, 2021).
In addition to their increased availability, wearables have demonstrated high levels of user compliance and satisfaction when proper education and follow-up are provided (Milani et al., 2017). Expert working groups have further emphasized that wearable technologies can be effectively integrated into clinical practice when specific implementation principles are followed (Smuck et al., 2021). Their applications extend across populations; for instance, sensors placed on infants have detected patterns of motion complexity that may help predict later development of autism spectrum disorder (Wilson et al., 2021). Similarly, wearable-derived physical activity data have been used to identify recovery trajectories in young children following surgeries such as appendectomy, offering objective measures of surgical recovery (De Boer et al., 2021). Together, these findings highlight the substantial benefits of incorporating wearable technology into digital phenotyping.
Despite these advantages, important limitations remain. Digital phenotyping that relies on personal digital devices excludes certain groups, including young children, older adults with functional impairments, and individuals with cognitive or behavioral challenges. Furthermore, many low-income individuals lack access to smartphones or reliable internet connections, limiting their ability to participate in digital data collection. Even with wearable devices, financial barriers may persist or worsen inequalities if access is not ensured equitably (Kiang et al., 2021).
Although funding concerns exist, reimbursement programs have begun supporting the use of remote monitoring devices and the clinical review of collected data, including coverage through Medicare, with expected expansion to other insurers (De Boer et al., 2023). However, most wearable technologies still depend on smartphones for data storage and transmission via Bluetooth and cloud-based systems. As a result, smartphones remain an essential component of digital phenotyping. To promote broader inclusion, proactive strategies such as providing smartphones and wearables at low or no cost through insurance coverage or public programs may be necessary to reduce financial and technological barriers.
In conclusion, digital phenotyping represents a promising and objective method for capturing real-world behavioral and physiological data, thereby advancing the understanding of mental health conditions and disease processes. While smartphones established the foundation of this field, wearable devices significantly enhance data collection capabilities and broaden clinical applications across diverse populations. However, challenges related to access, cost, and technological dependence must be addressed to avoid widening health disparities. By combining technological innovation with equitable device provision and supportive reimbursement systems, digital phenotyping can continue to evolve toward more inclusive and effective healthcare solutions.
References:
Barnett, I., Torous, J., Staples, P., Sandoval, L., Keshavan, M., & Onnela, J. P. (2018). Relapse prediction in schizophrenia through digital phenotyping: A pilot study. Neuropsychopharmacology, 43(8), 1660–1666.
Business Wire. (2021, February 8). Global smart wearable market – market to grow by 19.48% from 2021–2026 [Press release]. Business Wire. https://www.businesswire.com/news/home/20210208005342/en/Global-Smart-Wearable-Market---Market-to-Grow-by-19.48-from-2021---2026---ResearchAndMarkets.com
De Boer, C., Ghomrawi, H., Many, B., Bouchard, M. E., Linton, S., Figueroa, A., ... & Abdullah, F. (2021). Utility of wearable sensors to assess postoperative recovery in pediatric patients after appendectomy. Journal of Surgical Research, 263, 160–166.
De Boer, C., Ghomrawi, H., Zeineddin, S., Linton, S., Kwon, S., & Abdullah, F. (2023). A call to expand the scope of digital phenotyping. Journal of Medical Internet Research, 25, e39546.
Fuller, D., Colwell, E., Low, J., Orychock, K., Tobin, M. A., Simango, B., ... & Taylor, N. G. (2020). Reliability and validity of commercially available wearable devices for measuring steps, energy expenditure, and heart rate: Systematic review. JMIR mHealth and uHealth, 8(9), e18694.
Kiang, M. V., Chen, J. T., Krieger, N., Buckee, C. O., Alexander, M. J., Baker, J. T., ... & Onnela, J. P. (2021). Sociodemographic characteristics of missing data in digital phenotyping. Scientific Reports, 11(1), 15408.
Milani, R. V., Lavie, C. J., Bober, R. M., Milani, A. R., & Ventura, H. O. (2017). Improving hypertension control and patient engagement using digital tools. The American Journal of Medicine, 130(1), 14–20.
Niknejad, N., Ismail, W. B., Mardani, A., Liao, H., & Ghani, I. (2020). A comprehensive overview of smart wearables: The state of the art literature, recent advances, and future challenges. Engineering Applications of Artificial Intelligence, 90, 103529.
Onnela, J. P. (2021). Digital phenotyping and Beiwe research platform. Harvard T.H. Chan School of Public Health. https://www.hsph.harvard.edu/onnela-lab/beiwe-research-platform/
Smuck, M., Odonkor, C. A., Wilt, J. K., Schmidt, N., & Swiernik, M. A. (2021). The emerging clinical role of wearables: Factors for successful implementation in healthcare. NPJ Digital Medicine, 4(1), 45.
Torous, J., Kiang, M. V., Lorme, J., & Onnela, J. P. (2016). New tools for new research in psychiatry: A scalable and customizable platform to empower data-driven smartphone research. JMIR Mental Health, 3(2), e5165.
Wilson, R. B., Vangala, S., Elashoff, D., Safari, T., & Smith, B. A. (2021). Using wearable sensor technology to measure motion complexity in infants at high familial risk for autism spectrum disorder. Sensors, 21(2), 616.
